github.com/apache/arrow/go/v14@v14.0.2/internal/utils/_lib/transpose_ints_avx2_amd64.s (about)

     1  	.text
     2  	.intel_syntax noprefix
     3  	.file	"transpose_ints.c"
     4  	.globl	transpose_uint8_uint8_avx2      # -- Begin function transpose_uint8_uint8_avx2
     5  	.p2align	4, 0x90
     6  	.type	transpose_uint8_uint8_avx2,@function
     7  transpose_uint8_uint8_avx2:             # @transpose_uint8_uint8_avx2
     8  # %bb.0:
     9  	push	rbp
    10  	mov	rbp, rsp
    11  	and	rsp, -8
    12  	cmp	edx, 4
    13  	jl	.LBB0_1
    14  	.p2align	4, 0x90
    15  .LBB0_5:                                # =>This Inner Loop Header: Depth=1
    16  	mov	eax, edx
    17  	movzx	edx, byte ptr [rdi]
    18  	movzx	edx, byte ptr [rcx + 4*rdx]
    19  	mov	byte ptr [rsi], dl
    20  	movzx	edx, byte ptr [rdi + 1]
    21  	movzx	edx, byte ptr [rcx + 4*rdx]
    22  	mov	byte ptr [rsi + 1], dl
    23  	movzx	edx, byte ptr [rdi + 2]
    24  	movzx	edx, byte ptr [rcx + 4*rdx]
    25  	mov	byte ptr [rsi + 2], dl
    26  	movzx	edx, byte ptr [rdi + 3]
    27  	movzx	edx, byte ptr [rcx + 4*rdx]
    28  	mov	byte ptr [rsi + 3], dl
    29  	lea	edx, [rax - 4]
    30  	add	rdi, 4
    31  	add	rsi, 4
    32  	cmp	eax, 7
    33  	jg	.LBB0_5
    34  .LBB0_1:
    35  	test	edx, edx
    36  	jle	.LBB0_4
    37  # %bb.2:
    38  	add	edx, 1
    39  	xor	r8d, r8d
    40  	.p2align	4, 0x90
    41  .LBB0_3:                                # =>This Inner Loop Header: Depth=1
    42  	movzx	eax, byte ptr [rdi + r8]
    43  	movzx	eax, byte ptr [rcx + 4*rax]
    44  	mov	byte ptr [rsi + r8], al
    45  	add	r8, 1
    46  	add	edx, -1
    47  	cmp	edx, 1
    48  	jg	.LBB0_3
    49  .LBB0_4:
    50  	mov	rsp, rbp
    51  	pop	rbp
    52  	ret
    53  .Lfunc_end0:
    54  	.size	transpose_uint8_uint8_avx2, .Lfunc_end0-transpose_uint8_uint8_avx2
    55                                          # -- End function
    56  	.globl	transpose_int8_uint8_avx2       # -- Begin function transpose_int8_uint8_avx2
    57  	.p2align	4, 0x90
    58  	.type	transpose_int8_uint8_avx2,@function
    59  transpose_int8_uint8_avx2:              # @transpose_int8_uint8_avx2
    60  # %bb.0:
    61  	push	rbp
    62  	mov	rbp, rsp
    63  	and	rsp, -8
    64  	cmp	edx, 4
    65  	jl	.LBB1_1
    66  	.p2align	4, 0x90
    67  .LBB1_5:                                # =>This Inner Loop Header: Depth=1
    68  	mov	eax, edx
    69  	movsx	rdx, byte ptr [rdi]
    70  	movzx	edx, byte ptr [rcx + 4*rdx]
    71  	mov	byte ptr [rsi], dl
    72  	movsx	rdx, byte ptr [rdi + 1]
    73  	movzx	edx, byte ptr [rcx + 4*rdx]
    74  	mov	byte ptr [rsi + 1], dl
    75  	movsx	rdx, byte ptr [rdi + 2]
    76  	movzx	edx, byte ptr [rcx + 4*rdx]
    77  	mov	byte ptr [rsi + 2], dl
    78  	movsx	rdx, byte ptr [rdi + 3]
    79  	movzx	edx, byte ptr [rcx + 4*rdx]
    80  	mov	byte ptr [rsi + 3], dl
    81  	lea	edx, [rax - 4]
    82  	add	rdi, 4
    83  	add	rsi, 4
    84  	cmp	eax, 7
    85  	jg	.LBB1_5
    86  .LBB1_1:
    87  	test	edx, edx
    88  	jle	.LBB1_4
    89  # %bb.2:
    90  	add	edx, 1
    91  	xor	r8d, r8d
    92  	.p2align	4, 0x90
    93  .LBB1_3:                                # =>This Inner Loop Header: Depth=1
    94  	movsx	rax, byte ptr [rdi + r8]
    95  	movzx	eax, byte ptr [rcx + 4*rax]
    96  	mov	byte ptr [rsi + r8], al
    97  	add	r8, 1
    98  	add	edx, -1
    99  	cmp	edx, 1
   100  	jg	.LBB1_3
   101  .LBB1_4:
   102  	mov	rsp, rbp
   103  	pop	rbp
   104  	ret
   105  .Lfunc_end1:
   106  	.size	transpose_int8_uint8_avx2, .Lfunc_end1-transpose_int8_uint8_avx2
   107                                          # -- End function
   108  	.globl	transpose_uint16_uint8_avx2     # -- Begin function transpose_uint16_uint8_avx2
   109  	.p2align	4, 0x90
   110  	.type	transpose_uint16_uint8_avx2,@function
   111  transpose_uint16_uint8_avx2:            # @transpose_uint16_uint8_avx2
   112  # %bb.0:
   113  	push	rbp
   114  	mov	rbp, rsp
   115  	and	rsp, -8
   116  	cmp	edx, 4
   117  	jl	.LBB2_1
   118  	.p2align	4, 0x90
   119  .LBB2_5:                                # =>This Inner Loop Header: Depth=1
   120  	mov	eax, edx
   121  	movzx	edx, word ptr [rdi]
   122  	movzx	edx, byte ptr [rcx + 4*rdx]
   123  	mov	byte ptr [rsi], dl
   124  	movzx	edx, word ptr [rdi + 2]
   125  	movzx	edx, byte ptr [rcx + 4*rdx]
   126  	mov	byte ptr [rsi + 1], dl
   127  	movzx	edx, word ptr [rdi + 4]
   128  	movzx	edx, byte ptr [rcx + 4*rdx]
   129  	mov	byte ptr [rsi + 2], dl
   130  	movzx	edx, word ptr [rdi + 6]
   131  	movzx	edx, byte ptr [rcx + 4*rdx]
   132  	mov	byte ptr [rsi + 3], dl
   133  	lea	edx, [rax - 4]
   134  	add	rdi, 8
   135  	add	rsi, 4
   136  	cmp	eax, 7
   137  	jg	.LBB2_5
   138  .LBB2_1:
   139  	test	edx, edx
   140  	jle	.LBB2_4
   141  # %bb.2:
   142  	add	edx, 1
   143  	xor	r8d, r8d
   144  	.p2align	4, 0x90
   145  .LBB2_3:                                # =>This Inner Loop Header: Depth=1
   146  	movzx	eax, word ptr [rdi + 2*r8]
   147  	movzx	eax, byte ptr [rcx + 4*rax]
   148  	mov	byte ptr [rsi + r8], al
   149  	add	r8, 1
   150  	add	edx, -1
   151  	cmp	edx, 1
   152  	jg	.LBB2_3
   153  .LBB2_4:
   154  	mov	rsp, rbp
   155  	pop	rbp
   156  	ret
   157  .Lfunc_end2:
   158  	.size	transpose_uint16_uint8_avx2, .Lfunc_end2-transpose_uint16_uint8_avx2
   159                                          # -- End function
   160  	.globl	transpose_int16_uint8_avx2      # -- Begin function transpose_int16_uint8_avx2
   161  	.p2align	4, 0x90
   162  	.type	transpose_int16_uint8_avx2,@function
   163  transpose_int16_uint8_avx2:             # @transpose_int16_uint8_avx2
   164  # %bb.0:
   165  	push	rbp
   166  	mov	rbp, rsp
   167  	and	rsp, -8
   168  	cmp	edx, 4
   169  	jl	.LBB3_1
   170  	.p2align	4, 0x90
   171  .LBB3_5:                                # =>This Inner Loop Header: Depth=1
   172  	mov	eax, edx
   173  	movsx	rdx, word ptr [rdi]
   174  	movzx	edx, byte ptr [rcx + 4*rdx]
   175  	mov	byte ptr [rsi], dl
   176  	movsx	rdx, word ptr [rdi + 2]
   177  	movzx	edx, byte ptr [rcx + 4*rdx]
   178  	mov	byte ptr [rsi + 1], dl
   179  	movsx	rdx, word ptr [rdi + 4]
   180  	movzx	edx, byte ptr [rcx + 4*rdx]
   181  	mov	byte ptr [rsi + 2], dl
   182  	movsx	rdx, word ptr [rdi + 6]
   183  	movzx	edx, byte ptr [rcx + 4*rdx]
   184  	mov	byte ptr [rsi + 3], dl
   185  	lea	edx, [rax - 4]
   186  	add	rdi, 8
   187  	add	rsi, 4
   188  	cmp	eax, 7
   189  	jg	.LBB3_5
   190  .LBB3_1:
   191  	test	edx, edx
   192  	jle	.LBB3_4
   193  # %bb.2:
   194  	add	edx, 1
   195  	xor	r8d, r8d
   196  	.p2align	4, 0x90
   197  .LBB3_3:                                # =>This Inner Loop Header: Depth=1
   198  	movsx	rax, word ptr [rdi + 2*r8]
   199  	movzx	eax, byte ptr [rcx + 4*rax]
   200  	mov	byte ptr [rsi + r8], al
   201  	add	r8, 1
   202  	add	edx, -1
   203  	cmp	edx, 1
   204  	jg	.LBB3_3
   205  .LBB3_4:
   206  	mov	rsp, rbp
   207  	pop	rbp
   208  	ret
   209  .Lfunc_end3:
   210  	.size	transpose_int16_uint8_avx2, .Lfunc_end3-transpose_int16_uint8_avx2
   211                                          # -- End function
   212  	.globl	transpose_uint32_uint8_avx2     # -- Begin function transpose_uint32_uint8_avx2
   213  	.p2align	4, 0x90
   214  	.type	transpose_uint32_uint8_avx2,@function
   215  transpose_uint32_uint8_avx2:            # @transpose_uint32_uint8_avx2
   216  # %bb.0:
   217  	push	rbp
   218  	mov	rbp, rsp
   219  	and	rsp, -8
   220  	cmp	edx, 4
   221  	jl	.LBB4_1
   222  	.p2align	4, 0x90
   223  .LBB4_5:                                # =>This Inner Loop Header: Depth=1
   224  	mov	eax, edx
   225  	mov	edx, dword ptr [rdi]
   226  	movzx	edx, byte ptr [rcx + 4*rdx]
   227  	mov	byte ptr [rsi], dl
   228  	mov	edx, dword ptr [rdi + 4]
   229  	movzx	edx, byte ptr [rcx + 4*rdx]
   230  	mov	byte ptr [rsi + 1], dl
   231  	mov	edx, dword ptr [rdi + 8]
   232  	movzx	edx, byte ptr [rcx + 4*rdx]
   233  	mov	byte ptr [rsi + 2], dl
   234  	mov	edx, dword ptr [rdi + 12]
   235  	movzx	edx, byte ptr [rcx + 4*rdx]
   236  	mov	byte ptr [rsi + 3], dl
   237  	lea	edx, [rax - 4]
   238  	add	rdi, 16
   239  	add	rsi, 4
   240  	cmp	eax, 7
   241  	jg	.LBB4_5
   242  .LBB4_1:
   243  	test	edx, edx
   244  	jle	.LBB4_4
   245  # %bb.2:
   246  	add	edx, 1
   247  	xor	r8d, r8d
   248  	.p2align	4, 0x90
   249  .LBB4_3:                                # =>This Inner Loop Header: Depth=1
   250  	mov	eax, dword ptr [rdi + 4*r8]
   251  	movzx	eax, byte ptr [rcx + 4*rax]
   252  	mov	byte ptr [rsi + r8], al
   253  	add	r8, 1
   254  	add	edx, -1
   255  	cmp	edx, 1
   256  	jg	.LBB4_3
   257  .LBB4_4:
   258  	mov	rsp, rbp
   259  	pop	rbp
   260  	ret
   261  .Lfunc_end4:
   262  	.size	transpose_uint32_uint8_avx2, .Lfunc_end4-transpose_uint32_uint8_avx2
   263                                          # -- End function
   264  	.globl	transpose_int32_uint8_avx2      # -- Begin function transpose_int32_uint8_avx2
   265  	.p2align	4, 0x90
   266  	.type	transpose_int32_uint8_avx2,@function
   267  transpose_int32_uint8_avx2:             # @transpose_int32_uint8_avx2
   268  # %bb.0:
   269  	push	rbp
   270  	mov	rbp, rsp
   271  	and	rsp, -8
   272  	cmp	edx, 4
   273  	jl	.LBB5_1
   274  	.p2align	4, 0x90
   275  .LBB5_5:                                # =>This Inner Loop Header: Depth=1
   276  	mov	eax, edx
   277  	movsxd	rdx, dword ptr [rdi]
   278  	movzx	edx, byte ptr [rcx + 4*rdx]
   279  	mov	byte ptr [rsi], dl
   280  	movsxd	rdx, dword ptr [rdi + 4]
   281  	movzx	edx, byte ptr [rcx + 4*rdx]
   282  	mov	byte ptr [rsi + 1], dl
   283  	movsxd	rdx, dword ptr [rdi + 8]
   284  	movzx	edx, byte ptr [rcx + 4*rdx]
   285  	mov	byte ptr [rsi + 2], dl
   286  	movsxd	rdx, dword ptr [rdi + 12]
   287  	movzx	edx, byte ptr [rcx + 4*rdx]
   288  	mov	byte ptr [rsi + 3], dl
   289  	lea	edx, [rax - 4]
   290  	add	rdi, 16
   291  	add	rsi, 4
   292  	cmp	eax, 7
   293  	jg	.LBB5_5
   294  .LBB5_1:
   295  	test	edx, edx
   296  	jle	.LBB5_4
   297  # %bb.2:
   298  	add	edx, 1
   299  	xor	r8d, r8d
   300  	.p2align	4, 0x90
   301  .LBB5_3:                                # =>This Inner Loop Header: Depth=1
   302  	movsxd	rax, dword ptr [rdi + 4*r8]
   303  	movzx	eax, byte ptr [rcx + 4*rax]
   304  	mov	byte ptr [rsi + r8], al
   305  	add	r8, 1
   306  	add	edx, -1
   307  	cmp	edx, 1
   308  	jg	.LBB5_3
   309  .LBB5_4:
   310  	mov	rsp, rbp
   311  	pop	rbp
   312  	ret
   313  .Lfunc_end5:
   314  	.size	transpose_int32_uint8_avx2, .Lfunc_end5-transpose_int32_uint8_avx2
   315                                          # -- End function
   316  	.globl	transpose_uint64_uint8_avx2     # -- Begin function transpose_uint64_uint8_avx2
   317  	.p2align	4, 0x90
   318  	.type	transpose_uint64_uint8_avx2,@function
   319  transpose_uint64_uint8_avx2:            # @transpose_uint64_uint8_avx2
   320  # %bb.0:
   321  	push	rbp
   322  	mov	rbp, rsp
   323  	and	rsp, -8
   324  	cmp	edx, 4
   325  	jl	.LBB6_1
   326  	.p2align	4, 0x90
   327  .LBB6_5:                                # =>This Inner Loop Header: Depth=1
   328  	mov	eax, edx
   329  	mov	rdx, qword ptr [rdi]
   330  	movzx	edx, byte ptr [rcx + 4*rdx]
   331  	mov	byte ptr [rsi], dl
   332  	mov	rdx, qword ptr [rdi + 8]
   333  	movzx	edx, byte ptr [rcx + 4*rdx]
   334  	mov	byte ptr [rsi + 1], dl
   335  	mov	rdx, qword ptr [rdi + 16]
   336  	movzx	edx, byte ptr [rcx + 4*rdx]
   337  	mov	byte ptr [rsi + 2], dl
   338  	mov	rdx, qword ptr [rdi + 24]
   339  	movzx	edx, byte ptr [rcx + 4*rdx]
   340  	mov	byte ptr [rsi + 3], dl
   341  	lea	edx, [rax - 4]
   342  	add	rdi, 32
   343  	add	rsi, 4
   344  	cmp	eax, 7
   345  	jg	.LBB6_5
   346  .LBB6_1:
   347  	test	edx, edx
   348  	jle	.LBB6_4
   349  # %bb.2:
   350  	add	edx, 1
   351  	xor	r8d, r8d
   352  	.p2align	4, 0x90
   353  .LBB6_3:                                # =>This Inner Loop Header: Depth=1
   354  	mov	rax, qword ptr [rdi + 8*r8]
   355  	movzx	eax, byte ptr [rcx + 4*rax]
   356  	mov	byte ptr [rsi + r8], al
   357  	add	r8, 1
   358  	add	edx, -1
   359  	cmp	edx, 1
   360  	jg	.LBB6_3
   361  .LBB6_4:
   362  	mov	rsp, rbp
   363  	pop	rbp
   364  	ret
   365  .Lfunc_end6:
   366  	.size	transpose_uint64_uint8_avx2, .Lfunc_end6-transpose_uint64_uint8_avx2
   367                                          # -- End function
   368  	.globl	transpose_int64_uint8_avx2      # -- Begin function transpose_int64_uint8_avx2
   369  	.p2align	4, 0x90
   370  	.type	transpose_int64_uint8_avx2,@function
   371  transpose_int64_uint8_avx2:             # @transpose_int64_uint8_avx2
   372  # %bb.0:
   373  	push	rbp
   374  	mov	rbp, rsp
   375  	and	rsp, -8
   376  	cmp	edx, 4
   377  	jl	.LBB7_1
   378  	.p2align	4, 0x90
   379  .LBB7_5:                                # =>This Inner Loop Header: Depth=1
   380  	mov	eax, edx
   381  	mov	rdx, qword ptr [rdi]
   382  	movzx	edx, byte ptr [rcx + 4*rdx]
   383  	mov	byte ptr [rsi], dl
   384  	mov	rdx, qword ptr [rdi + 8]
   385  	movzx	edx, byte ptr [rcx + 4*rdx]
   386  	mov	byte ptr [rsi + 1], dl
   387  	mov	rdx, qword ptr [rdi + 16]
   388  	movzx	edx, byte ptr [rcx + 4*rdx]
   389  	mov	byte ptr [rsi + 2], dl
   390  	mov	rdx, qword ptr [rdi + 24]
   391  	movzx	edx, byte ptr [rcx + 4*rdx]
   392  	mov	byte ptr [rsi + 3], dl
   393  	lea	edx, [rax - 4]
   394  	add	rdi, 32
   395  	add	rsi, 4
   396  	cmp	eax, 7
   397  	jg	.LBB7_5
   398  .LBB7_1:
   399  	test	edx, edx
   400  	jle	.LBB7_4
   401  # %bb.2:
   402  	add	edx, 1
   403  	xor	r8d, r8d
   404  	.p2align	4, 0x90
   405  .LBB7_3:                                # =>This Inner Loop Header: Depth=1
   406  	mov	rax, qword ptr [rdi + 8*r8]
   407  	movzx	eax, byte ptr [rcx + 4*rax]
   408  	mov	byte ptr [rsi + r8], al
   409  	add	r8, 1
   410  	add	edx, -1
   411  	cmp	edx, 1
   412  	jg	.LBB7_3
   413  .LBB7_4:
   414  	mov	rsp, rbp
   415  	pop	rbp
   416  	ret
   417  .Lfunc_end7:
   418  	.size	transpose_int64_uint8_avx2, .Lfunc_end7-transpose_int64_uint8_avx2
   419                                          # -- End function
   420  	.globl	transpose_uint8_int8_avx2       # -- Begin function transpose_uint8_int8_avx2
   421  	.p2align	4, 0x90
   422  	.type	transpose_uint8_int8_avx2,@function
   423  transpose_uint8_int8_avx2:              # @transpose_uint8_int8_avx2
   424  # %bb.0:
   425  	push	rbp
   426  	mov	rbp, rsp
   427  	and	rsp, -8
   428  	cmp	edx, 4
   429  	jl	.LBB8_1
   430  	.p2align	4, 0x90
   431  .LBB8_5:                                # =>This Inner Loop Header: Depth=1
   432  	mov	eax, edx
   433  	movzx	edx, byte ptr [rdi]
   434  	movzx	edx, byte ptr [rcx + 4*rdx]
   435  	mov	byte ptr [rsi], dl
   436  	movzx	edx, byte ptr [rdi + 1]
   437  	movzx	edx, byte ptr [rcx + 4*rdx]
   438  	mov	byte ptr [rsi + 1], dl
   439  	movzx	edx, byte ptr [rdi + 2]
   440  	movzx	edx, byte ptr [rcx + 4*rdx]
   441  	mov	byte ptr [rsi + 2], dl
   442  	movzx	edx, byte ptr [rdi + 3]
   443  	movzx	edx, byte ptr [rcx + 4*rdx]
   444  	mov	byte ptr [rsi + 3], dl
   445  	lea	edx, [rax - 4]
   446  	add	rdi, 4
   447  	add	rsi, 4
   448  	cmp	eax, 7
   449  	jg	.LBB8_5
   450  .LBB8_1:
   451  	test	edx, edx
   452  	jle	.LBB8_4
   453  # %bb.2:
   454  	add	edx, 1
   455  	xor	r8d, r8d
   456  	.p2align	4, 0x90
   457  .LBB8_3:                                # =>This Inner Loop Header: Depth=1
   458  	movzx	eax, byte ptr [rdi + r8]
   459  	movzx	eax, byte ptr [rcx + 4*rax]
   460  	mov	byte ptr [rsi + r8], al
   461  	add	r8, 1
   462  	add	edx, -1
   463  	cmp	edx, 1
   464  	jg	.LBB8_3
   465  .LBB8_4:
   466  	mov	rsp, rbp
   467  	pop	rbp
   468  	ret
   469  .Lfunc_end8:
   470  	.size	transpose_uint8_int8_avx2, .Lfunc_end8-transpose_uint8_int8_avx2
   471                                          # -- End function
   472  	.globl	transpose_int8_int8_avx2        # -- Begin function transpose_int8_int8_avx2
   473  	.p2align	4, 0x90
   474  	.type	transpose_int8_int8_avx2,@function
   475  transpose_int8_int8_avx2:               # @transpose_int8_int8_avx2
   476  # %bb.0:
   477  	push	rbp
   478  	mov	rbp, rsp
   479  	and	rsp, -8
   480  	cmp	edx, 4
   481  	jl	.LBB9_1
   482  	.p2align	4, 0x90
   483  .LBB9_5:                                # =>This Inner Loop Header: Depth=1
   484  	mov	eax, edx
   485  	movsx	rdx, byte ptr [rdi]
   486  	movzx	edx, byte ptr [rcx + 4*rdx]
   487  	mov	byte ptr [rsi], dl
   488  	movsx	rdx, byte ptr [rdi + 1]
   489  	movzx	edx, byte ptr [rcx + 4*rdx]
   490  	mov	byte ptr [rsi + 1], dl
   491  	movsx	rdx, byte ptr [rdi + 2]
   492  	movzx	edx, byte ptr [rcx + 4*rdx]
   493  	mov	byte ptr [rsi + 2], dl
   494  	movsx	rdx, byte ptr [rdi + 3]
   495  	movzx	edx, byte ptr [rcx + 4*rdx]
   496  	mov	byte ptr [rsi + 3], dl
   497  	lea	edx, [rax - 4]
   498  	add	rdi, 4
   499  	add	rsi, 4
   500  	cmp	eax, 7
   501  	jg	.LBB9_5
   502  .LBB9_1:
   503  	test	edx, edx
   504  	jle	.LBB9_4
   505  # %bb.2:
   506  	add	edx, 1
   507  	xor	r8d, r8d
   508  	.p2align	4, 0x90
   509  .LBB9_3:                                # =>This Inner Loop Header: Depth=1
   510  	movsx	rax, byte ptr [rdi + r8]
   511  	movzx	eax, byte ptr [rcx + 4*rax]
   512  	mov	byte ptr [rsi + r8], al
   513  	add	r8, 1
   514  	add	edx, -1
   515  	cmp	edx, 1
   516  	jg	.LBB9_3
   517  .LBB9_4:
   518  	mov	rsp, rbp
   519  	pop	rbp
   520  	ret
   521  .Lfunc_end9:
   522  	.size	transpose_int8_int8_avx2, .Lfunc_end9-transpose_int8_int8_avx2
   523                                          # -- End function
   524  	.globl	transpose_uint16_int8_avx2      # -- Begin function transpose_uint16_int8_avx2
   525  	.p2align	4, 0x90
   526  	.type	transpose_uint16_int8_avx2,@function
   527  transpose_uint16_int8_avx2:             # @transpose_uint16_int8_avx2
   528  # %bb.0:
   529  	push	rbp
   530  	mov	rbp, rsp
   531  	and	rsp, -8
   532  	cmp	edx, 4
   533  	jl	.LBB10_1
   534  	.p2align	4, 0x90
   535  .LBB10_5:                               # =>This Inner Loop Header: Depth=1
   536  	mov	eax, edx
   537  	movzx	edx, word ptr [rdi]
   538  	movzx	edx, byte ptr [rcx + 4*rdx]
   539  	mov	byte ptr [rsi], dl
   540  	movzx	edx, word ptr [rdi + 2]
   541  	movzx	edx, byte ptr [rcx + 4*rdx]
   542  	mov	byte ptr [rsi + 1], dl
   543  	movzx	edx, word ptr [rdi + 4]
   544  	movzx	edx, byte ptr [rcx + 4*rdx]
   545  	mov	byte ptr [rsi + 2], dl
   546  	movzx	edx, word ptr [rdi + 6]
   547  	movzx	edx, byte ptr [rcx + 4*rdx]
   548  	mov	byte ptr [rsi + 3], dl
   549  	lea	edx, [rax - 4]
   550  	add	rdi, 8
   551  	add	rsi, 4
   552  	cmp	eax, 7
   553  	jg	.LBB10_5
   554  .LBB10_1:
   555  	test	edx, edx
   556  	jle	.LBB10_4
   557  # %bb.2:
   558  	add	edx, 1
   559  	xor	r8d, r8d
   560  	.p2align	4, 0x90
   561  .LBB10_3:                               # =>This Inner Loop Header: Depth=1
   562  	movzx	eax, word ptr [rdi + 2*r8]
   563  	movzx	eax, byte ptr [rcx + 4*rax]
   564  	mov	byte ptr [rsi + r8], al
   565  	add	r8, 1
   566  	add	edx, -1
   567  	cmp	edx, 1
   568  	jg	.LBB10_3
   569  .LBB10_4:
   570  	mov	rsp, rbp
   571  	pop	rbp
   572  	ret
   573  .Lfunc_end10:
   574  	.size	transpose_uint16_int8_avx2, .Lfunc_end10-transpose_uint16_int8_avx2
   575                                          # -- End function
   576  	.globl	transpose_int16_int8_avx2       # -- Begin function transpose_int16_int8_avx2
   577  	.p2align	4, 0x90
   578  	.type	transpose_int16_int8_avx2,@function
   579  transpose_int16_int8_avx2:              # @transpose_int16_int8_avx2
   580  # %bb.0:
   581  	push	rbp
   582  	mov	rbp, rsp
   583  	and	rsp, -8
   584  	cmp	edx, 4
   585  	jl	.LBB11_1
   586  	.p2align	4, 0x90
   587  .LBB11_5:                               # =>This Inner Loop Header: Depth=1
   588  	mov	eax, edx
   589  	movsx	rdx, word ptr [rdi]
   590  	movzx	edx, byte ptr [rcx + 4*rdx]
   591  	mov	byte ptr [rsi], dl
   592  	movsx	rdx, word ptr [rdi + 2]
   593  	movzx	edx, byte ptr [rcx + 4*rdx]
   594  	mov	byte ptr [rsi + 1], dl
   595  	movsx	rdx, word ptr [rdi + 4]
   596  	movzx	edx, byte ptr [rcx + 4*rdx]
   597  	mov	byte ptr [rsi + 2], dl
   598  	movsx	rdx, word ptr [rdi + 6]
   599  	movzx	edx, byte ptr [rcx + 4*rdx]
   600  	mov	byte ptr [rsi + 3], dl
   601  	lea	edx, [rax - 4]
   602  	add	rdi, 8
   603  	add	rsi, 4
   604  	cmp	eax, 7
   605  	jg	.LBB11_5
   606  .LBB11_1:
   607  	test	edx, edx
   608  	jle	.LBB11_4
   609  # %bb.2:
   610  	add	edx, 1
   611  	xor	r8d, r8d
   612  	.p2align	4, 0x90
   613  .LBB11_3:                               # =>This Inner Loop Header: Depth=1
   614  	movsx	rax, word ptr [rdi + 2*r8]
   615  	movzx	eax, byte ptr [rcx + 4*rax]
   616  	mov	byte ptr [rsi + r8], al
   617  	add	r8, 1
   618  	add	edx, -1
   619  	cmp	edx, 1
   620  	jg	.LBB11_3
   621  .LBB11_4:
   622  	mov	rsp, rbp
   623  	pop	rbp
   624  	ret
   625  .Lfunc_end11:
   626  	.size	transpose_int16_int8_avx2, .Lfunc_end11-transpose_int16_int8_avx2
   627                                          # -- End function
   628  	.globl	transpose_uint32_int8_avx2      # -- Begin function transpose_uint32_int8_avx2
   629  	.p2align	4, 0x90
   630  	.type	transpose_uint32_int8_avx2,@function
   631  transpose_uint32_int8_avx2:             # @transpose_uint32_int8_avx2
   632  # %bb.0:
   633  	push	rbp
   634  	mov	rbp, rsp
   635  	and	rsp, -8
   636  	cmp	edx, 4
   637  	jl	.LBB12_1
   638  	.p2align	4, 0x90
   639  .LBB12_5:                               # =>This Inner Loop Header: Depth=1
   640  	mov	eax, edx
   641  	mov	edx, dword ptr [rdi]
   642  	movzx	edx, byte ptr [rcx + 4*rdx]
   643  	mov	byte ptr [rsi], dl
   644  	mov	edx, dword ptr [rdi + 4]
   645  	movzx	edx, byte ptr [rcx + 4*rdx]
   646  	mov	byte ptr [rsi + 1], dl
   647  	mov	edx, dword ptr [rdi + 8]
   648  	movzx	edx, byte ptr [rcx + 4*rdx]
   649  	mov	byte ptr [rsi + 2], dl
   650  	mov	edx, dword ptr [rdi + 12]
   651  	movzx	edx, byte ptr [rcx + 4*rdx]
   652  	mov	byte ptr [rsi + 3], dl
   653  	lea	edx, [rax - 4]
   654  	add	rdi, 16
   655  	add	rsi, 4
   656  	cmp	eax, 7
   657  	jg	.LBB12_5
   658  .LBB12_1:
   659  	test	edx, edx
   660  	jle	.LBB12_4
   661  # %bb.2:
   662  	add	edx, 1
   663  	xor	r8d, r8d
   664  	.p2align	4, 0x90
   665  .LBB12_3:                               # =>This Inner Loop Header: Depth=1
   666  	mov	eax, dword ptr [rdi + 4*r8]
   667  	movzx	eax, byte ptr [rcx + 4*rax]
   668  	mov	byte ptr [rsi + r8], al
   669  	add	r8, 1
   670  	add	edx, -1
   671  	cmp	edx, 1
   672  	jg	.LBB12_3
   673  .LBB12_4:
   674  	mov	rsp, rbp
   675  	pop	rbp
   676  	ret
   677  .Lfunc_end12:
   678  	.size	transpose_uint32_int8_avx2, .Lfunc_end12-transpose_uint32_int8_avx2
   679                                          # -- End function
   680  	.globl	transpose_int32_int8_avx2       # -- Begin function transpose_int32_int8_avx2
   681  	.p2align	4, 0x90
   682  	.type	transpose_int32_int8_avx2,@function
   683  transpose_int32_int8_avx2:              # @transpose_int32_int8_avx2
   684  # %bb.0:
   685  	push	rbp
   686  	mov	rbp, rsp
   687  	and	rsp, -8
   688  	cmp	edx, 4
   689  	jl	.LBB13_1
   690  	.p2align	4, 0x90
   691  .LBB13_5:                               # =>This Inner Loop Header: Depth=1
   692  	mov	eax, edx
   693  	movsxd	rdx, dword ptr [rdi]
   694  	movzx	edx, byte ptr [rcx + 4*rdx]
   695  	mov	byte ptr [rsi], dl
   696  	movsxd	rdx, dword ptr [rdi + 4]
   697  	movzx	edx, byte ptr [rcx + 4*rdx]
   698  	mov	byte ptr [rsi + 1], dl
   699  	movsxd	rdx, dword ptr [rdi + 8]
   700  	movzx	edx, byte ptr [rcx + 4*rdx]
   701  	mov	byte ptr [rsi + 2], dl
   702  	movsxd	rdx, dword ptr [rdi + 12]
   703  	movzx	edx, byte ptr [rcx + 4*rdx]
   704  	mov	byte ptr [rsi + 3], dl
   705  	lea	edx, [rax - 4]
   706  	add	rdi, 16
   707  	add	rsi, 4
   708  	cmp	eax, 7
   709  	jg	.LBB13_5
   710  .LBB13_1:
   711  	test	edx, edx
   712  	jle	.LBB13_4
   713  # %bb.2:
   714  	add	edx, 1
   715  	xor	r8d, r8d
   716  	.p2align	4, 0x90
   717  .LBB13_3:                               # =>This Inner Loop Header: Depth=1
   718  	movsxd	rax, dword ptr [rdi + 4*r8]
   719  	movzx	eax, byte ptr [rcx + 4*rax]
   720  	mov	byte ptr [rsi + r8], al
   721  	add	r8, 1
   722  	add	edx, -1
   723  	cmp	edx, 1
   724  	jg	.LBB13_3
   725  .LBB13_4:
   726  	mov	rsp, rbp
   727  	pop	rbp
   728  	ret
   729  .Lfunc_end13:
   730  	.size	transpose_int32_int8_avx2, .Lfunc_end13-transpose_int32_int8_avx2
   731                                          # -- End function
   732  	.globl	transpose_uint64_int8_avx2      # -- Begin function transpose_uint64_int8_avx2
   733  	.p2align	4, 0x90
   734  	.type	transpose_uint64_int8_avx2,@function
   735  transpose_uint64_int8_avx2:             # @transpose_uint64_int8_avx2
   736  # %bb.0:
   737  	push	rbp
   738  	mov	rbp, rsp
   739  	and	rsp, -8
   740  	cmp	edx, 4
   741  	jl	.LBB14_1
   742  	.p2align	4, 0x90
   743  .LBB14_5:                               # =>This Inner Loop Header: Depth=1
   744  	mov	eax, edx
   745  	mov	rdx, qword ptr [rdi]
   746  	movzx	edx, byte ptr [rcx + 4*rdx]
   747  	mov	byte ptr [rsi], dl
   748  	mov	rdx, qword ptr [rdi + 8]
   749  	movzx	edx, byte ptr [rcx + 4*rdx]
   750  	mov	byte ptr [rsi + 1], dl
   751  	mov	rdx, qword ptr [rdi + 16]
   752  	movzx	edx, byte ptr [rcx + 4*rdx]
   753  	mov	byte ptr [rsi + 2], dl
   754  	mov	rdx, qword ptr [rdi + 24]
   755  	movzx	edx, byte ptr [rcx + 4*rdx]
   756  	mov	byte ptr [rsi + 3], dl
   757  	lea	edx, [rax - 4]
   758  	add	rdi, 32
   759  	add	rsi, 4
   760  	cmp	eax, 7
   761  	jg	.LBB14_5
   762  .LBB14_1:
   763  	test	edx, edx
   764  	jle	.LBB14_4
   765  # %bb.2:
   766  	add	edx, 1
   767  	xor	r8d, r8d
   768  	.p2align	4, 0x90
   769  .LBB14_3:                               # =>This Inner Loop Header: Depth=1
   770  	mov	rax, qword ptr [rdi + 8*r8]
   771  	movzx	eax, byte ptr [rcx + 4*rax]
   772  	mov	byte ptr [rsi + r8], al
   773  	add	r8, 1
   774  	add	edx, -1
   775  	cmp	edx, 1
   776  	jg	.LBB14_3
   777  .LBB14_4:
   778  	mov	rsp, rbp
   779  	pop	rbp
   780  	ret
   781  .Lfunc_end14:
   782  	.size	transpose_uint64_int8_avx2, .Lfunc_end14-transpose_uint64_int8_avx2
   783                                          # -- End function
   784  	.globl	transpose_int64_int8_avx2       # -- Begin function transpose_int64_int8_avx2
   785  	.p2align	4, 0x90
   786  	.type	transpose_int64_int8_avx2,@function
   787  transpose_int64_int8_avx2:              # @transpose_int64_int8_avx2
   788  # %bb.0:
   789  	push	rbp
   790  	mov	rbp, rsp
   791  	and	rsp, -8
   792  	cmp	edx, 4
   793  	jl	.LBB15_1
   794  	.p2align	4, 0x90
   795  .LBB15_5:                               # =>This Inner Loop Header: Depth=1
   796  	mov	eax, edx
   797  	mov	rdx, qword ptr [rdi]
   798  	movzx	edx, byte ptr [rcx + 4*rdx]
   799  	mov	byte ptr [rsi], dl
   800  	mov	rdx, qword ptr [rdi + 8]
   801  	movzx	edx, byte ptr [rcx + 4*rdx]
   802  	mov	byte ptr [rsi + 1], dl
   803  	mov	rdx, qword ptr [rdi + 16]
   804  	movzx	edx, byte ptr [rcx + 4*rdx]
   805  	mov	byte ptr [rsi + 2], dl
   806  	mov	rdx, qword ptr [rdi + 24]
   807  	movzx	edx, byte ptr [rcx + 4*rdx]
   808  	mov	byte ptr [rsi + 3], dl
   809  	lea	edx, [rax - 4]
   810  	add	rdi, 32
   811  	add	rsi, 4
   812  	cmp	eax, 7
   813  	jg	.LBB15_5
   814  .LBB15_1:
   815  	test	edx, edx
   816  	jle	.LBB15_4
   817  # %bb.2:
   818  	add	edx, 1
   819  	xor	r8d, r8d
   820  	.p2align	4, 0x90
   821  .LBB15_3:                               # =>This Inner Loop Header: Depth=1
   822  	mov	rax, qword ptr [rdi + 8*r8]
   823  	movzx	eax, byte ptr [rcx + 4*rax]
   824  	mov	byte ptr [rsi + r8], al
   825  	add	r8, 1
   826  	add	edx, -1
   827  	cmp	edx, 1
   828  	jg	.LBB15_3
   829  .LBB15_4:
   830  	mov	rsp, rbp
   831  	pop	rbp
   832  	ret
   833  .Lfunc_end15:
   834  	.size	transpose_int64_int8_avx2, .Lfunc_end15-transpose_int64_int8_avx2
   835                                          # -- End function
   836  	.globl	transpose_uint8_uint16_avx2     # -- Begin function transpose_uint8_uint16_avx2
   837  	.p2align	4, 0x90
   838  	.type	transpose_uint8_uint16_avx2,@function
   839  transpose_uint8_uint16_avx2:            # @transpose_uint8_uint16_avx2
   840  # %bb.0:
   841  	push	rbp
   842  	mov	rbp, rsp
   843  	and	rsp, -8
   844  	cmp	edx, 4
   845  	jl	.LBB16_1
   846  	.p2align	4, 0x90
   847  .LBB16_5:                               # =>This Inner Loop Header: Depth=1
   848  	mov	eax, edx
   849  	movzx	edx, byte ptr [rdi]
   850  	movzx	edx, word ptr [rcx + 4*rdx]
   851  	mov	word ptr [rsi], dx
   852  	movzx	edx, byte ptr [rdi + 1]
   853  	movzx	edx, word ptr [rcx + 4*rdx]
   854  	mov	word ptr [rsi + 2], dx
   855  	movzx	edx, byte ptr [rdi + 2]
   856  	movzx	edx, word ptr [rcx + 4*rdx]
   857  	mov	word ptr [rsi + 4], dx
   858  	movzx	edx, byte ptr [rdi + 3]
   859  	movzx	edx, word ptr [rcx + 4*rdx]
   860  	mov	word ptr [rsi + 6], dx
   861  	lea	edx, [rax - 4]
   862  	add	rdi, 4
   863  	add	rsi, 8
   864  	cmp	eax, 7
   865  	jg	.LBB16_5
   866  .LBB16_1:
   867  	test	edx, edx
   868  	jle	.LBB16_4
   869  # %bb.2:
   870  	add	edx, 1
   871  	xor	r8d, r8d
   872  	.p2align	4, 0x90
   873  .LBB16_3:                               # =>This Inner Loop Header: Depth=1
   874  	movzx	eax, byte ptr [rdi + r8]
   875  	movzx	eax, word ptr [rcx + 4*rax]
   876  	mov	word ptr [rsi + 2*r8], ax
   877  	add	r8, 1
   878  	add	edx, -1
   879  	cmp	edx, 1
   880  	jg	.LBB16_3
   881  .LBB16_4:
   882  	mov	rsp, rbp
   883  	pop	rbp
   884  	ret
   885  .Lfunc_end16:
   886  	.size	transpose_uint8_uint16_avx2, .Lfunc_end16-transpose_uint8_uint16_avx2
   887                                          # -- End function
   888  	.globl	transpose_int8_uint16_avx2      # -- Begin function transpose_int8_uint16_avx2
   889  	.p2align	4, 0x90
   890  	.type	transpose_int8_uint16_avx2,@function
   891  transpose_int8_uint16_avx2:             # @transpose_int8_uint16_avx2
   892  # %bb.0:
   893  	push	rbp
   894  	mov	rbp, rsp
   895  	and	rsp, -8
   896  	cmp	edx, 4
   897  	jl	.LBB17_1
   898  	.p2align	4, 0x90
   899  .LBB17_5:                               # =>This Inner Loop Header: Depth=1
   900  	mov	eax, edx
   901  	movsx	rdx, byte ptr [rdi]
   902  	movzx	edx, word ptr [rcx + 4*rdx]
   903  	mov	word ptr [rsi], dx
   904  	movsx	rdx, byte ptr [rdi + 1]
   905  	movzx	edx, word ptr [rcx + 4*rdx]
   906  	mov	word ptr [rsi + 2], dx
   907  	movsx	rdx, byte ptr [rdi + 2]
   908  	movzx	edx, word ptr [rcx + 4*rdx]
   909  	mov	word ptr [rsi + 4], dx
   910  	movsx	rdx, byte ptr [rdi + 3]
   911  	movzx	edx, word ptr [rcx + 4*rdx]
   912  	mov	word ptr [rsi + 6], dx
   913  	lea	edx, [rax - 4]
   914  	add	rdi, 4
   915  	add	rsi, 8
   916  	cmp	eax, 7
   917  	jg	.LBB17_5
   918  .LBB17_1:
   919  	test	edx, edx
   920  	jle	.LBB17_4
   921  # %bb.2:
   922  	add	edx, 1
   923  	xor	r8d, r8d
   924  	.p2align	4, 0x90
   925  .LBB17_3:                               # =>This Inner Loop Header: Depth=1
   926  	movsx	rax, byte ptr [rdi + r8]
   927  	movzx	eax, word ptr [rcx + 4*rax]
   928  	mov	word ptr [rsi + 2*r8], ax
   929  	add	r8, 1
   930  	add	edx, -1
   931  	cmp	edx, 1
   932  	jg	.LBB17_3
   933  .LBB17_4:
   934  	mov	rsp, rbp
   935  	pop	rbp
   936  	ret
   937  .Lfunc_end17:
   938  	.size	transpose_int8_uint16_avx2, .Lfunc_end17-transpose_int8_uint16_avx2
   939                                          # -- End function
   940  	.globl	transpose_uint16_uint16_avx2    # -- Begin function transpose_uint16_uint16_avx2
   941  	.p2align	4, 0x90
   942  	.type	transpose_uint16_uint16_avx2,@function
   943  transpose_uint16_uint16_avx2:           # @transpose_uint16_uint16_avx2
   944  # %bb.0:
   945  	push	rbp
   946  	mov	rbp, rsp
   947  	and	rsp, -8
   948  	cmp	edx, 4
   949  	jl	.LBB18_1
   950  	.p2align	4, 0x90
   951  .LBB18_5:                               # =>This Inner Loop Header: Depth=1
   952  	mov	eax, edx
   953  	movzx	edx, word ptr [rdi]
   954  	movzx	edx, word ptr [rcx + 4*rdx]
   955  	mov	word ptr [rsi], dx
   956  	movzx	edx, word ptr [rdi + 2]
   957  	movzx	edx, word ptr [rcx + 4*rdx]
   958  	mov	word ptr [rsi + 2], dx
   959  	movzx	edx, word ptr [rdi + 4]
   960  	movzx	edx, word ptr [rcx + 4*rdx]
   961  	mov	word ptr [rsi + 4], dx
   962  	movzx	edx, word ptr [rdi + 6]
   963  	movzx	edx, word ptr [rcx + 4*rdx]
   964  	mov	word ptr [rsi + 6], dx
   965  	lea	edx, [rax - 4]
   966  	add	rdi, 8
   967  	add	rsi, 8
   968  	cmp	eax, 7
   969  	jg	.LBB18_5
   970  .LBB18_1:
   971  	test	edx, edx
   972  	jle	.LBB18_4
   973  # %bb.2:
   974  	add	edx, 1
   975  	xor	r8d, r8d
   976  	.p2align	4, 0x90
   977  .LBB18_3:                               # =>This Inner Loop Header: Depth=1
   978  	movzx	eax, word ptr [rdi + r8]
   979  	movzx	eax, word ptr [rcx + 4*rax]
   980  	mov	word ptr [rsi + r8], ax
   981  	add	r8, 2
   982  	add	edx, -1
   983  	cmp	edx, 1
   984  	jg	.LBB18_3
   985  .LBB18_4:
   986  	mov	rsp, rbp
   987  	pop	rbp
   988  	ret
   989  .Lfunc_end18:
   990  	.size	transpose_uint16_uint16_avx2, .Lfunc_end18-transpose_uint16_uint16_avx2
   991                                          # -- End function
   992  	.globl	transpose_int16_uint16_avx2     # -- Begin function transpose_int16_uint16_avx2
   993  	.p2align	4, 0x90
   994  	.type	transpose_int16_uint16_avx2,@function
   995  transpose_int16_uint16_avx2:            # @transpose_int16_uint16_avx2
   996  # %bb.0:
   997  	push	rbp
   998  	mov	rbp, rsp
   999  	and	rsp, -8
  1000  	cmp	edx, 4
  1001  	jl	.LBB19_1
  1002  	.p2align	4, 0x90
  1003  .LBB19_5:                               # =>This Inner Loop Header: Depth=1
  1004  	mov	eax, edx
  1005  	movsx	rdx, word ptr [rdi]
  1006  	movzx	edx, word ptr [rcx + 4*rdx]
  1007  	mov	word ptr [rsi], dx
  1008  	movsx	rdx, word ptr [rdi + 2]
  1009  	movzx	edx, word ptr [rcx + 4*rdx]
  1010  	mov	word ptr [rsi + 2], dx
  1011  	movsx	rdx, word ptr [rdi + 4]
  1012  	movzx	edx, word ptr [rcx + 4*rdx]
  1013  	mov	word ptr [rsi + 4], dx
  1014  	movsx	rdx, word ptr [rdi + 6]
  1015  	movzx	edx, word ptr [rcx + 4*rdx]
  1016  	mov	word ptr [rsi + 6], dx
  1017  	lea	edx, [rax - 4]
  1018  	add	rdi, 8
  1019  	add	rsi, 8
  1020  	cmp	eax, 7
  1021  	jg	.LBB19_5
  1022  .LBB19_1:
  1023  	test	edx, edx
  1024  	jle	.LBB19_4
  1025  # %bb.2:
  1026  	add	edx, 1
  1027  	xor	r8d, r8d
  1028  	.p2align	4, 0x90
  1029  .LBB19_3:                               # =>This Inner Loop Header: Depth=1
  1030  	movsx	rax, word ptr [rdi + r8]
  1031  	movzx	eax, word ptr [rcx + 4*rax]
  1032  	mov	word ptr [rsi + r8], ax
  1033  	add	r8, 2
  1034  	add	edx, -1
  1035  	cmp	edx, 1
  1036  	jg	.LBB19_3
  1037  .LBB19_4:
  1038  	mov	rsp, rbp
  1039  	pop	rbp
  1040  	ret
  1041  .Lfunc_end19:
  1042  	.size	transpose_int16_uint16_avx2, .Lfunc_end19-transpose_int16_uint16_avx2
  1043                                          # -- End function
  1044  	.globl	transpose_uint32_uint16_avx2    # -- Begin function transpose_uint32_uint16_avx2
  1045  	.p2align	4, 0x90
  1046  	.type	transpose_uint32_uint16_avx2,@function
  1047  transpose_uint32_uint16_avx2:           # @transpose_uint32_uint16_avx2
  1048  # %bb.0:
  1049  	push	rbp
  1050  	mov	rbp, rsp
  1051  	and	rsp, -8
  1052  	cmp	edx, 4
  1053  	jl	.LBB20_1
  1054  	.p2align	4, 0x90
  1055  .LBB20_5:                               # =>This Inner Loop Header: Depth=1
  1056  	mov	eax, edx
  1057  	mov	edx, dword ptr [rdi]
  1058  	movzx	edx, word ptr [rcx + 4*rdx]
  1059  	mov	word ptr [rsi], dx
  1060  	mov	edx, dword ptr [rdi + 4]
  1061  	movzx	edx, word ptr [rcx + 4*rdx]
  1062  	mov	word ptr [rsi + 2], dx
  1063  	mov	edx, dword ptr [rdi + 8]
  1064  	movzx	edx, word ptr [rcx + 4*rdx]
  1065  	mov	word ptr [rsi + 4], dx
  1066  	mov	edx, dword ptr [rdi + 12]
  1067  	movzx	edx, word ptr [rcx + 4*rdx]
  1068  	mov	word ptr [rsi + 6], dx
  1069  	lea	edx, [rax - 4]
  1070  	add	rdi, 16
  1071  	add	rsi, 8
  1072  	cmp	eax, 7
  1073  	jg	.LBB20_5
  1074  .LBB20_1:
  1075  	test	edx, edx
  1076  	jle	.LBB20_4
  1077  # %bb.2:
  1078  	add	edx, 1
  1079  	xor	r8d, r8d
  1080  	.p2align	4, 0x90
  1081  .LBB20_3:                               # =>This Inner Loop Header: Depth=1
  1082  	mov	eax, dword ptr [rdi + 2*r8]
  1083  	movzx	eax, word ptr [rcx + 4*rax]
  1084  	mov	word ptr [rsi + r8], ax
  1085  	add	r8, 2
  1086  	add	edx, -1
  1087  	cmp	edx, 1
  1088  	jg	.LBB20_3
  1089  .LBB20_4:
  1090  	mov	rsp, rbp
  1091  	pop	rbp
  1092  	ret
  1093  .Lfunc_end20:
  1094  	.size	transpose_uint32_uint16_avx2, .Lfunc_end20-transpose_uint32_uint16_avx2
  1095                                          # -- End function
  1096  	.globl	transpose_int32_uint16_avx2     # -- Begin function transpose_int32_uint16_avx2
  1097  	.p2align	4, 0x90
  1098  	.type	transpose_int32_uint16_avx2,@function
  1099  transpose_int32_uint16_avx2:            # @transpose_int32_uint16_avx2
  1100  # %bb.0:
  1101  	push	rbp
  1102  	mov	rbp, rsp
  1103  	and	rsp, -8
  1104  	cmp	edx, 4
  1105  	jl	.LBB21_1
  1106  	.p2align	4, 0x90
  1107  .LBB21_5:                               # =>This Inner Loop Header: Depth=1
  1108  	mov	eax, edx
  1109  	movsxd	rdx, dword ptr [rdi]
  1110  	movzx	edx, word ptr [rcx + 4*rdx]
  1111  	mov	word ptr [rsi], dx
  1112  	movsxd	rdx, dword ptr [rdi + 4]
  1113  	movzx	edx, word ptr [rcx + 4*rdx]
  1114  	mov	word ptr [rsi + 2], dx
  1115  	movsxd	rdx, dword ptr [rdi + 8]
  1116  	movzx	edx, word ptr [rcx + 4*rdx]
  1117  	mov	word ptr [rsi + 4], dx
  1118  	movsxd	rdx, dword ptr [rdi + 12]
  1119  	movzx	edx, word ptr [rcx + 4*rdx]
  1120  	mov	word ptr [rsi + 6], dx
  1121  	lea	edx, [rax - 4]
  1122  	add	rdi, 16
  1123  	add	rsi, 8
  1124  	cmp	eax, 7
  1125  	jg	.LBB21_5
  1126  .LBB21_1:
  1127  	test	edx, edx
  1128  	jle	.LBB21_4
  1129  # %bb.2:
  1130  	add	edx, 1
  1131  	xor	r8d, r8d
  1132  	.p2align	4, 0x90
  1133  .LBB21_3:                               # =>This Inner Loop Header: Depth=1
  1134  	movsxd	rax, dword ptr [rdi + 2*r8]
  1135  	movzx	eax, word ptr [rcx + 4*rax]
  1136  	mov	word ptr [rsi + r8], ax
  1137  	add	r8, 2
  1138  	add	edx, -1
  1139  	cmp	edx, 1
  1140  	jg	.LBB21_3
  1141  .LBB21_4:
  1142  	mov	rsp, rbp
  1143  	pop	rbp
  1144  	ret
  1145  .Lfunc_end21:
  1146  	.size	transpose_int32_uint16_avx2, .Lfunc_end21-transpose_int32_uint16_avx2
  1147                                          # -- End function
  1148  	.globl	transpose_uint64_uint16_avx2    # -- Begin function transpose_uint64_uint16_avx2
  1149  	.p2align	4, 0x90
  1150  	.type	transpose_uint64_uint16_avx2,@function
  1151  transpose_uint64_uint16_avx2:           # @transpose_uint64_uint16_avx2
  1152  # %bb.0:
  1153  	push	rbp
  1154  	mov	rbp, rsp
  1155  	and	rsp, -8
  1156  	cmp	edx, 4
  1157  	jl	.LBB22_1
  1158  	.p2align	4, 0x90
  1159  .LBB22_5:                               # =>This Inner Loop Header: Depth=1
  1160  	mov	eax, edx
  1161  	mov	rdx, qword ptr [rdi]
  1162  	movzx	edx, word ptr [rcx + 4*rdx]
  1163  	mov	word ptr [rsi], dx
  1164  	mov	rdx, qword ptr [rdi + 8]
  1165  	movzx	edx, word ptr [rcx + 4*rdx]
  1166  	mov	word ptr [rsi + 2], dx
  1167  	mov	rdx, qword ptr [rdi + 16]
  1168  	movzx	edx, word ptr [rcx + 4*rdx]
  1169  	mov	word ptr [rsi + 4], dx
  1170  	mov	rdx, qword ptr [rdi + 24]
  1171  	movzx	edx, word ptr [rcx + 4*rdx]
  1172  	mov	word ptr [rsi + 6], dx
  1173  	lea	edx, [rax - 4]
  1174  	add	rdi, 32
  1175  	add	rsi, 8
  1176  	cmp	eax, 7
  1177  	jg	.LBB22_5
  1178  .LBB22_1:
  1179  	test	edx, edx
  1180  	jle	.LBB22_4
  1181  # %bb.2:
  1182  	add	edx, 1
  1183  	xor	r8d, r8d
  1184  	.p2align	4, 0x90
  1185  .LBB22_3:                               # =>This Inner Loop Header: Depth=1
  1186  	mov	rax, qword ptr [rdi + 4*r8]
  1187  	movzx	eax, word ptr [rcx + 4*rax]
  1188  	mov	word ptr [rsi + r8], ax
  1189  	add	r8, 2
  1190  	add	edx, -1
  1191  	cmp	edx, 1
  1192  	jg	.LBB22_3
  1193  .LBB22_4:
  1194  	mov	rsp, rbp
  1195  	pop	rbp
  1196  	ret
  1197  .Lfunc_end22:
  1198  	.size	transpose_uint64_uint16_avx2, .Lfunc_end22-transpose_uint64_uint16_avx2
  1199                                          # -- End function
  1200  	.globl	transpose_int64_uint16_avx2     # -- Begin function transpose_int64_uint16_avx2
  1201  	.p2align	4, 0x90
  1202  	.type	transpose_int64_uint16_avx2,@function
  1203  transpose_int64_uint16_avx2:            # @transpose_int64_uint16_avx2
  1204  # %bb.0:
  1205  	push	rbp
  1206  	mov	rbp, rsp
  1207  	and	rsp, -8
  1208  	cmp	edx, 4
  1209  	jl	.LBB23_1
  1210  	.p2align	4, 0x90
  1211  .LBB23_5:                               # =>This Inner Loop Header: Depth=1
  1212  	mov	eax, edx
  1213  	mov	rdx, qword ptr [rdi]
  1214  	movzx	edx, word ptr [rcx + 4*rdx]
  1215  	mov	word ptr [rsi], dx
  1216  	mov	rdx, qword ptr [rdi + 8]
  1217  	movzx	edx, word ptr [rcx + 4*rdx]
  1218  	mov	word ptr [rsi + 2], dx
  1219  	mov	rdx, qword ptr [rdi + 16]
  1220  	movzx	edx, word ptr [rcx + 4*rdx]
  1221  	mov	word ptr [rsi + 4], dx
  1222  	mov	rdx, qword ptr [rdi + 24]
  1223  	movzx	edx, word ptr [rcx + 4*rdx]
  1224  	mov	word ptr [rsi + 6], dx
  1225  	lea	edx, [rax - 4]
  1226  	add	rdi, 32
  1227  	add	rsi, 8
  1228  	cmp	eax, 7
  1229  	jg	.LBB23_5
  1230  .LBB23_1:
  1231  	test	edx, edx
  1232  	jle	.LBB23_4
  1233  # %bb.2:
  1234  	add	edx, 1
  1235  	xor	r8d, r8d
  1236  	.p2align	4, 0x90
  1237  .LBB23_3:                               # =>This Inner Loop Header: Depth=1
  1238  	mov	rax, qword ptr [rdi + 4*r8]
  1239  	movzx	eax, word ptr [rcx + 4*rax]
  1240  	mov	word ptr [rsi + r8], ax
  1241  	add	r8, 2
  1242  	add	edx, -1
  1243  	cmp	edx, 1
  1244  	jg	.LBB23_3
  1245  .LBB23_4:
  1246  	mov	rsp, rbp
  1247  	pop	rbp
  1248  	ret
  1249  .Lfunc_end23:
  1250  	.size	transpose_int64_uint16_avx2, .Lfunc_end23-transpose_int64_uint16_avx2
  1251                                          # -- End function
  1252  	.globl	transpose_uint8_int16_avx2      # -- Begin function transpose_uint8_int16_avx2
  1253  	.p2align	4, 0x90
  1254  	.type	transpose_uint8_int16_avx2,@function
  1255  transpose_uint8_int16_avx2:             # @transpose_uint8_int16_avx2
  1256  # %bb.0:
  1257  	push	rbp
  1258  	mov	rbp, rsp
  1259  	and	rsp, -8
  1260  	cmp	edx, 4
  1261  	jl	.LBB24_1
  1262  	.p2align	4, 0x90
  1263  .LBB24_5:                               # =>This Inner Loop Header: Depth=1
  1264  	mov	eax, edx
  1265  	movzx	edx, byte ptr [rdi]
  1266  	movzx	edx, word ptr [rcx + 4*rdx]
  1267  	mov	word ptr [rsi], dx
  1268  	movzx	edx, byte ptr [rdi + 1]
  1269  	movzx	edx, word ptr [rcx + 4*rdx]
  1270  	mov	word ptr [rsi + 2], dx
  1271  	movzx	edx, byte ptr [rdi + 2]
  1272  	movzx	edx, word ptr [rcx + 4*rdx]
  1273  	mov	word ptr [rsi + 4], dx
  1274  	movzx	edx, byte ptr [rdi + 3]
  1275  	movzx	edx, word ptr [rcx + 4*rdx]
  1276  	mov	word ptr [rsi + 6], dx
  1277  	lea	edx, [rax - 4]
  1278  	add	rdi, 4
  1279  	add	rsi, 8
  1280  	cmp	eax, 7
  1281  	jg	.LBB24_5
  1282  .LBB24_1:
  1283  	test	edx, edx
  1284  	jle	.LBB24_4
  1285  # %bb.2:
  1286  	add	edx, 1
  1287  	xor	r8d, r8d
  1288  	.p2align	4, 0x90
  1289  .LBB24_3:                               # =>This Inner Loop Header: Depth=1
  1290  	movzx	eax, byte ptr [rdi + r8]
  1291  	movzx	eax, word ptr [rcx + 4*rax]
  1292  	mov	word ptr [rsi + 2*r8], ax
  1293  	add	r8, 1
  1294  	add	edx, -1
  1295  	cmp	edx, 1
  1296  	jg	.LBB24_3
  1297  .LBB24_4:
  1298  	mov	rsp, rbp
  1299  	pop	rbp
  1300  	ret
  1301  .Lfunc_end24:
  1302  	.size	transpose_uint8_int16_avx2, .Lfunc_end24-transpose_uint8_int16_avx2
  1303                                          # -- End function
  1304  	.globl	transpose_int8_int16_avx2       # -- Begin function transpose_int8_int16_avx2
  1305  	.p2align	4, 0x90
  1306  	.type	transpose_int8_int16_avx2,@function
  1307  transpose_int8_int16_avx2:              # @transpose_int8_int16_avx2
  1308  # %bb.0:
  1309  	push	rbp
  1310  	mov	rbp, rsp
  1311  	and	rsp, -8
  1312  	cmp	edx, 4
  1313  	jl	.LBB25_1
  1314  	.p2align	4, 0x90
  1315  .LBB25_5:                               # =>This Inner Loop Header: Depth=1
  1316  	mov	eax, edx
  1317  	movsx	rdx, byte ptr [rdi]
  1318  	movzx	edx, word ptr [rcx + 4*rdx]
  1319  	mov	word ptr [rsi], dx
  1320  	movsx	rdx, byte ptr [rdi + 1]
  1321  	movzx	edx, word ptr [rcx + 4*rdx]
  1322  	mov	word ptr [rsi + 2], dx
  1323  	movsx	rdx, byte ptr [rdi + 2]
  1324  	movzx	edx, word ptr [rcx + 4*rdx]
  1325  	mov	word ptr [rsi + 4], dx
  1326  	movsx	rdx, byte ptr [rdi + 3]
  1327  	movzx	edx, word ptr [rcx + 4*rdx]
  1328  	mov	word ptr [rsi + 6], dx
  1329  	lea	edx, [rax - 4]
  1330  	add	rdi, 4
  1331  	add	rsi, 8
  1332  	cmp	eax, 7
  1333  	jg	.LBB25_5
  1334  .LBB25_1:
  1335  	test	edx, edx
  1336  	jle	.LBB25_4
  1337  # %bb.2:
  1338  	add	edx, 1
  1339  	xor	r8d, r8d
  1340  	.p2align	4, 0x90
  1341  .LBB25_3:                               # =>This Inner Loop Header: Depth=1
  1342  	movsx	rax, byte ptr [rdi + r8]
  1343  	movzx	eax, word ptr [rcx + 4*rax]
  1344  	mov	word ptr [rsi + 2*r8], ax
  1345  	add	r8, 1
  1346  	add	edx, -1
  1347  	cmp	edx, 1
  1348  	jg	.LBB25_3
  1349  .LBB25_4:
  1350  	mov	rsp, rbp
  1351  	pop	rbp
  1352  	ret
  1353  .Lfunc_end25:
  1354  	.size	transpose_int8_int16_avx2, .Lfunc_end25-transpose_int8_int16_avx2
  1355                                          # -- End function
  1356  	.globl	transpose_uint16_int16_avx2     # -- Begin function transpose_uint16_int16_avx2
  1357  	.p2align	4, 0x90
  1358  	.type	transpose_uint16_int16_avx2,@function
  1359  transpose_uint16_int16_avx2:            # @transpose_uint16_int16_avx2
  1360  # %bb.0:
  1361  	push	rbp
  1362  	mov	rbp, rsp
  1363  	and	rsp, -8
  1364  	cmp	edx, 4
  1365  	jl	.LBB26_1
  1366  	.p2align	4, 0x90
  1367  .LBB26_5:                               # =>This Inner Loop Header: Depth=1
  1368  	mov	eax, edx
  1369  	movzx	edx, word ptr [rdi]
  1370  	movzx	edx, word ptr [rcx + 4*rdx]
  1371  	mov	word ptr [rsi], dx
  1372  	movzx	edx, word ptr [rdi + 2]
  1373  	movzx	edx, word ptr [rcx + 4*rdx]
  1374  	mov	word ptr [rsi + 2], dx
  1375  	movzx	edx, word ptr [rdi + 4]
  1376  	movzx	edx, word ptr [rcx + 4*rdx]
  1377  	mov	word ptr [rsi + 4], dx
  1378  	movzx	edx, word ptr [rdi + 6]
  1379  	movzx	edx, word ptr [rcx + 4*rdx]
  1380  	mov	word ptr [rsi + 6], dx
  1381  	lea	edx, [rax - 4]
  1382  	add	rdi, 8
  1383  	add	rsi, 8
  1384  	cmp	eax, 7
  1385  	jg	.LBB26_5
  1386  .LBB26_1:
  1387  	test	edx, edx
  1388  	jle	.LBB26_4
  1389  # %bb.2:
  1390  	add	edx, 1
  1391  	xor	r8d, r8d
  1392  	.p2align	4, 0x90
  1393  .LBB26_3:                               # =>This Inner Loop Header: Depth=1
  1394  	movzx	eax, word ptr [rdi + r8]
  1395  	movzx	eax, word ptr [rcx + 4*rax]
  1396  	mov	word ptr [rsi + r8], ax
  1397  	add	r8, 2
  1398  	add	edx, -1
  1399  	cmp	edx, 1
  1400  	jg	.LBB26_3
  1401  .LBB26_4:
  1402  	mov	rsp, rbp
  1403  	pop	rbp
  1404  	ret
  1405  .Lfunc_end26:
  1406  	.size	transpose_uint16_int16_avx2, .Lfunc_end26-transpose_uint16_int16_avx2
  1407                                          # -- End function
  1408  	.globl	transpose_int16_int16_avx2      # -- Begin function transpose_int16_int16_avx2
  1409  	.p2align	4, 0x90
  1410  	.type	transpose_int16_int16_avx2,@function
  1411  transpose_int16_int16_avx2:             # @transpose_int16_int16_avx2
  1412  # %bb.0:
  1413  	push	rbp
  1414  	mov	rbp, rsp
  1415  	and	rsp, -8
  1416  	cmp	edx, 4
  1417  	jl	.LBB27_1
  1418  	.p2align	4, 0x90
  1419  .LBB27_5:                               # =>This Inner Loop Header: Depth=1
  1420  	mov	eax, edx
  1421  	movsx	rdx, word ptr [rdi]
  1422  	movzx	edx, word ptr [rcx + 4*rdx]
  1423  	mov	word ptr [rsi], dx
  1424  	movsx	rdx, word ptr [rdi + 2]
  1425  	movzx	edx, word ptr [rcx + 4*rdx]
  1426  	mov	word ptr [rsi + 2], dx
  1427  	movsx	rdx, word ptr [rdi + 4]
  1428  	movzx	edx, word ptr [rcx + 4*rdx]
  1429  	mov	word ptr [rsi + 4], dx
  1430  	movsx	rdx, word ptr [rdi + 6]
  1431  	movzx	edx, word ptr [rcx + 4*rdx]
  1432  	mov	word ptr [rsi + 6], dx
  1433  	lea	edx, [rax - 4]
  1434  	add	rdi, 8
  1435  	add	rsi, 8
  1436  	cmp	eax, 7
  1437  	jg	.LBB27_5
  1438  .LBB27_1:
  1439  	test	edx, edx
  1440  	jle	.LBB27_4
  1441  # %bb.2:
  1442  	add	edx, 1
  1443  	xor	r8d, r8d
  1444  	.p2align	4, 0x90
  1445  .LBB27_3:                               # =>This Inner Loop Header: Depth=1
  1446  	movsx	rax, word ptr [rdi + r8]
  1447  	movzx	eax, word ptr [rcx + 4*rax]
  1448  	mov	word ptr [rsi + r8], ax
  1449  	add	r8, 2
  1450  	add	edx, -1
  1451  	cmp	edx, 1
  1452  	jg	.LBB27_3
  1453  .LBB27_4:
  1454  	mov	rsp, rbp
  1455  	pop	rbp
  1456  	ret
  1457  .Lfunc_end27:
  1458  	.size	transpose_int16_int16_avx2, .Lfunc_end27-transpose_int16_int16_avx2
  1459                                          # -- End function
  1460  	.globl	transpose_uint32_int16_avx2     # -- Begin function transpose_uint32_int16_avx2
  1461  	.p2align	4, 0x90
  1462  	.type	transpose_uint32_int16_avx2,@function
  1463  transpose_uint32_int16_avx2:            # @transpose_uint32_int16_avx2
  1464  # %bb.0:
  1465  	push	rbp
  1466  	mov	rbp, rsp
  1467  	and	rsp, -8
  1468  	cmp	edx, 4
  1469  	jl	.LBB28_1
  1470  	.p2align	4, 0x90
  1471  .LBB28_5:                               # =>This Inner Loop Header: Depth=1
  1472  	mov	eax, edx
  1473  	mov	edx, dword ptr [rdi]
  1474  	movzx	edx, word ptr [rcx + 4*rdx]
  1475  	mov	word ptr [rsi], dx
  1476  	mov	edx, dword ptr [rdi + 4]
  1477  	movzx	edx, word ptr [rcx + 4*rdx]
  1478  	mov	word ptr [rsi + 2], dx
  1479  	mov	edx, dword ptr [rdi + 8]
  1480  	movzx	edx, word ptr [rcx + 4*rdx]
  1481  	mov	word ptr [rsi + 4], dx
  1482  	mov	edx, dword ptr [rdi + 12]
  1483  	movzx	edx, word ptr [rcx + 4*rdx]
  1484  	mov	word ptr [rsi + 6], dx
  1485  	lea	edx, [rax - 4]
  1486  	add	rdi, 16
  1487  	add	rsi, 8
  1488  	cmp	eax, 7
  1489  	jg	.LBB28_5
  1490  .LBB28_1:
  1491  	test	edx, edx
  1492  	jle	.LBB28_4
  1493  # %bb.2:
  1494  	add	edx, 1
  1495  	xor	r8d, r8d
  1496  	.p2align	4, 0x90
  1497  .LBB28_3:                               # =>This Inner Loop Header: Depth=1
  1498  	mov	eax, dword ptr [rdi + 2*r8]
  1499  	movzx	eax, word ptr [rcx + 4*rax]
  1500  	mov	word ptr [rsi + r8], ax
  1501  	add	r8, 2
  1502  	add	edx, -1
  1503  	cmp	edx, 1
  1504  	jg	.LBB28_3
  1505  .LBB28_4:
  1506  	mov	rsp, rbp
  1507  	pop	rbp
  1508  	ret
  1509  .Lfunc_end28:
  1510  	.size	transpose_uint32_int16_avx2, .Lfunc_end28-transpose_uint32_int16_avx2
  1511                                          # -- End function
  1512  	.globl	transpose_int32_int16_avx2      # -- Begin function transpose_int32_int16_avx2
  1513  	.p2align	4, 0x90
  1514  	.type	transpose_int32_int16_avx2,@function
  1515  transpose_int32_int16_avx2:             # @transpose_int32_int16_avx2
  1516  # %bb.0:
  1517  	push	rbp
  1518  	mov	rbp, rsp
  1519  	and	rsp, -8
  1520  	cmp	edx, 4
  1521  	jl	.LBB29_1
  1522  	.p2align	4, 0x90
  1523  .LBB29_5:                               # =>This Inner Loop Header: Depth=1
  1524  	mov	eax, edx
  1525  	movsxd	rdx, dword ptr [rdi]
  1526  	movzx	edx, word ptr [rcx + 4*rdx]
  1527  	mov	word ptr [rsi], dx
  1528  	movsxd	rdx, dword ptr [rdi + 4]
  1529  	movzx	edx, word ptr [rcx + 4*rdx]
  1530  	mov	word ptr [rsi + 2], dx
  1531  	movsxd	rdx, dword ptr [rdi + 8]
  1532  	movzx	edx, word ptr [rcx + 4*rdx]
  1533  	mov	word ptr [rsi + 4], dx
  1534  	movsxd	rdx, dword ptr [rdi + 12]
  1535  	movzx	edx, word ptr [rcx + 4*rdx]
  1536  	mov	word ptr [rsi + 6], dx
  1537  	lea	edx, [rax - 4]
  1538  	add	rdi, 16
  1539  	add	rsi, 8
  1540  	cmp	eax, 7
  1541  	jg	.LBB29_5
  1542  .LBB29_1:
  1543  	test	edx, edx
  1544  	jle	.LBB29_4
  1545  # %bb.2:
  1546  	add	edx, 1
  1547  	xor	r8d, r8d
  1548  	.p2align	4, 0x90
  1549  .LBB29_3:                               # =>This Inner Loop Header: Depth=1
  1550  	movsxd	rax, dword ptr [rdi + 2*r8]
  1551  	movzx	eax, word ptr [rcx + 4*rax]
  1552  	mov	word ptr [rsi + r8], ax
  1553  	add	r8, 2
  1554  	add	edx, -1
  1555  	cmp	edx, 1
  1556  	jg	.LBB29_3
  1557  .LBB29_4:
  1558  	mov	rsp, rbp
  1559  	pop	rbp
  1560  	ret
  1561  .Lfunc_end29:
  1562  	.size	transpose_int32_int16_avx2, .Lfunc_end29-transpose_int32_int16_avx2
  1563                                          # -- End function
  1564  	.globl	transpose_uint64_int16_avx2     # -- Begin function transpose_uint64_int16_avx2
  1565  	.p2align	4, 0x90
  1566  	.type	transpose_uint64_int16_avx2,@function
  1567  transpose_uint64_int16_avx2:            # @transpose_uint64_int16_avx2
  1568  # %bb.0:
  1569  	push	rbp
  1570  	mov	rbp, rsp
  1571  	and	rsp, -8
  1572  	cmp	edx, 4
  1573  	jl	.LBB30_1
  1574  	.p2align	4, 0x90
  1575  .LBB30_5:                               # =>This Inner Loop Header: Depth=1
  1576  	mov	eax, edx
  1577  	mov	rdx, qword ptr [rdi]
  1578  	movzx	edx, word ptr [rcx + 4*rdx]
  1579  	mov	word ptr [rsi], dx
  1580  	mov	rdx, qword ptr [rdi + 8]
  1581  	movzx	edx, word ptr [rcx + 4*rdx]
  1582  	mov	word ptr [rsi + 2], dx
  1583  	mov	rdx, qword ptr [rdi + 16]
  1584  	movzx	edx, word ptr [rcx + 4*rdx]
  1585  	mov	word ptr [rsi + 4], dx
  1586  	mov	rdx, qword ptr [rdi + 24]
  1587  	movzx	edx, word ptr [rcx + 4*rdx]
  1588  	mov	word ptr [rsi + 6], dx
  1589  	lea	edx, [rax - 4]
  1590  	add	rdi, 32
  1591  	add	rsi, 8
  1592  	cmp	eax, 7
  1593  	jg	.LBB30_5
  1594  .LBB30_1:
  1595  	test	edx, edx
  1596  	jle	.LBB30_4
  1597  # %bb.2:
  1598  	add	edx, 1
  1599  	xor	r8d, r8d
  1600  	.p2align	4, 0x90
  1601  .LBB30_3:                               # =>This Inner Loop Header: Depth=1
  1602  	mov	rax, qword ptr [rdi + 4*r8]
  1603  	movzx	eax, word ptr [rcx + 4*rax]
  1604  	mov	word ptr [rsi + r8], ax
  1605  	add	r8, 2
  1606  	add	edx, -1
  1607  	cmp	edx, 1
  1608  	jg	.LBB30_3
  1609  .LBB30_4:
  1610  	mov	rsp, rbp
  1611  	pop	rbp
  1612  	ret
  1613  .Lfunc_end30:
  1614  	.size	transpose_uint64_int16_avx2, .Lfunc_end30-transpose_uint64_int16_avx2
  1615                                          # -- End function
  1616  	.globl	transpose_int64_int16_avx2      # -- Begin function transpose_int64_int16_avx2
  1617  	.p2align	4, 0x90
  1618  	.type	transpose_int64_int16_avx2,@function
  1619  transpose_int64_int16_avx2:             # @transpose_int64_int16_avx2
  1620  # %bb.0:
  1621  	push	rbp
  1622  	mov	rbp, rsp
  1623  	and	rsp, -8
  1624  	cmp	edx, 4
  1625  	jl	.LBB31_1
  1626  	.p2align	4, 0x90
  1627  .LBB31_5:                               # =>This Inner Loop Header: Depth=1
  1628  	mov	eax, edx
  1629  	mov	rdx, qword ptr [rdi]
  1630  	movzx	edx, word ptr [rcx + 4*rdx]
  1631  	mov	word ptr [rsi], dx
  1632  	mov	rdx, qword ptr [rdi + 8]
  1633  	movzx	edx, word ptr [rcx + 4*rdx]
  1634  	mov	word ptr [rsi + 2], dx
  1635  	mov	rdx, qword ptr [rdi + 16]
  1636  	movzx	edx, word ptr [rcx + 4*rdx]
  1637  	mov	word ptr [rsi + 4], dx
  1638  	mov	rdx, qword ptr [rdi + 24]
  1639  	movzx	edx, word ptr [rcx + 4*rdx]
  1640  	mov	word ptr [rsi + 6], dx
  1641  	lea	edx, [rax - 4]
  1642  	add	rdi, 32
  1643  	add	rsi, 8
  1644  	cmp	eax, 7
  1645  	jg	.LBB31_5
  1646  .LBB31_1:
  1647  	test	edx, edx
  1648  	jle	.LBB31_4
  1649  # %bb.2:
  1650  	add	edx, 1
  1651  	xor	r8d, r8d
  1652  	.p2align	4, 0x90
  1653  .LBB31_3:                               # =>This Inner Loop Header: Depth=1
  1654  	mov	rax, qword ptr [rdi + 4*r8]
  1655  	movzx	eax, word ptr [rcx + 4*rax]
  1656  	mov	word ptr [rsi + r8], ax
  1657  	add	r8, 2
  1658  	add	edx, -1
  1659  	cmp	edx, 1
  1660  	jg	.LBB31_3
  1661  .LBB31_4:
  1662  	mov	rsp, rbp
  1663  	pop	rbp
  1664  	ret
  1665  .Lfunc_end31:
  1666  	.size	transpose_int64_int16_avx2, .Lfunc_end31-transpose_int64_int16_avx2
  1667                                          # -- End function
  1668  	.globl	transpose_uint8_uint32_avx2     # -- Begin function transpose_uint8_uint32_avx2
  1669  	.p2align	4, 0x90
  1670  	.type	transpose_uint8_uint32_avx2,@function
  1671  transpose_uint8_uint32_avx2:            # @transpose_uint8_uint32_avx2
  1672  # %bb.0:
  1673  	push	rbp
  1674  	mov	rbp, rsp
  1675  	and	rsp, -8
  1676  	cmp	edx, 4
  1677  	jl	.LBB32_1
  1678  	.p2align	4, 0x90
  1679  .LBB32_5:                               # =>This Inner Loop Header: Depth=1
  1680  	mov	eax, edx
  1681  	movzx	edx, byte ptr [rdi]
  1682  	mov	edx, dword ptr [rcx + 4*rdx]
  1683  	mov	dword ptr [rsi], edx
  1684  	movzx	edx, byte ptr [rdi + 1]
  1685  	mov	edx, dword ptr [rcx + 4*rdx]
  1686  	mov	dword ptr [rsi + 4], edx
  1687  	movzx	edx, byte ptr [rdi + 2]
  1688  	mov	edx, dword ptr [rcx + 4*rdx]
  1689  	mov	dword ptr [rsi + 8], edx
  1690  	movzx	edx, byte ptr [rdi + 3]
  1691  	mov	edx, dword ptr [rcx + 4*rdx]
  1692  	mov	dword ptr [rsi + 12], edx
  1693  	lea	edx, [rax - 4]
  1694  	add	rdi, 4
  1695  	add	rsi, 16
  1696  	cmp	eax, 7
  1697  	jg	.LBB32_5
  1698  .LBB32_1:
  1699  	test	edx, edx
  1700  	jle	.LBB32_4
  1701  # %bb.2:
  1702  	add	edx, 1
  1703  	xor	r8d, r8d
  1704  	.p2align	4, 0x90
  1705  .LBB32_3:                               # =>This Inner Loop Header: Depth=1
  1706  	movzx	eax, byte ptr [rdi + r8]
  1707  	mov	eax, dword ptr [rcx + 4*rax]
  1708  	mov	dword ptr [rsi + 4*r8], eax
  1709  	add	r8, 1
  1710  	add	edx, -1
  1711  	cmp	edx, 1
  1712  	jg	.LBB32_3
  1713  .LBB32_4:
  1714  	mov	rsp, rbp
  1715  	pop	rbp
  1716  	ret
  1717  .Lfunc_end32:
  1718  	.size	transpose_uint8_uint32_avx2, .Lfunc_end32-transpose_uint8_uint32_avx2
  1719                                          # -- End function
  1720  	.globl	transpose_int8_uint32_avx2      # -- Begin function transpose_int8_uint32_avx2
  1721  	.p2align	4, 0x90
  1722  	.type	transpose_int8_uint32_avx2,@function
  1723  transpose_int8_uint32_avx2:             # @transpose_int8_uint32_avx2
  1724  # %bb.0:
  1725  	push	rbp
  1726  	mov	rbp, rsp
  1727  	and	rsp, -8
  1728  	cmp	edx, 4
  1729  	jl	.LBB33_1
  1730  	.p2align	4, 0x90
  1731  .LBB33_5:                               # =>This Inner Loop Header: Depth=1
  1732  	mov	eax, edx
  1733  	movsx	rdx, byte ptr [rdi]
  1734  	mov	edx, dword ptr [rcx + 4*rdx]
  1735  	mov	dword ptr [rsi], edx
  1736  	movsx	rdx, byte ptr [rdi + 1]
  1737  	mov	edx, dword ptr [rcx + 4*rdx]
  1738  	mov	dword ptr [rsi + 4], edx
  1739  	movsx	rdx, byte ptr [rdi + 2]
  1740  	mov	edx, dword ptr [rcx + 4*rdx]
  1741  	mov	dword ptr [rsi + 8], edx
  1742  	movsx	rdx, byte ptr [rdi + 3]
  1743  	mov	edx, dword ptr [rcx + 4*rdx]
  1744  	mov	dword ptr [rsi + 12], edx
  1745  	lea	edx, [rax - 4]
  1746  	add	rdi, 4
  1747  	add	rsi, 16
  1748  	cmp	eax, 7
  1749  	jg	.LBB33_5
  1750  .LBB33_1:
  1751  	test	edx, edx
  1752  	jle	.LBB33_4
  1753  # %bb.2:
  1754  	add	edx, 1
  1755  	xor	r8d, r8d
  1756  	.p2align	4, 0x90
  1757  .LBB33_3:                               # =>This Inner Loop Header: Depth=1
  1758  	movsx	rax, byte ptr [rdi + r8]
  1759  	mov	eax, dword ptr [rcx + 4*rax]
  1760  	mov	dword ptr [rsi + 4*r8], eax
  1761  	add	r8, 1
  1762  	add	edx, -1
  1763  	cmp	edx, 1
  1764  	jg	.LBB33_3
  1765  .LBB33_4:
  1766  	mov	rsp, rbp
  1767  	pop	rbp
  1768  	ret
  1769  .Lfunc_end33:
  1770  	.size	transpose_int8_uint32_avx2, .Lfunc_end33-transpose_int8_uint32_avx2
  1771                                          # -- End function
  1772  	.globl	transpose_uint16_uint32_avx2    # -- Begin function transpose_uint16_uint32_avx2
  1773  	.p2align	4, 0x90
  1774  	.type	transpose_uint16_uint32_avx2,@function
  1775  transpose_uint16_uint32_avx2:           # @transpose_uint16_uint32_avx2
  1776  # %bb.0:
  1777  	push	rbp
  1778  	mov	rbp, rsp
  1779  	and	rsp, -8
  1780  	cmp	edx, 4
  1781  	jl	.LBB34_1
  1782  	.p2align	4, 0x90
  1783  .LBB34_5:                               # =>This Inner Loop Header: Depth=1
  1784  	mov	eax, edx
  1785  	movzx	edx, word ptr [rdi]
  1786  	mov	edx, dword ptr [rcx + 4*rdx]
  1787  	mov	dword ptr [rsi], edx
  1788  	movzx	edx, word ptr [rdi + 2]
  1789  	mov	edx, dword ptr [rcx + 4*rdx]
  1790  	mov	dword ptr [rsi + 4], edx
  1791  	movzx	edx, word ptr [rdi + 4]
  1792  	mov	edx, dword ptr [rcx + 4*rdx]
  1793  	mov	dword ptr [rsi + 8], edx
  1794  	movzx	edx, word ptr [rdi + 6]
  1795  	mov	edx, dword ptr [rcx + 4*rdx]
  1796  	mov	dword ptr [rsi + 12], edx
  1797  	lea	edx, [rax - 4]
  1798  	add	rdi, 8
  1799  	add	rsi, 16
  1800  	cmp	eax, 7
  1801  	jg	.LBB34_5
  1802  .LBB34_1:
  1803  	test	edx, edx
  1804  	jle	.LBB34_4
  1805  # %bb.2:
  1806  	add	edx, 1
  1807  	xor	r8d, r8d
  1808  	.p2align	4, 0x90
  1809  .LBB34_3:                               # =>This Inner Loop Header: Depth=1
  1810  	movzx	eax, word ptr [rdi + r8]
  1811  	mov	eax, dword ptr [rcx + 4*rax]
  1812  	mov	dword ptr [rsi + 2*r8], eax
  1813  	add	r8, 2
  1814  	add	edx, -1
  1815  	cmp	edx, 1
  1816  	jg	.LBB34_3
  1817  .LBB34_4:
  1818  	mov	rsp, rbp
  1819  	pop	rbp
  1820  	ret
  1821  .Lfunc_end34:
  1822  	.size	transpose_uint16_uint32_avx2, .Lfunc_end34-transpose_uint16_uint32_avx2
  1823                                          # -- End function
  1824  	.globl	transpose_int16_uint32_avx2     # -- Begin function transpose_int16_uint32_avx2
  1825  	.p2align	4, 0x90
  1826  	.type	transpose_int16_uint32_avx2,@function
  1827  transpose_int16_uint32_avx2:            # @transpose_int16_uint32_avx2
  1828  # %bb.0:
  1829  	push	rbp
  1830  	mov	rbp, rsp
  1831  	and	rsp, -8
  1832  	cmp	edx, 4
  1833  	jl	.LBB35_1
  1834  	.p2align	4, 0x90
  1835  .LBB35_5:                               # =>This Inner Loop Header: Depth=1
  1836  	mov	eax, edx
  1837  	movsx	rdx, word ptr [rdi]
  1838  	mov	edx, dword ptr [rcx + 4*rdx]
  1839  	mov	dword ptr [rsi], edx
  1840  	movsx	rdx, word ptr [rdi + 2]
  1841  	mov	edx, dword ptr [rcx + 4*rdx]
  1842  	mov	dword ptr [rsi + 4], edx
  1843  	movsx	rdx, word ptr [rdi + 4]
  1844  	mov	edx, dword ptr [rcx + 4*rdx]
  1845  	mov	dword ptr [rsi + 8], edx
  1846  	movsx	rdx, word ptr [rdi + 6]
  1847  	mov	edx, dword ptr [rcx + 4*rdx]
  1848  	mov	dword ptr [rsi + 12], edx
  1849  	lea	edx, [rax - 4]
  1850  	add	rdi, 8
  1851  	add	rsi, 16
  1852  	cmp	eax, 7
  1853  	jg	.LBB35_5
  1854  .LBB35_1:
  1855  	test	edx, edx
  1856  	jle	.LBB35_4
  1857  # %bb.2:
  1858  	add	edx, 1
  1859  	xor	r8d, r8d
  1860  	.p2align	4, 0x90
  1861  .LBB35_3:                               # =>This Inner Loop Header: Depth=1
  1862  	movsx	rax, word ptr [rdi + r8]
  1863  	mov	eax, dword ptr [rcx + 4*rax]
  1864  	mov	dword ptr [rsi + 2*r8], eax
  1865  	add	r8, 2
  1866  	add	edx, -1
  1867  	cmp	edx, 1
  1868  	jg	.LBB35_3
  1869  .LBB35_4:
  1870  	mov	rsp, rbp
  1871  	pop	rbp
  1872  	ret
  1873  .Lfunc_end35:
  1874  	.size	transpose_int16_uint32_avx2, .Lfunc_end35-transpose_int16_uint32_avx2
  1875                                          # -- End function
  1876  	.globl	transpose_uint32_uint32_avx2    # -- Begin function transpose_uint32_uint32_avx2
  1877  	.p2align	4, 0x90
  1878  	.type	transpose_uint32_uint32_avx2,@function
  1879  transpose_uint32_uint32_avx2:           # @transpose_uint32_uint32_avx2
  1880  # %bb.0:
  1881  	push	rbp
  1882  	mov	rbp, rsp
  1883  	and	rsp, -8
  1884  	cmp	edx, 4
  1885  	jl	.LBB36_1
  1886  	.p2align	4, 0x90
  1887  .LBB36_5:                               # =>This Inner Loop Header: Depth=1
  1888  	mov	eax, edx
  1889  	mov	edx, dword ptr [rdi]
  1890  	mov	edx, dword ptr [rcx + 4*rdx]
  1891  	mov	dword ptr [rsi], edx
  1892  	mov	edx, dword ptr [rdi + 4]
  1893  	mov	edx, dword ptr [rcx + 4*rdx]
  1894  	mov	dword ptr [rsi + 4], edx
  1895  	mov	edx, dword ptr [rdi + 8]
  1896  	mov	edx, dword ptr [rcx + 4*rdx]
  1897  	mov	dword ptr [rsi + 8], edx
  1898  	mov	edx, dword ptr [rdi + 12]
  1899  	mov	edx, dword ptr [rcx + 4*rdx]
  1900  	mov	dword ptr [rsi + 12], edx
  1901  	lea	edx, [rax - 4]
  1902  	add	rdi, 16
  1903  	add	rsi, 16
  1904  	cmp	eax, 7
  1905  	jg	.LBB36_5
  1906  .LBB36_1:
  1907  	test	edx, edx
  1908  	jle	.LBB36_4
  1909  # %bb.2:
  1910  	add	edx, 1
  1911  	xor	r8d, r8d
  1912  	.p2align	4, 0x90
  1913  .LBB36_3:                               # =>This Inner Loop Header: Depth=1
  1914  	mov	eax, dword ptr [rdi + r8]
  1915  	mov	eax, dword ptr [rcx + 4*rax]
  1916  	mov	dword ptr [rsi + r8], eax
  1917  	add	r8, 4
  1918  	add	edx, -1
  1919  	cmp	edx, 1
  1920  	jg	.LBB36_3
  1921  .LBB36_4:
  1922  	mov	rsp, rbp
  1923  	pop	rbp
  1924  	ret
  1925  .Lfunc_end36:
  1926  	.size	transpose_uint32_uint32_avx2, .Lfunc_end36-transpose_uint32_uint32_avx2
  1927                                          # -- End function
  1928  	.globl	transpose_int32_uint32_avx2     # -- Begin function transpose_int32_uint32_avx2
  1929  	.p2align	4, 0x90
  1930  	.type	transpose_int32_uint32_avx2,@function
  1931  transpose_int32_uint32_avx2:            # @transpose_int32_uint32_avx2
  1932  # %bb.0:
  1933  	push	rbp
  1934  	mov	rbp, rsp
  1935  	and	rsp, -8
  1936  	cmp	edx, 4
  1937  	jl	.LBB37_1
  1938  	.p2align	4, 0x90
  1939  .LBB37_5:                               # =>This Inner Loop Header: Depth=1
  1940  	mov	eax, edx
  1941  	movsxd	rdx, dword ptr [rdi]
  1942  	mov	edx, dword ptr [rcx + 4*rdx]
  1943  	mov	dword ptr [rsi], edx
  1944  	movsxd	rdx, dword ptr [rdi + 4]
  1945  	mov	edx, dword ptr [rcx + 4*rdx]
  1946  	mov	dword ptr [rsi + 4], edx
  1947  	movsxd	rdx, dword ptr [rdi + 8]
  1948  	mov	edx, dword ptr [rcx + 4*rdx]
  1949  	mov	dword ptr [rsi + 8], edx
  1950  	movsxd	rdx, dword ptr [rdi + 12]
  1951  	mov	edx, dword ptr [rcx + 4*rdx]
  1952  	mov	dword ptr [rsi + 12], edx
  1953  	lea	edx, [rax - 4]
  1954  	add	rdi, 16
  1955  	add	rsi, 16
  1956  	cmp	eax, 7
  1957  	jg	.LBB37_5
  1958  .LBB37_1:
  1959  	test	edx, edx
  1960  	jle	.LBB37_4
  1961  # %bb.2:
  1962  	add	edx, 1
  1963  	xor	r8d, r8d
  1964  	.p2align	4, 0x90
  1965  .LBB37_3:                               # =>This Inner Loop Header: Depth=1
  1966  	movsxd	rax, dword ptr [rdi + r8]
  1967  	mov	eax, dword ptr [rcx + 4*rax]
  1968  	mov	dword ptr [rsi + r8], eax
  1969  	add	r8, 4
  1970  	add	edx, -1
  1971  	cmp	edx, 1
  1972  	jg	.LBB37_3
  1973  .LBB37_4:
  1974  	mov	rsp, rbp
  1975  	pop	rbp
  1976  	ret
  1977  .Lfunc_end37:
  1978  	.size	transpose_int32_uint32_avx2, .Lfunc_end37-transpose_int32_uint32_avx2
  1979                                          # -- End function
  1980  	.globl	transpose_uint64_uint32_avx2    # -- Begin function transpose_uint64_uint32_avx2
  1981  	.p2align	4, 0x90
  1982  	.type	transpose_uint64_uint32_avx2,@function
  1983  transpose_uint64_uint32_avx2:           # @transpose_uint64_uint32_avx2
  1984  # %bb.0:
  1985  	push	rbp
  1986  	mov	rbp, rsp
  1987  	and	rsp, -8
  1988  	cmp	edx, 4
  1989  	jl	.LBB38_1
  1990  	.p2align	4, 0x90
  1991  .LBB38_5:                               # =>This Inner Loop Header: Depth=1
  1992  	mov	eax, edx
  1993  	mov	rdx, qword ptr [rdi]
  1994  	mov	edx, dword ptr [rcx + 4*rdx]
  1995  	mov	dword ptr [rsi], edx
  1996  	mov	rdx, qword ptr [rdi + 8]
  1997  	mov	edx, dword ptr [rcx + 4*rdx]
  1998  	mov	dword ptr [rsi + 4], edx
  1999  	mov	rdx, qword ptr [rdi + 16]
  2000  	mov	edx, dword ptr [rcx + 4*rdx]
  2001  	mov	dword ptr [rsi + 8], edx
  2002  	mov	rdx, qword ptr [rdi + 24]
  2003  	mov	edx, dword ptr [rcx + 4*rdx]
  2004  	mov	dword ptr [rsi + 12], edx
  2005  	lea	edx, [rax - 4]
  2006  	add	rdi, 32
  2007  	add	rsi, 16
  2008  	cmp	eax, 7
  2009  	jg	.LBB38_5
  2010  .LBB38_1:
  2011  	test	edx, edx
  2012  	jle	.LBB38_4
  2013  # %bb.2:
  2014  	add	edx, 1
  2015  	xor	r8d, r8d
  2016  	.p2align	4, 0x90
  2017  .LBB38_3:                               # =>This Inner Loop Header: Depth=1
  2018  	mov	rax, qword ptr [rdi + 2*r8]
  2019  	mov	eax, dword ptr [rcx + 4*rax]
  2020  	mov	dword ptr [rsi + r8], eax
  2021  	add	r8, 4
  2022  	add	edx, -1
  2023  	cmp	edx, 1
  2024  	jg	.LBB38_3
  2025  .LBB38_4:
  2026  	mov	rsp, rbp
  2027  	pop	rbp
  2028  	ret
  2029  .Lfunc_end38:
  2030  	.size	transpose_uint64_uint32_avx2, .Lfunc_end38-transpose_uint64_uint32_avx2
  2031                                          # -- End function
  2032  	.globl	transpose_int64_uint32_avx2     # -- Begin function transpose_int64_uint32_avx2
  2033  	.p2align	4, 0x90
  2034  	.type	transpose_int64_uint32_avx2,@function
  2035  transpose_int64_uint32_avx2:            # @transpose_int64_uint32_avx2
  2036  # %bb.0:
  2037  	push	rbp
  2038  	mov	rbp, rsp
  2039  	and	rsp, -8
  2040  	cmp	edx, 4
  2041  	jl	.LBB39_1
  2042  	.p2align	4, 0x90
  2043  .LBB39_5:                               # =>This Inner Loop Header: Depth=1
  2044  	mov	eax, edx
  2045  	mov	rdx, qword ptr [rdi]
  2046  	mov	edx, dword ptr [rcx + 4*rdx]
  2047  	mov	dword ptr [rsi], edx
  2048  	mov	rdx, qword ptr [rdi + 8]
  2049  	mov	edx, dword ptr [rcx + 4*rdx]
  2050  	mov	dword ptr [rsi + 4], edx
  2051  	mov	rdx, qword ptr [rdi + 16]
  2052  	mov	edx, dword ptr [rcx + 4*rdx]
  2053  	mov	dword ptr [rsi + 8], edx
  2054  	mov	rdx, qword ptr [rdi + 24]
  2055  	mov	edx, dword ptr [rcx + 4*rdx]
  2056  	mov	dword ptr [rsi + 12], edx
  2057  	lea	edx, [rax - 4]
  2058  	add	rdi, 32
  2059  	add	rsi, 16
  2060  	cmp	eax, 7
  2061  	jg	.LBB39_5
  2062  .LBB39_1:
  2063  	test	edx, edx
  2064  	jle	.LBB39_4
  2065  # %bb.2:
  2066  	add	edx, 1
  2067  	xor	r8d, r8d
  2068  	.p2align	4, 0x90
  2069  .LBB39_3:                               # =>This Inner Loop Header: Depth=1
  2070  	mov	rax, qword ptr [rdi + 2*r8]
  2071  	mov	eax, dword ptr [rcx + 4*rax]
  2072  	mov	dword ptr [rsi + r8], eax
  2073  	add	r8, 4
  2074  	add	edx, -1
  2075  	cmp	edx, 1
  2076  	jg	.LBB39_3
  2077  .LBB39_4:
  2078  	mov	rsp, rbp
  2079  	pop	rbp
  2080  	ret
  2081  .Lfunc_end39:
  2082  	.size	transpose_int64_uint32_avx2, .Lfunc_end39-transpose_int64_uint32_avx2
  2083                                          # -- End function
  2084  	.globl	transpose_uint8_int32_avx2      # -- Begin function transpose_uint8_int32_avx2
  2085  	.p2align	4, 0x90
  2086  	.type	transpose_uint8_int32_avx2,@function
  2087  transpose_uint8_int32_avx2:             # @transpose_uint8_int32_avx2
  2088  # %bb.0:
  2089  	push	rbp
  2090  	mov	rbp, rsp
  2091  	and	rsp, -8
  2092  	cmp	edx, 4
  2093  	jl	.LBB40_1
  2094  	.p2align	4, 0x90
  2095  .LBB40_5:                               # =>This Inner Loop Header: Depth=1
  2096  	mov	eax, edx
  2097  	movzx	edx, byte ptr [rdi]
  2098  	mov	edx, dword ptr [rcx + 4*rdx]
  2099  	mov	dword ptr [rsi], edx
  2100  	movzx	edx, byte ptr [rdi + 1]
  2101  	mov	edx, dword ptr [rcx + 4*rdx]
  2102  	mov	dword ptr [rsi + 4], edx
  2103  	movzx	edx, byte ptr [rdi + 2]
  2104  	mov	edx, dword ptr [rcx + 4*rdx]
  2105  	mov	dword ptr [rsi + 8], edx
  2106  	movzx	edx, byte ptr [rdi + 3]
  2107  	mov	edx, dword ptr [rcx + 4*rdx]
  2108  	mov	dword ptr [rsi + 12], edx
  2109  	lea	edx, [rax - 4]
  2110  	add	rdi, 4
  2111  	add	rsi, 16
  2112  	cmp	eax, 7
  2113  	jg	.LBB40_5
  2114  .LBB40_1:
  2115  	test	edx, edx
  2116  	jle	.LBB40_4
  2117  # %bb.2:
  2118  	add	edx, 1
  2119  	xor	r8d, r8d
  2120  	.p2align	4, 0x90
  2121  .LBB40_3:                               # =>This Inner Loop Header: Depth=1
  2122  	movzx	eax, byte ptr [rdi + r8]
  2123  	mov	eax, dword ptr [rcx + 4*rax]
  2124  	mov	dword ptr [rsi + 4*r8], eax
  2125  	add	r8, 1
  2126  	add	edx, -1
  2127  	cmp	edx, 1
  2128  	jg	.LBB40_3
  2129  .LBB40_4:
  2130  	mov	rsp, rbp
  2131  	pop	rbp
  2132  	ret
  2133  .Lfunc_end40:
  2134  	.size	transpose_uint8_int32_avx2, .Lfunc_end40-transpose_uint8_int32_avx2
  2135                                          # -- End function
  2136  	.globl	transpose_int8_int32_avx2       # -- Begin function transpose_int8_int32_avx2
  2137  	.p2align	4, 0x90
  2138  	.type	transpose_int8_int32_avx2,@function
  2139  transpose_int8_int32_avx2:              # @transpose_int8_int32_avx2
  2140  # %bb.0:
  2141  	push	rbp
  2142  	mov	rbp, rsp
  2143  	and	rsp, -8
  2144  	cmp	edx, 4
  2145  	jl	.LBB41_1
  2146  	.p2align	4, 0x90
  2147  .LBB41_5:                               # =>This Inner Loop Header: Depth=1
  2148  	mov	eax, edx
  2149  	movsx	rdx, byte ptr [rdi]
  2150  	mov	edx, dword ptr [rcx + 4*rdx]
  2151  	mov	dword ptr [rsi], edx
  2152  	movsx	rdx, byte ptr [rdi + 1]
  2153  	mov	edx, dword ptr [rcx + 4*rdx]
  2154  	mov	dword ptr [rsi + 4], edx
  2155  	movsx	rdx, byte ptr [rdi + 2]
  2156  	mov	edx, dword ptr [rcx + 4*rdx]
  2157  	mov	dword ptr [rsi + 8], edx
  2158  	movsx	rdx, byte ptr [rdi + 3]
  2159  	mov	edx, dword ptr [rcx + 4*rdx]
  2160  	mov	dword ptr [rsi + 12], edx
  2161  	lea	edx, [rax - 4]
  2162  	add	rdi, 4
  2163  	add	rsi, 16
  2164  	cmp	eax, 7
  2165  	jg	.LBB41_5
  2166  .LBB41_1:
  2167  	test	edx, edx
  2168  	jle	.LBB41_4
  2169  # %bb.2:
  2170  	add	edx, 1
  2171  	xor	r8d, r8d
  2172  	.p2align	4, 0x90
  2173  .LBB41_3:                               # =>This Inner Loop Header: Depth=1
  2174  	movsx	rax, byte ptr [rdi + r8]
  2175  	mov	eax, dword ptr [rcx + 4*rax]
  2176  	mov	dword ptr [rsi + 4*r8], eax
  2177  	add	r8, 1
  2178  	add	edx, -1
  2179  	cmp	edx, 1
  2180  	jg	.LBB41_3
  2181  .LBB41_4:
  2182  	mov	rsp, rbp
  2183  	pop	rbp
  2184  	ret
  2185  .Lfunc_end41:
  2186  	.size	transpose_int8_int32_avx2, .Lfunc_end41-transpose_int8_int32_avx2
  2187                                          # -- End function
  2188  	.globl	transpose_uint16_int32_avx2     # -- Begin function transpose_uint16_int32_avx2
  2189  	.p2align	4, 0x90
  2190  	.type	transpose_uint16_int32_avx2,@function
  2191  transpose_uint16_int32_avx2:            # @transpose_uint16_int32_avx2
  2192  # %bb.0:
  2193  	push	rbp
  2194  	mov	rbp, rsp
  2195  	and	rsp, -8
  2196  	cmp	edx, 4
  2197  	jl	.LBB42_1
  2198  	.p2align	4, 0x90
  2199  .LBB42_5:                               # =>This Inner Loop Header: Depth=1
  2200  	mov	eax, edx
  2201  	movzx	edx, word ptr [rdi]
  2202  	mov	edx, dword ptr [rcx + 4*rdx]
  2203  	mov	dword ptr [rsi], edx
  2204  	movzx	edx, word ptr [rdi + 2]
  2205  	mov	edx, dword ptr [rcx + 4*rdx]
  2206  	mov	dword ptr [rsi + 4], edx
  2207  	movzx	edx, word ptr [rdi + 4]
  2208  	mov	edx, dword ptr [rcx + 4*rdx]
  2209  	mov	dword ptr [rsi + 8], edx
  2210  	movzx	edx, word ptr [rdi + 6]
  2211  	mov	edx, dword ptr [rcx + 4*rdx]
  2212  	mov	dword ptr [rsi + 12], edx
  2213  	lea	edx, [rax - 4]
  2214  	add	rdi, 8
  2215  	add	rsi, 16
  2216  	cmp	eax, 7
  2217  	jg	.LBB42_5
  2218  .LBB42_1:
  2219  	test	edx, edx
  2220  	jle	.LBB42_4
  2221  # %bb.2:
  2222  	add	edx, 1
  2223  	xor	r8d, r8d
  2224  	.p2align	4, 0x90
  2225  .LBB42_3:                               # =>This Inner Loop Header: Depth=1
  2226  	movzx	eax, word ptr [rdi + r8]
  2227  	mov	eax, dword ptr [rcx + 4*rax]
  2228  	mov	dword ptr [rsi + 2*r8], eax
  2229  	add	r8, 2
  2230  	add	edx, -1
  2231  	cmp	edx, 1
  2232  	jg	.LBB42_3
  2233  .LBB42_4:
  2234  	mov	rsp, rbp
  2235  	pop	rbp
  2236  	ret
  2237  .Lfunc_end42:
  2238  	.size	transpose_uint16_int32_avx2, .Lfunc_end42-transpose_uint16_int32_avx2
  2239                                          # -- End function
  2240  	.globl	transpose_int16_int32_avx2      # -- Begin function transpose_int16_int32_avx2
  2241  	.p2align	4, 0x90
  2242  	.type	transpose_int16_int32_avx2,@function
  2243  transpose_int16_int32_avx2:             # @transpose_int16_int32_avx2
  2244  # %bb.0:
  2245  	push	rbp
  2246  	mov	rbp, rsp
  2247  	and	rsp, -8
  2248  	cmp	edx, 4
  2249  	jl	.LBB43_1
  2250  	.p2align	4, 0x90
  2251  .LBB43_5:                               # =>This Inner Loop Header: Depth=1
  2252  	mov	eax, edx
  2253  	movsx	rdx, word ptr [rdi]
  2254  	mov	edx, dword ptr [rcx + 4*rdx]
  2255  	mov	dword ptr [rsi], edx
  2256  	movsx	rdx, word ptr [rdi + 2]
  2257  	mov	edx, dword ptr [rcx + 4*rdx]
  2258  	mov	dword ptr [rsi + 4], edx
  2259  	movsx	rdx, word ptr [rdi + 4]
  2260  	mov	edx, dword ptr [rcx + 4*rdx]
  2261  	mov	dword ptr [rsi + 8], edx
  2262  	movsx	rdx, word ptr [rdi + 6]
  2263  	mov	edx, dword ptr [rcx + 4*rdx]
  2264  	mov	dword ptr [rsi + 12], edx
  2265  	lea	edx, [rax - 4]
  2266  	add	rdi, 8
  2267  	add	rsi, 16
  2268  	cmp	eax, 7
  2269  	jg	.LBB43_5
  2270  .LBB43_1:
  2271  	test	edx, edx
  2272  	jle	.LBB43_4
  2273  # %bb.2:
  2274  	add	edx, 1
  2275  	xor	r8d, r8d
  2276  	.p2align	4, 0x90
  2277  .LBB43_3:                               # =>This Inner Loop Header: Depth=1
  2278  	movsx	rax, word ptr [rdi + r8]
  2279  	mov	eax, dword ptr [rcx + 4*rax]
  2280  	mov	dword ptr [rsi + 2*r8], eax
  2281  	add	r8, 2
  2282  	add	edx, -1
  2283  	cmp	edx, 1
  2284  	jg	.LBB43_3
  2285  .LBB43_4:
  2286  	mov	rsp, rbp
  2287  	pop	rbp
  2288  	ret
  2289  .Lfunc_end43:
  2290  	.size	transpose_int16_int32_avx2, .Lfunc_end43-transpose_int16_int32_avx2
  2291                                          # -- End function
  2292  	.globl	transpose_uint32_int32_avx2     # -- Begin function transpose_uint32_int32_avx2
  2293  	.p2align	4, 0x90
  2294  	.type	transpose_uint32_int32_avx2,@function
  2295  transpose_uint32_int32_avx2:            # @transpose_uint32_int32_avx2
  2296  # %bb.0:
  2297  	push	rbp
  2298  	mov	rbp, rsp
  2299  	and	rsp, -8
  2300  	cmp	edx, 4
  2301  	jl	.LBB44_1
  2302  	.p2align	4, 0x90
  2303  .LBB44_5:                               # =>This Inner Loop Header: Depth=1
  2304  	mov	eax, edx
  2305  	mov	edx, dword ptr [rdi]
  2306  	mov	edx, dword ptr [rcx + 4*rdx]
  2307  	mov	dword ptr [rsi], edx
  2308  	mov	edx, dword ptr [rdi + 4]
  2309  	mov	edx, dword ptr [rcx + 4*rdx]
  2310  	mov	dword ptr [rsi + 4], edx
  2311  	mov	edx, dword ptr [rdi + 8]
  2312  	mov	edx, dword ptr [rcx + 4*rdx]
  2313  	mov	dword ptr [rsi + 8], edx
  2314  	mov	edx, dword ptr [rdi + 12]
  2315  	mov	edx, dword ptr [rcx + 4*rdx]
  2316  	mov	dword ptr [rsi + 12], edx
  2317  	lea	edx, [rax - 4]
  2318  	add	rdi, 16
  2319  	add	rsi, 16
  2320  	cmp	eax, 7
  2321  	jg	.LBB44_5
  2322  .LBB44_1:
  2323  	test	edx, edx
  2324  	jle	.LBB44_4
  2325  # %bb.2:
  2326  	add	edx, 1
  2327  	xor	r8d, r8d
  2328  	.p2align	4, 0x90
  2329  .LBB44_3:                               # =>This Inner Loop Header: Depth=1
  2330  	mov	eax, dword ptr [rdi + r8]
  2331  	mov	eax, dword ptr [rcx + 4*rax]
  2332  	mov	dword ptr [rsi + r8], eax
  2333  	add	r8, 4
  2334  	add	edx, -1
  2335  	cmp	edx, 1
  2336  	jg	.LBB44_3
  2337  .LBB44_4:
  2338  	mov	rsp, rbp
  2339  	pop	rbp
  2340  	ret
  2341  .Lfunc_end44:
  2342  	.size	transpose_uint32_int32_avx2, .Lfunc_end44-transpose_uint32_int32_avx2
  2343                                          # -- End function
  2344  	.globl	transpose_int32_int32_avx2      # -- Begin function transpose_int32_int32_avx2
  2345  	.p2align	4, 0x90
  2346  	.type	transpose_int32_int32_avx2,@function
  2347  transpose_int32_int32_avx2:             # @transpose_int32_int32_avx2
  2348  # %bb.0:
  2349  	push	rbp
  2350  	mov	rbp, rsp
  2351  	and	rsp, -8
  2352  	cmp	edx, 4
  2353  	jl	.LBB45_1
  2354  	.p2align	4, 0x90
  2355  .LBB45_5:                               # =>This Inner Loop Header: Depth=1
  2356  	mov	eax, edx
  2357  	movsxd	rdx, dword ptr [rdi]
  2358  	mov	edx, dword ptr [rcx + 4*rdx]
  2359  	mov	dword ptr [rsi], edx
  2360  	movsxd	rdx, dword ptr [rdi + 4]
  2361  	mov	edx, dword ptr [rcx + 4*rdx]
  2362  	mov	dword ptr [rsi + 4], edx
  2363  	movsxd	rdx, dword ptr [rdi + 8]
  2364  	mov	edx, dword ptr [rcx + 4*rdx]
  2365  	mov	dword ptr [rsi + 8], edx
  2366  	movsxd	rdx, dword ptr [rdi + 12]
  2367  	mov	edx, dword ptr [rcx + 4*rdx]
  2368  	mov	dword ptr [rsi + 12], edx
  2369  	lea	edx, [rax - 4]
  2370  	add	rdi, 16
  2371  	add	rsi, 16
  2372  	cmp	eax, 7
  2373  	jg	.LBB45_5
  2374  .LBB45_1:
  2375  	test	edx, edx
  2376  	jle	.LBB45_4
  2377  # %bb.2:
  2378  	add	edx, 1
  2379  	xor	r8d, r8d
  2380  	.p2align	4, 0x90
  2381  .LBB45_3:                               # =>This Inner Loop Header: Depth=1
  2382  	movsxd	rax, dword ptr [rdi + r8]
  2383  	mov	eax, dword ptr [rcx + 4*rax]
  2384  	mov	dword ptr [rsi + r8], eax
  2385  	add	r8, 4
  2386  	add	edx, -1
  2387  	cmp	edx, 1
  2388  	jg	.LBB45_3
  2389  .LBB45_4:
  2390  	mov	rsp, rbp
  2391  	pop	rbp
  2392  	ret
  2393  .Lfunc_end45:
  2394  	.size	transpose_int32_int32_avx2, .Lfunc_end45-transpose_int32_int32_avx2
  2395                                          # -- End function
  2396  	.globl	transpose_uint64_int32_avx2     # -- Begin function transpose_uint64_int32_avx2
  2397  	.p2align	4, 0x90
  2398  	.type	transpose_uint64_int32_avx2,@function
  2399  transpose_uint64_int32_avx2:            # @transpose_uint64_int32_avx2
  2400  # %bb.0:
  2401  	push	rbp
  2402  	mov	rbp, rsp
  2403  	and	rsp, -8
  2404  	cmp	edx, 4
  2405  	jl	.LBB46_1
  2406  	.p2align	4, 0x90
  2407  .LBB46_5:                               # =>This Inner Loop Header: Depth=1
  2408  	mov	eax, edx
  2409  	mov	rdx, qword ptr [rdi]
  2410  	mov	edx, dword ptr [rcx + 4*rdx]
  2411  	mov	dword ptr [rsi], edx
  2412  	mov	rdx, qword ptr [rdi + 8]
  2413  	mov	edx, dword ptr [rcx + 4*rdx]
  2414  	mov	dword ptr [rsi + 4], edx
  2415  	mov	rdx, qword ptr [rdi + 16]
  2416  	mov	edx, dword ptr [rcx + 4*rdx]
  2417  	mov	dword ptr [rsi + 8], edx
  2418  	mov	rdx, qword ptr [rdi + 24]
  2419  	mov	edx, dword ptr [rcx + 4*rdx]
  2420  	mov	dword ptr [rsi + 12], edx
  2421  	lea	edx, [rax - 4]
  2422  	add	rdi, 32
  2423  	add	rsi, 16
  2424  	cmp	eax, 7
  2425  	jg	.LBB46_5
  2426  .LBB46_1:
  2427  	test	edx, edx
  2428  	jle	.LBB46_4
  2429  # %bb.2:
  2430  	add	edx, 1
  2431  	xor	r8d, r8d
  2432  	.p2align	4, 0x90
  2433  .LBB46_3:                               # =>This Inner Loop Header: Depth=1
  2434  	mov	rax, qword ptr [rdi + 2*r8]
  2435  	mov	eax, dword ptr [rcx + 4*rax]
  2436  	mov	dword ptr [rsi + r8], eax
  2437  	add	r8, 4
  2438  	add	edx, -1
  2439  	cmp	edx, 1
  2440  	jg	.LBB46_3
  2441  .LBB46_4:
  2442  	mov	rsp, rbp
  2443  	pop	rbp
  2444  	ret
  2445  .Lfunc_end46:
  2446  	.size	transpose_uint64_int32_avx2, .Lfunc_end46-transpose_uint64_int32_avx2
  2447                                          # -- End function
  2448  	.globl	transpose_int64_int32_avx2      # -- Begin function transpose_int64_int32_avx2
  2449  	.p2align	4, 0x90
  2450  	.type	transpose_int64_int32_avx2,@function
  2451  transpose_int64_int32_avx2:             # @transpose_int64_int32_avx2
  2452  # %bb.0:
  2453  	push	rbp
  2454  	mov	rbp, rsp
  2455  	and	rsp, -8
  2456  	cmp	edx, 4
  2457  	jl	.LBB47_1
  2458  	.p2align	4, 0x90
  2459  .LBB47_5:                               # =>This Inner Loop Header: Depth=1
  2460  	mov	eax, edx
  2461  	mov	rdx, qword ptr [rdi]
  2462  	mov	edx, dword ptr [rcx + 4*rdx]
  2463  	mov	dword ptr [rsi], edx
  2464  	mov	rdx, qword ptr [rdi + 8]
  2465  	mov	edx, dword ptr [rcx + 4*rdx]
  2466  	mov	dword ptr [rsi + 4], edx
  2467  	mov	rdx, qword ptr [rdi + 16]
  2468  	mov	edx, dword ptr [rcx + 4*rdx]
  2469  	mov	dword ptr [rsi + 8], edx
  2470  	mov	rdx, qword ptr [rdi + 24]
  2471  	mov	edx, dword ptr [rcx + 4*rdx]
  2472  	mov	dword ptr [rsi + 12], edx
  2473  	lea	edx, [rax - 4]
  2474  	add	rdi, 32
  2475  	add	rsi, 16
  2476  	cmp	eax, 7
  2477  	jg	.LBB47_5
  2478  .LBB47_1:
  2479  	test	edx, edx
  2480  	jle	.LBB47_4
  2481  # %bb.2:
  2482  	add	edx, 1
  2483  	xor	r8d, r8d
  2484  	.p2align	4, 0x90
  2485  .LBB47_3:                               # =>This Inner Loop Header: Depth=1
  2486  	mov	rax, qword ptr [rdi + 2*r8]
  2487  	mov	eax, dword ptr [rcx + 4*rax]
  2488  	mov	dword ptr [rsi + r8], eax
  2489  	add	r8, 4
  2490  	add	edx, -1
  2491  	cmp	edx, 1
  2492  	jg	.LBB47_3
  2493  .LBB47_4:
  2494  	mov	rsp, rbp
  2495  	pop	rbp
  2496  	ret
  2497  .Lfunc_end47:
  2498  	.size	transpose_int64_int32_avx2, .Lfunc_end47-transpose_int64_int32_avx2
  2499                                          # -- End function
  2500  	.globl	transpose_uint8_uint64_avx2     # -- Begin function transpose_uint8_uint64_avx2
  2501  	.p2align	4, 0x90
  2502  	.type	transpose_uint8_uint64_avx2,@function
  2503  transpose_uint8_uint64_avx2:            # @transpose_uint8_uint64_avx2
  2504  # %bb.0:
  2505  	push	rbp
  2506  	mov	rbp, rsp
  2507  	and	rsp, -8
  2508  	cmp	edx, 4
  2509  	jl	.LBB48_1
  2510  	.p2align	4, 0x90
  2511  .LBB48_5:                               # =>This Inner Loop Header: Depth=1
  2512  	mov	eax, edx
  2513  	movzx	edx, byte ptr [rdi]
  2514  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2515  	mov	qword ptr [rsi], rdx
  2516  	movzx	edx, byte ptr [rdi + 1]
  2517  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2518  	mov	qword ptr [rsi + 8], rdx
  2519  	movzx	edx, byte ptr [rdi + 2]
  2520  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2521  	mov	qword ptr [rsi + 16], rdx
  2522  	movzx	edx, byte ptr [rdi + 3]
  2523  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2524  	mov	qword ptr [rsi + 24], rdx
  2525  	lea	edx, [rax - 4]
  2526  	add	rdi, 4
  2527  	add	rsi, 32
  2528  	cmp	eax, 7
  2529  	jg	.LBB48_5
  2530  .LBB48_1:
  2531  	test	edx, edx
  2532  	jle	.LBB48_4
  2533  # %bb.2:
  2534  	add	edx, 1
  2535  	xor	r8d, r8d
  2536  	.p2align	4, 0x90
  2537  .LBB48_3:                               # =>This Inner Loop Header: Depth=1
  2538  	movzx	eax, byte ptr [rdi + r8]
  2539  	movsxd	rax, dword ptr [rcx + 4*rax]
  2540  	mov	qword ptr [rsi + 8*r8], rax
  2541  	add	r8, 1
  2542  	add	edx, -1
  2543  	cmp	edx, 1
  2544  	jg	.LBB48_3
  2545  .LBB48_4:
  2546  	mov	rsp, rbp
  2547  	pop	rbp
  2548  	ret
  2549  .Lfunc_end48:
  2550  	.size	transpose_uint8_uint64_avx2, .Lfunc_end48-transpose_uint8_uint64_avx2
  2551                                          # -- End function
  2552  	.globl	transpose_int8_uint64_avx2      # -- Begin function transpose_int8_uint64_avx2
  2553  	.p2align	4, 0x90
  2554  	.type	transpose_int8_uint64_avx2,@function
  2555  transpose_int8_uint64_avx2:             # @transpose_int8_uint64_avx2
  2556  # %bb.0:
  2557  	push	rbp
  2558  	mov	rbp, rsp
  2559  	and	rsp, -8
  2560  	cmp	edx, 4
  2561  	jl	.LBB49_1
  2562  	.p2align	4, 0x90
  2563  .LBB49_5:                               # =>This Inner Loop Header: Depth=1
  2564  	mov	eax, edx
  2565  	movsx	rdx, byte ptr [rdi]
  2566  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2567  	mov	qword ptr [rsi], rdx
  2568  	movsx	rdx, byte ptr [rdi + 1]
  2569  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2570  	mov	qword ptr [rsi + 8], rdx
  2571  	movsx	rdx, byte ptr [rdi + 2]
  2572  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2573  	mov	qword ptr [rsi + 16], rdx
  2574  	movsx	rdx, byte ptr [rdi + 3]
  2575  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2576  	mov	qword ptr [rsi + 24], rdx
  2577  	lea	edx, [rax - 4]
  2578  	add	rdi, 4
  2579  	add	rsi, 32
  2580  	cmp	eax, 7
  2581  	jg	.LBB49_5
  2582  .LBB49_1:
  2583  	test	edx, edx
  2584  	jle	.LBB49_4
  2585  # %bb.2:
  2586  	add	edx, 1
  2587  	xor	r8d, r8d
  2588  	.p2align	4, 0x90
  2589  .LBB49_3:                               # =>This Inner Loop Header: Depth=1
  2590  	movsx	rax, byte ptr [rdi + r8]
  2591  	movsxd	rax, dword ptr [rcx + 4*rax]
  2592  	mov	qword ptr [rsi + 8*r8], rax
  2593  	add	r8, 1
  2594  	add	edx, -1
  2595  	cmp	edx, 1
  2596  	jg	.LBB49_3
  2597  .LBB49_4:
  2598  	mov	rsp, rbp
  2599  	pop	rbp
  2600  	ret
  2601  .Lfunc_end49:
  2602  	.size	transpose_int8_uint64_avx2, .Lfunc_end49-transpose_int8_uint64_avx2
  2603                                          # -- End function
  2604  	.globl	transpose_uint16_uint64_avx2    # -- Begin function transpose_uint16_uint64_avx2
  2605  	.p2align	4, 0x90
  2606  	.type	transpose_uint16_uint64_avx2,@function
  2607  transpose_uint16_uint64_avx2:           # @transpose_uint16_uint64_avx2
  2608  # %bb.0:
  2609  	push	rbp
  2610  	mov	rbp, rsp
  2611  	and	rsp, -8
  2612  	cmp	edx, 4
  2613  	jl	.LBB50_1
  2614  	.p2align	4, 0x90
  2615  .LBB50_5:                               # =>This Inner Loop Header: Depth=1
  2616  	mov	eax, edx
  2617  	movzx	edx, word ptr [rdi]
  2618  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2619  	mov	qword ptr [rsi], rdx
  2620  	movzx	edx, word ptr [rdi + 2]
  2621  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2622  	mov	qword ptr [rsi + 8], rdx
  2623  	movzx	edx, word ptr [rdi + 4]
  2624  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2625  	mov	qword ptr [rsi + 16], rdx
  2626  	movzx	edx, word ptr [rdi + 6]
  2627  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2628  	mov	qword ptr [rsi + 24], rdx
  2629  	lea	edx, [rax - 4]
  2630  	add	rdi, 8
  2631  	add	rsi, 32
  2632  	cmp	eax, 7
  2633  	jg	.LBB50_5
  2634  .LBB50_1:
  2635  	test	edx, edx
  2636  	jle	.LBB50_4
  2637  # %bb.2:
  2638  	add	edx, 1
  2639  	xor	r8d, r8d
  2640  	.p2align	4, 0x90
  2641  .LBB50_3:                               # =>This Inner Loop Header: Depth=1
  2642  	movzx	eax, word ptr [rdi + r8]
  2643  	movsxd	rax, dword ptr [rcx + 4*rax]
  2644  	mov	qword ptr [rsi + 4*r8], rax
  2645  	add	r8, 2
  2646  	add	edx, -1
  2647  	cmp	edx, 1
  2648  	jg	.LBB50_3
  2649  .LBB50_4:
  2650  	mov	rsp, rbp
  2651  	pop	rbp
  2652  	ret
  2653  .Lfunc_end50:
  2654  	.size	transpose_uint16_uint64_avx2, .Lfunc_end50-transpose_uint16_uint64_avx2
  2655                                          # -- End function
  2656  	.globl	transpose_int16_uint64_avx2     # -- Begin function transpose_int16_uint64_avx2
  2657  	.p2align	4, 0x90
  2658  	.type	transpose_int16_uint64_avx2,@function
  2659  transpose_int16_uint64_avx2:            # @transpose_int16_uint64_avx2
  2660  # %bb.0:
  2661  	push	rbp
  2662  	mov	rbp, rsp
  2663  	and	rsp, -8
  2664  	cmp	edx, 4
  2665  	jl	.LBB51_1
  2666  	.p2align	4, 0x90
  2667  .LBB51_5:                               # =>This Inner Loop Header: Depth=1
  2668  	mov	eax, edx
  2669  	movsx	rdx, word ptr [rdi]
  2670  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2671  	mov	qword ptr [rsi], rdx
  2672  	movsx	rdx, word ptr [rdi + 2]
  2673  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2674  	mov	qword ptr [rsi + 8], rdx
  2675  	movsx	rdx, word ptr [rdi + 4]
  2676  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2677  	mov	qword ptr [rsi + 16], rdx
  2678  	movsx	rdx, word ptr [rdi + 6]
  2679  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2680  	mov	qword ptr [rsi + 24], rdx
  2681  	lea	edx, [rax - 4]
  2682  	add	rdi, 8
  2683  	add	rsi, 32
  2684  	cmp	eax, 7
  2685  	jg	.LBB51_5
  2686  .LBB51_1:
  2687  	test	edx, edx
  2688  	jle	.LBB51_4
  2689  # %bb.2:
  2690  	add	edx, 1
  2691  	xor	r8d, r8d
  2692  	.p2align	4, 0x90
  2693  .LBB51_3:                               # =>This Inner Loop Header: Depth=1
  2694  	movsx	rax, word ptr [rdi + r8]
  2695  	movsxd	rax, dword ptr [rcx + 4*rax]
  2696  	mov	qword ptr [rsi + 4*r8], rax
  2697  	add	r8, 2
  2698  	add	edx, -1
  2699  	cmp	edx, 1
  2700  	jg	.LBB51_3
  2701  .LBB51_4:
  2702  	mov	rsp, rbp
  2703  	pop	rbp
  2704  	ret
  2705  .Lfunc_end51:
  2706  	.size	transpose_int16_uint64_avx2, .Lfunc_end51-transpose_int16_uint64_avx2
  2707                                          # -- End function
  2708  	.globl	transpose_uint32_uint64_avx2    # -- Begin function transpose_uint32_uint64_avx2
  2709  	.p2align	4, 0x90
  2710  	.type	transpose_uint32_uint64_avx2,@function
  2711  transpose_uint32_uint64_avx2:           # @transpose_uint32_uint64_avx2
  2712  # %bb.0:
  2713  	push	rbp
  2714  	mov	rbp, rsp
  2715  	and	rsp, -8
  2716  	cmp	edx, 4
  2717  	jl	.LBB52_1
  2718  	.p2align	4, 0x90
  2719  .LBB52_5:                               # =>This Inner Loop Header: Depth=1
  2720  	mov	eax, edx
  2721  	mov	edx, dword ptr [rdi]
  2722  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2723  	mov	qword ptr [rsi], rdx
  2724  	mov	edx, dword ptr [rdi + 4]
  2725  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2726  	mov	qword ptr [rsi + 8], rdx
  2727  	mov	edx, dword ptr [rdi + 8]
  2728  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2729  	mov	qword ptr [rsi + 16], rdx
  2730  	mov	edx, dword ptr [rdi + 12]
  2731  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2732  	mov	qword ptr [rsi + 24], rdx
  2733  	lea	edx, [rax - 4]
  2734  	add	rdi, 16
  2735  	add	rsi, 32
  2736  	cmp	eax, 7
  2737  	jg	.LBB52_5
  2738  .LBB52_1:
  2739  	test	edx, edx
  2740  	jle	.LBB52_4
  2741  # %bb.2:
  2742  	add	edx, 1
  2743  	xor	r8d, r8d
  2744  	.p2align	4, 0x90
  2745  .LBB52_3:                               # =>This Inner Loop Header: Depth=1
  2746  	mov	eax, dword ptr [rdi + r8]
  2747  	movsxd	rax, dword ptr [rcx + 4*rax]
  2748  	mov	qword ptr [rsi + 2*r8], rax
  2749  	add	r8, 4
  2750  	add	edx, -1
  2751  	cmp	edx, 1
  2752  	jg	.LBB52_3
  2753  .LBB52_4:
  2754  	mov	rsp, rbp
  2755  	pop	rbp
  2756  	ret
  2757  .Lfunc_end52:
  2758  	.size	transpose_uint32_uint64_avx2, .Lfunc_end52-transpose_uint32_uint64_avx2
  2759                                          # -- End function
  2760  	.globl	transpose_int32_uint64_avx2     # -- Begin function transpose_int32_uint64_avx2
  2761  	.p2align	4, 0x90
  2762  	.type	transpose_int32_uint64_avx2,@function
  2763  transpose_int32_uint64_avx2:            # @transpose_int32_uint64_avx2
  2764  # %bb.0:
  2765  	push	rbp
  2766  	mov	rbp, rsp
  2767  	and	rsp, -8
  2768  	cmp	edx, 4
  2769  	jl	.LBB53_1
  2770  	.p2align	4, 0x90
  2771  .LBB53_5:                               # =>This Inner Loop Header: Depth=1
  2772  	mov	eax, edx
  2773  	movsxd	rdx, dword ptr [rdi]
  2774  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2775  	mov	qword ptr [rsi], rdx
  2776  	movsxd	rdx, dword ptr [rdi + 4]
  2777  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2778  	mov	qword ptr [rsi + 8], rdx
  2779  	movsxd	rdx, dword ptr [rdi + 8]
  2780  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2781  	mov	qword ptr [rsi + 16], rdx
  2782  	movsxd	rdx, dword ptr [rdi + 12]
  2783  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2784  	mov	qword ptr [rsi + 24], rdx
  2785  	lea	edx, [rax - 4]
  2786  	add	rdi, 16
  2787  	add	rsi, 32
  2788  	cmp	eax, 7
  2789  	jg	.LBB53_5
  2790  .LBB53_1:
  2791  	test	edx, edx
  2792  	jle	.LBB53_4
  2793  # %bb.2:
  2794  	add	edx, 1
  2795  	xor	r8d, r8d
  2796  	.p2align	4, 0x90
  2797  .LBB53_3:                               # =>This Inner Loop Header: Depth=1
  2798  	movsxd	rax, dword ptr [rdi + r8]
  2799  	movsxd	rax, dword ptr [rcx + 4*rax]
  2800  	mov	qword ptr [rsi + 2*r8], rax
  2801  	add	r8, 4
  2802  	add	edx, -1
  2803  	cmp	edx, 1
  2804  	jg	.LBB53_3
  2805  .LBB53_4:
  2806  	mov	rsp, rbp
  2807  	pop	rbp
  2808  	ret
  2809  .Lfunc_end53:
  2810  	.size	transpose_int32_uint64_avx2, .Lfunc_end53-transpose_int32_uint64_avx2
  2811                                          # -- End function
  2812  	.globl	transpose_uint64_uint64_avx2    # -- Begin function transpose_uint64_uint64_avx2
  2813  	.p2align	4, 0x90
  2814  	.type	transpose_uint64_uint64_avx2,@function
  2815  transpose_uint64_uint64_avx2:           # @transpose_uint64_uint64_avx2
  2816  # %bb.0:
  2817  	push	rbp
  2818  	mov	rbp, rsp
  2819  	and	rsp, -8
  2820  	cmp	edx, 4
  2821  	jl	.LBB54_1
  2822  	.p2align	4, 0x90
  2823  .LBB54_5:                               # =>This Inner Loop Header: Depth=1
  2824  	mov	eax, edx
  2825  	mov	rdx, qword ptr [rdi]
  2826  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2827  	mov	qword ptr [rsi], rdx
  2828  	mov	rdx, qword ptr [rdi + 8]
  2829  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2830  	mov	qword ptr [rsi + 8], rdx
  2831  	mov	rdx, qword ptr [rdi + 16]
  2832  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2833  	mov	qword ptr [rsi + 16], rdx
  2834  	mov	rdx, qword ptr [rdi + 24]
  2835  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2836  	mov	qword ptr [rsi + 24], rdx
  2837  	lea	edx, [rax - 4]
  2838  	add	rdi, 32
  2839  	add	rsi, 32
  2840  	cmp	eax, 7
  2841  	jg	.LBB54_5
  2842  .LBB54_1:
  2843  	test	edx, edx
  2844  	jle	.LBB54_4
  2845  # %bb.2:
  2846  	add	edx, 1
  2847  	xor	r8d, r8d
  2848  	.p2align	4, 0x90
  2849  .LBB54_3:                               # =>This Inner Loop Header: Depth=1
  2850  	mov	rax, qword ptr [rdi + r8]
  2851  	movsxd	rax, dword ptr [rcx + 4*rax]
  2852  	mov	qword ptr [rsi + r8], rax
  2853  	add	r8, 8
  2854  	add	edx, -1
  2855  	cmp	edx, 1
  2856  	jg	.LBB54_3
  2857  .LBB54_4:
  2858  	mov	rsp, rbp
  2859  	pop	rbp
  2860  	ret
  2861  .Lfunc_end54:
  2862  	.size	transpose_uint64_uint64_avx2, .Lfunc_end54-transpose_uint64_uint64_avx2
  2863                                          # -- End function
  2864  	.globl	transpose_int64_uint64_avx2     # -- Begin function transpose_int64_uint64_avx2
  2865  	.p2align	4, 0x90
  2866  	.type	transpose_int64_uint64_avx2,@function
  2867  transpose_int64_uint64_avx2:            # @transpose_int64_uint64_avx2
  2868  # %bb.0:
  2869  	push	rbp
  2870  	mov	rbp, rsp
  2871  	and	rsp, -8
  2872  	cmp	edx, 4
  2873  	jl	.LBB55_1
  2874  	.p2align	4, 0x90
  2875  .LBB55_5:                               # =>This Inner Loop Header: Depth=1
  2876  	mov	eax, edx
  2877  	mov	rdx, qword ptr [rdi]
  2878  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2879  	mov	qword ptr [rsi], rdx
  2880  	mov	rdx, qword ptr [rdi + 8]
  2881  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2882  	mov	qword ptr [rsi + 8], rdx
  2883  	mov	rdx, qword ptr [rdi + 16]
  2884  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2885  	mov	qword ptr [rsi + 16], rdx
  2886  	mov	rdx, qword ptr [rdi + 24]
  2887  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2888  	mov	qword ptr [rsi + 24], rdx
  2889  	lea	edx, [rax - 4]
  2890  	add	rdi, 32
  2891  	add	rsi, 32
  2892  	cmp	eax, 7
  2893  	jg	.LBB55_5
  2894  .LBB55_1:
  2895  	test	edx, edx
  2896  	jle	.LBB55_4
  2897  # %bb.2:
  2898  	add	edx, 1
  2899  	xor	r8d, r8d
  2900  	.p2align	4, 0x90
  2901  .LBB55_3:                               # =>This Inner Loop Header: Depth=1
  2902  	mov	rax, qword ptr [rdi + r8]
  2903  	movsxd	rax, dword ptr [rcx + 4*rax]
  2904  	mov	qword ptr [rsi + r8], rax
  2905  	add	r8, 8
  2906  	add	edx, -1
  2907  	cmp	edx, 1
  2908  	jg	.LBB55_3
  2909  .LBB55_4:
  2910  	mov	rsp, rbp
  2911  	pop	rbp
  2912  	ret
  2913  .Lfunc_end55:
  2914  	.size	transpose_int64_uint64_avx2, .Lfunc_end55-transpose_int64_uint64_avx2
  2915                                          # -- End function
  2916  	.globl	transpose_uint8_int64_avx2      # -- Begin function transpose_uint8_int64_avx2
  2917  	.p2align	4, 0x90
  2918  	.type	transpose_uint8_int64_avx2,@function
  2919  transpose_uint8_int64_avx2:             # @transpose_uint8_int64_avx2
  2920  # %bb.0:
  2921  	push	rbp
  2922  	mov	rbp, rsp
  2923  	and	rsp, -8
  2924  	cmp	edx, 4
  2925  	jl	.LBB56_1
  2926  	.p2align	4, 0x90
  2927  .LBB56_5:                               # =>This Inner Loop Header: Depth=1
  2928  	mov	eax, edx
  2929  	movzx	edx, byte ptr [rdi]
  2930  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2931  	mov	qword ptr [rsi], rdx
  2932  	movzx	edx, byte ptr [rdi + 1]
  2933  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2934  	mov	qword ptr [rsi + 8], rdx
  2935  	movzx	edx, byte ptr [rdi + 2]
  2936  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2937  	mov	qword ptr [rsi + 16], rdx
  2938  	movzx	edx, byte ptr [rdi + 3]
  2939  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2940  	mov	qword ptr [rsi + 24], rdx
  2941  	lea	edx, [rax - 4]
  2942  	add	rdi, 4
  2943  	add	rsi, 32
  2944  	cmp	eax, 7
  2945  	jg	.LBB56_5
  2946  .LBB56_1:
  2947  	test	edx, edx
  2948  	jle	.LBB56_4
  2949  # %bb.2:
  2950  	add	edx, 1
  2951  	xor	r8d, r8d
  2952  	.p2align	4, 0x90
  2953  .LBB56_3:                               # =>This Inner Loop Header: Depth=1
  2954  	movzx	eax, byte ptr [rdi + r8]
  2955  	movsxd	rax, dword ptr [rcx + 4*rax]
  2956  	mov	qword ptr [rsi + 8*r8], rax
  2957  	add	r8, 1
  2958  	add	edx, -1
  2959  	cmp	edx, 1
  2960  	jg	.LBB56_3
  2961  .LBB56_4:
  2962  	mov	rsp, rbp
  2963  	pop	rbp
  2964  	ret
  2965  .Lfunc_end56:
  2966  	.size	transpose_uint8_int64_avx2, .Lfunc_end56-transpose_uint8_int64_avx2
  2967                                          # -- End function
  2968  	.globl	transpose_int8_int64_avx2       # -- Begin function transpose_int8_int64_avx2
  2969  	.p2align	4, 0x90
  2970  	.type	transpose_int8_int64_avx2,@function
  2971  transpose_int8_int64_avx2:              # @transpose_int8_int64_avx2
  2972  # %bb.0:
  2973  	push	rbp
  2974  	mov	rbp, rsp
  2975  	and	rsp, -8
  2976  	cmp	edx, 4
  2977  	jl	.LBB57_1
  2978  	.p2align	4, 0x90
  2979  .LBB57_5:                               # =>This Inner Loop Header: Depth=1
  2980  	mov	eax, edx
  2981  	movsx	rdx, byte ptr [rdi]
  2982  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2983  	mov	qword ptr [rsi], rdx
  2984  	movsx	rdx, byte ptr [rdi + 1]
  2985  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2986  	mov	qword ptr [rsi + 8], rdx
  2987  	movsx	rdx, byte ptr [rdi + 2]
  2988  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2989  	mov	qword ptr [rsi + 16], rdx
  2990  	movsx	rdx, byte ptr [rdi + 3]
  2991  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  2992  	mov	qword ptr [rsi + 24], rdx
  2993  	lea	edx, [rax - 4]
  2994  	add	rdi, 4
  2995  	add	rsi, 32
  2996  	cmp	eax, 7
  2997  	jg	.LBB57_5
  2998  .LBB57_1:
  2999  	test	edx, edx
  3000  	jle	.LBB57_4
  3001  # %bb.2:
  3002  	add	edx, 1
  3003  	xor	r8d, r8d
  3004  	.p2align	4, 0x90
  3005  .LBB57_3:                               # =>This Inner Loop Header: Depth=1
  3006  	movsx	rax, byte ptr [rdi + r8]
  3007  	movsxd	rax, dword ptr [rcx + 4*rax]
  3008  	mov	qword ptr [rsi + 8*r8], rax
  3009  	add	r8, 1
  3010  	add	edx, -1
  3011  	cmp	edx, 1
  3012  	jg	.LBB57_3
  3013  .LBB57_4:
  3014  	mov	rsp, rbp
  3015  	pop	rbp
  3016  	ret
  3017  .Lfunc_end57:
  3018  	.size	transpose_int8_int64_avx2, .Lfunc_end57-transpose_int8_int64_avx2
  3019                                          # -- End function
  3020  	.globl	transpose_uint16_int64_avx2     # -- Begin function transpose_uint16_int64_avx2
  3021  	.p2align	4, 0x90
  3022  	.type	transpose_uint16_int64_avx2,@function
  3023  transpose_uint16_int64_avx2:            # @transpose_uint16_int64_avx2
  3024  # %bb.0:
  3025  	push	rbp
  3026  	mov	rbp, rsp
  3027  	and	rsp, -8
  3028  	cmp	edx, 4
  3029  	jl	.LBB58_1
  3030  	.p2align	4, 0x90
  3031  .LBB58_5:                               # =>This Inner Loop Header: Depth=1
  3032  	mov	eax, edx
  3033  	movzx	edx, word ptr [rdi]
  3034  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3035  	mov	qword ptr [rsi], rdx
  3036  	movzx	edx, word ptr [rdi + 2]
  3037  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3038  	mov	qword ptr [rsi + 8], rdx
  3039  	movzx	edx, word ptr [rdi + 4]
  3040  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3041  	mov	qword ptr [rsi + 16], rdx
  3042  	movzx	edx, word ptr [rdi + 6]
  3043  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3044  	mov	qword ptr [rsi + 24], rdx
  3045  	lea	edx, [rax - 4]
  3046  	add	rdi, 8
  3047  	add	rsi, 32
  3048  	cmp	eax, 7
  3049  	jg	.LBB58_5
  3050  .LBB58_1:
  3051  	test	edx, edx
  3052  	jle	.LBB58_4
  3053  # %bb.2:
  3054  	add	edx, 1
  3055  	xor	r8d, r8d
  3056  	.p2align	4, 0x90
  3057  .LBB58_3:                               # =>This Inner Loop Header: Depth=1
  3058  	movzx	eax, word ptr [rdi + r8]
  3059  	movsxd	rax, dword ptr [rcx + 4*rax]
  3060  	mov	qword ptr [rsi + 4*r8], rax
  3061  	add	r8, 2
  3062  	add	edx, -1
  3063  	cmp	edx, 1
  3064  	jg	.LBB58_3
  3065  .LBB58_4:
  3066  	mov	rsp, rbp
  3067  	pop	rbp
  3068  	ret
  3069  .Lfunc_end58:
  3070  	.size	transpose_uint16_int64_avx2, .Lfunc_end58-transpose_uint16_int64_avx2
  3071                                          # -- End function
  3072  	.globl	transpose_int16_int64_avx2      # -- Begin function transpose_int16_int64_avx2
  3073  	.p2align	4, 0x90
  3074  	.type	transpose_int16_int64_avx2,@function
  3075  transpose_int16_int64_avx2:             # @transpose_int16_int64_avx2
  3076  # %bb.0:
  3077  	push	rbp
  3078  	mov	rbp, rsp
  3079  	and	rsp, -8
  3080  	cmp	edx, 4
  3081  	jl	.LBB59_1
  3082  	.p2align	4, 0x90
  3083  .LBB59_5:                               # =>This Inner Loop Header: Depth=1
  3084  	mov	eax, edx
  3085  	movsx	rdx, word ptr [rdi]
  3086  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3087  	mov	qword ptr [rsi], rdx
  3088  	movsx	rdx, word ptr [rdi + 2]
  3089  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3090  	mov	qword ptr [rsi + 8], rdx
  3091  	movsx	rdx, word ptr [rdi + 4]
  3092  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3093  	mov	qword ptr [rsi + 16], rdx
  3094  	movsx	rdx, word ptr [rdi + 6]
  3095  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3096  	mov	qword ptr [rsi + 24], rdx
  3097  	lea	edx, [rax - 4]
  3098  	add	rdi, 8
  3099  	add	rsi, 32
  3100  	cmp	eax, 7
  3101  	jg	.LBB59_5
  3102  .LBB59_1:
  3103  	test	edx, edx
  3104  	jle	.LBB59_4
  3105  # %bb.2:
  3106  	add	edx, 1
  3107  	xor	r8d, r8d
  3108  	.p2align	4, 0x90
  3109  .LBB59_3:                               # =>This Inner Loop Header: Depth=1
  3110  	movsx	rax, word ptr [rdi + r8]
  3111  	movsxd	rax, dword ptr [rcx + 4*rax]
  3112  	mov	qword ptr [rsi + 4*r8], rax
  3113  	add	r8, 2
  3114  	add	edx, -1
  3115  	cmp	edx, 1
  3116  	jg	.LBB59_3
  3117  .LBB59_4:
  3118  	mov	rsp, rbp
  3119  	pop	rbp
  3120  	ret
  3121  .Lfunc_end59:
  3122  	.size	transpose_int16_int64_avx2, .Lfunc_end59-transpose_int16_int64_avx2
  3123                                          # -- End function
  3124  	.globl	transpose_uint32_int64_avx2     # -- Begin function transpose_uint32_int64_avx2
  3125  	.p2align	4, 0x90
  3126  	.type	transpose_uint32_int64_avx2,@function
  3127  transpose_uint32_int64_avx2:            # @transpose_uint32_int64_avx2
  3128  # %bb.0:
  3129  	push	rbp
  3130  	mov	rbp, rsp
  3131  	and	rsp, -8
  3132  	cmp	edx, 4
  3133  	jl	.LBB60_1
  3134  	.p2align	4, 0x90
  3135  .LBB60_5:                               # =>This Inner Loop Header: Depth=1
  3136  	mov	eax, edx
  3137  	mov	edx, dword ptr [rdi]
  3138  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3139  	mov	qword ptr [rsi], rdx
  3140  	mov	edx, dword ptr [rdi + 4]
  3141  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3142  	mov	qword ptr [rsi + 8], rdx
  3143  	mov	edx, dword ptr [rdi + 8]
  3144  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3145  	mov	qword ptr [rsi + 16], rdx
  3146  	mov	edx, dword ptr [rdi + 12]
  3147  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3148  	mov	qword ptr [rsi + 24], rdx
  3149  	lea	edx, [rax - 4]
  3150  	add	rdi, 16
  3151  	add	rsi, 32
  3152  	cmp	eax, 7
  3153  	jg	.LBB60_5
  3154  .LBB60_1:
  3155  	test	edx, edx
  3156  	jle	.LBB60_4
  3157  # %bb.2:
  3158  	add	edx, 1
  3159  	xor	r8d, r8d
  3160  	.p2align	4, 0x90
  3161  .LBB60_3:                               # =>This Inner Loop Header: Depth=1
  3162  	mov	eax, dword ptr [rdi + r8]
  3163  	movsxd	rax, dword ptr [rcx + 4*rax]
  3164  	mov	qword ptr [rsi + 2*r8], rax
  3165  	add	r8, 4
  3166  	add	edx, -1
  3167  	cmp	edx, 1
  3168  	jg	.LBB60_3
  3169  .LBB60_4:
  3170  	mov	rsp, rbp
  3171  	pop	rbp
  3172  	ret
  3173  .Lfunc_end60:
  3174  	.size	transpose_uint32_int64_avx2, .Lfunc_end60-transpose_uint32_int64_avx2
  3175                                          # -- End function
  3176  	.globl	transpose_int32_int64_avx2      # -- Begin function transpose_int32_int64_avx2
  3177  	.p2align	4, 0x90
  3178  	.type	transpose_int32_int64_avx2,@function
  3179  transpose_int32_int64_avx2:             # @transpose_int32_int64_avx2
  3180  # %bb.0:
  3181  	push	rbp
  3182  	mov	rbp, rsp
  3183  	and	rsp, -8
  3184  	cmp	edx, 4
  3185  	jl	.LBB61_1
  3186  	.p2align	4, 0x90
  3187  .LBB61_5:                               # =>This Inner Loop Header: Depth=1
  3188  	mov	eax, edx
  3189  	movsxd	rdx, dword ptr [rdi]
  3190  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3191  	mov	qword ptr [rsi], rdx
  3192  	movsxd	rdx, dword ptr [rdi + 4]
  3193  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3194  	mov	qword ptr [rsi + 8], rdx
  3195  	movsxd	rdx, dword ptr [rdi + 8]
  3196  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3197  	mov	qword ptr [rsi + 16], rdx
  3198  	movsxd	rdx, dword ptr [rdi + 12]
  3199  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3200  	mov	qword ptr [rsi + 24], rdx
  3201  	lea	edx, [rax - 4]
  3202  	add	rdi, 16
  3203  	add	rsi, 32
  3204  	cmp	eax, 7
  3205  	jg	.LBB61_5
  3206  .LBB61_1:
  3207  	test	edx, edx
  3208  	jle	.LBB61_4
  3209  # %bb.2:
  3210  	add	edx, 1
  3211  	xor	r8d, r8d
  3212  	.p2align	4, 0x90
  3213  .LBB61_3:                               # =>This Inner Loop Header: Depth=1
  3214  	movsxd	rax, dword ptr [rdi + r8]
  3215  	movsxd	rax, dword ptr [rcx + 4*rax]
  3216  	mov	qword ptr [rsi + 2*r8], rax
  3217  	add	r8, 4
  3218  	add	edx, -1
  3219  	cmp	edx, 1
  3220  	jg	.LBB61_3
  3221  .LBB61_4:
  3222  	mov	rsp, rbp
  3223  	pop	rbp
  3224  	ret
  3225  .Lfunc_end61:
  3226  	.size	transpose_int32_int64_avx2, .Lfunc_end61-transpose_int32_int64_avx2
  3227                                          # -- End function
  3228  	.globl	transpose_uint64_int64_avx2     # -- Begin function transpose_uint64_int64_avx2
  3229  	.p2align	4, 0x90
  3230  	.type	transpose_uint64_int64_avx2,@function
  3231  transpose_uint64_int64_avx2:            # @transpose_uint64_int64_avx2
  3232  # %bb.0:
  3233  	push	rbp
  3234  	mov	rbp, rsp
  3235  	and	rsp, -8
  3236  	cmp	edx, 4
  3237  	jl	.LBB62_1
  3238  	.p2align	4, 0x90
  3239  .LBB62_5:                               # =>This Inner Loop Header: Depth=1
  3240  	mov	eax, edx
  3241  	mov	rdx, qword ptr [rdi]
  3242  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3243  	mov	qword ptr [rsi], rdx
  3244  	mov	rdx, qword ptr [rdi + 8]
  3245  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3246  	mov	qword ptr [rsi + 8], rdx
  3247  	mov	rdx, qword ptr [rdi + 16]
  3248  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3249  	mov	qword ptr [rsi + 16], rdx
  3250  	mov	rdx, qword ptr [rdi + 24]
  3251  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3252  	mov	qword ptr [rsi + 24], rdx
  3253  	lea	edx, [rax - 4]
  3254  	add	rdi, 32
  3255  	add	rsi, 32
  3256  	cmp	eax, 7
  3257  	jg	.LBB62_5
  3258  .LBB62_1:
  3259  	test	edx, edx
  3260  	jle	.LBB62_4
  3261  # %bb.2:
  3262  	add	edx, 1
  3263  	xor	r8d, r8d
  3264  	.p2align	4, 0x90
  3265  .LBB62_3:                               # =>This Inner Loop Header: Depth=1
  3266  	mov	rax, qword ptr [rdi + r8]
  3267  	movsxd	rax, dword ptr [rcx + 4*rax]
  3268  	mov	qword ptr [rsi + r8], rax
  3269  	add	r8, 8
  3270  	add	edx, -1
  3271  	cmp	edx, 1
  3272  	jg	.LBB62_3
  3273  .LBB62_4:
  3274  	mov	rsp, rbp
  3275  	pop	rbp
  3276  	ret
  3277  .Lfunc_end62:
  3278  	.size	transpose_uint64_int64_avx2, .Lfunc_end62-transpose_uint64_int64_avx2
  3279                                          # -- End function
  3280  	.globl	transpose_int64_int64_avx2      # -- Begin function transpose_int64_int64_avx2
  3281  	.p2align	4, 0x90
  3282  	.type	transpose_int64_int64_avx2,@function
  3283  transpose_int64_int64_avx2:             # @transpose_int64_int64_avx2
  3284  # %bb.0:
  3285  	push	rbp
  3286  	mov	rbp, rsp
  3287  	and	rsp, -8
  3288  	cmp	edx, 4
  3289  	jl	.LBB63_1
  3290  	.p2align	4, 0x90
  3291  .LBB63_5:                               # =>This Inner Loop Header: Depth=1
  3292  	mov	eax, edx
  3293  	mov	rdx, qword ptr [rdi]
  3294  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3295  	mov	qword ptr [rsi], rdx
  3296  	mov	rdx, qword ptr [rdi + 8]
  3297  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3298  	mov	qword ptr [rsi + 8], rdx
  3299  	mov	rdx, qword ptr [rdi + 16]
  3300  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3301  	mov	qword ptr [rsi + 16], rdx
  3302  	mov	rdx, qword ptr [rdi + 24]
  3303  	movsxd	rdx, dword ptr [rcx + 4*rdx]
  3304  	mov	qword ptr [rsi + 24], rdx
  3305  	lea	edx, [rax - 4]
  3306  	add	rdi, 32
  3307  	add	rsi, 32
  3308  	cmp	eax, 7
  3309  	jg	.LBB63_5
  3310  .LBB63_1:
  3311  	test	edx, edx
  3312  	jle	.LBB63_4
  3313  # %bb.2:
  3314  	add	edx, 1
  3315  	xor	r8d, r8d
  3316  	.p2align	4, 0x90
  3317  .LBB63_3:                               # =>This Inner Loop Header: Depth=1
  3318  	mov	rax, qword ptr [rdi + r8]
  3319  	movsxd	rax, dword ptr [rcx + 4*rax]
  3320  	mov	qword ptr [rsi + r8], rax
  3321  	add	r8, 8
  3322  	add	edx, -1
  3323  	cmp	edx, 1
  3324  	jg	.LBB63_3
  3325  .LBB63_4:
  3326  	mov	rsp, rbp
  3327  	pop	rbp
  3328  	ret
  3329  .Lfunc_end63:
  3330  	.size	transpose_int64_int64_avx2, .Lfunc_end63-transpose_int64_int64_avx2
  3331                                          # -- End function
  3332  	.ident	"Ubuntu clang version 11.0.0-2~ubuntu20.04.1"
  3333  	.section	".note.GNU-stack","",@progbits
  3334  	.addrsig