x86_64: Fix svml_d_log24_core_avx2.S code formatting

This commit contains following formatting changes

1. Instructions proceeded by a tab.
2. Instruction less than 8 characters in length have a tab
   between it and the first operand.
3. Instruction greater than 7 characters in length have a
   space between it and the first operand.
4. Tabs after `#define`d names and their value.
5. 8 space at the beginning of line replaced by tab.
6. Indent comments with code.
7. Remove redundent .text section.
8. 1 space between line content and line comment.
9. Space after all commas.

Reviewed-by: Noah Goldstein <goldstein.w.n@gmail.com>
This commit is contained in:
Sunil K Pandey 2022-03-07 10:47:14 -08:00
parent 2d4d1dc9ab
commit 26b648892a

View File

@ -47,8 +47,7 @@
#include <sysdep.h>
.text
.section .text.avx2,"ax",@progbits
.section .text.avx2, "ax", @progbits
ENTRY(_ZGVdN4v_log2_avx2)
pushq %rbp
cfi_def_cfa_offset(16)
@ -60,51 +59,51 @@ ENTRY(_ZGVdN4v_log2_avx2)
lea Table_Lookup_Bias+__svml_dlog2_data_internal(%rip), %r8
vmovapd %ymm0, %ymm3
/* preserve mantissa, set input exponent to 2^(-10) */
/* preserve mantissa, set input exponent to 2^(-10) */
vandpd ExpMask+__svml_dlog2_data_internal(%rip), %ymm3, %ymm4
vorpd Two10+__svml_dlog2_data_internal(%rip), %ymm4, %ymm2
/* reciprocal approximation good to at least 11 bits */
/* reciprocal approximation good to at least 11 bits */
vcvtpd2ps %ymm2, %xmm5
/* exponent bits */
/* exponent bits */
vpsrlq $20, %ymm3, %ymm7
vmovupd One+__svml_dlog2_data_internal(%rip), %ymm14
vrcpps %xmm5, %xmm6
/* check range */
/* check range */
vcmplt_oqpd MinNorm+__svml_dlog2_data_internal(%rip), %ymm3, %ymm11
vcmpnle_uqpd MaxNorm+__svml_dlog2_data_internal(%rip), %ymm3, %ymm12
vcvtps2pd %xmm6, %ymm9
/* round reciprocal to nearest integer, will have 1+9 mantissa bits */
/* round reciprocal to nearest integer, will have 1+9 mantissa bits */
vroundpd $0, %ymm9, %ymm1
/* exponent */
/* exponent */
vmovupd Threshold+__svml_dlog2_data_internal(%rip), %ymm9
/*
/*
* prepare table index
* table lookup
*/
vpsrlq $40, %ymm1, %ymm15
/* argument reduction */
/* argument reduction */
vfmsub213pd %ymm14, %ymm1, %ymm2
/* polynomial */
/* polynomial */
vmovupd poly_coeff+__svml_dlog2_data_internal(%rip), %ymm14
vcmplt_oqpd %ymm1, %ymm9, %ymm1
vfmadd213pd poly_coeff+32+__svml_dlog2_data_internal(%rip), %ymm2, %ymm14
vorpd %ymm12, %ymm11, %ymm13
vmulpd %ymm2, %ymm2, %ymm12
/* combine and get argument value range mask */
/* combine and get argument value range mask */
vmovmskpd %ymm13, %eax
vextractf128 $1, %ymm7, %xmm8
vshufps $221, %xmm8, %xmm7, %xmm10
/* biased exponent in DP format */
/* biased exponent in DP format */
vcvtdq2pd %xmm10, %ymm0
vandpd Bias+__svml_dlog2_data_internal(%rip), %ymm1, %ymm10
vorpd Bias1+__svml_dlog2_data_internal(%rip), %ymm10, %ymm11
@ -123,22 +122,22 @@ ENTRY(_ZGVdN4v_log2_avx2)
vpextrd $2, %xmm6, %edi
movslq %ecx, %rcx
movslq %edi, %rdi
vmovsd (%r8,%rdx), %xmm4
vmovsd (%r8,%rsi), %xmm7
vmovhpd (%r8,%rcx), %xmm4, %xmm5
vmovhpd (%r8,%rdi), %xmm7, %xmm8
vmovsd (%r8, %rdx), %xmm4
vmovsd (%r8, %rsi), %xmm7
vmovhpd (%r8, %rcx), %xmm4, %xmm5
vmovhpd (%r8, %rdi), %xmm7, %xmm8
vinsertf128 $1, %xmm8, %ymm5, %ymm13
/* reconstruction */
/* reconstruction */
vaddpd %ymm14, %ymm13, %ymm0
vaddpd %ymm0, %ymm1, %ymm0
testl %eax, %eax
/* Go to special inputs processing branch */
/* Go to special inputs processing branch */
jne L(SPECIAL_VALUES_BRANCH)
# LOE rbx r12 r13 r14 r15 eax ymm0 ymm3
/* Restore registers
/* Restore registers
* and exit the function
*/
@ -151,7 +150,7 @@ L(EXIT):
cfi_def_cfa(6, 16)
cfi_offset(6, -16)
/* Branch to process
/* Branch to process
* special inputs
*/
@ -177,18 +176,18 @@ L(SPECIAL_VALUES_BRANCH):
.cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa0, 0xff, 0xff, 0xff, 0x22
# LOE rbx r15 r12d r13d
/* Range mask
/* Range mask
* bits check
*/
L(RANGEMASK_CHECK):
btl %r12d, %r13d
/* Call scalar math function */
/* Call scalar math function */
jc L(SCALAR_MATH_CALL)
# LOE rbx r15 r12d r13d
/* Special inputs
/* Special inputs
* processing loop
*/
@ -196,7 +195,7 @@ L(SPECIAL_VALUES_LOOP):
incl %r12d
cmpl $4, %r12d
/* Check bits in range mask */
/* Check bits in range mask */
jl L(RANGEMASK_CHECK)
# LOE rbx r15 r12d r13d
@ -208,7 +207,7 @@ L(SPECIAL_VALUES_LOOP):
cfi_restore(14)
vmovupd 64(%rsp), %ymm0
/* Go to exit */
/* Go to exit */
jmp L(EXIT)
/* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -80; DW_OP_plus) */
.cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xb0, 0xff, 0xff, 0xff, 0x22
@ -218,19 +217,19 @@ L(SPECIAL_VALUES_LOOP):
.cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa0, 0xff, 0xff, 0xff, 0x22
# LOE rbx r12 r13 r14 r15 ymm0
/* Scalar math fucntion call
/* Scalar math fucntion call
* to process special input
*/
L(SCALAR_MATH_CALL):
movl %r12d, %r14d
movsd 32(%rsp,%r14,8), %xmm0
movsd 32(%rsp, %r14, 8), %xmm0
call log2@PLT
# LOE rbx r14 r15 r12d r13d xmm0
movsd %xmm0, 64(%rsp,%r14,8)
movsd %xmm0, 64(%rsp, %r14, 8)
/* Process special inputs in loop */
/* Process special inputs in loop */
jmp L(SPECIAL_VALUES_LOOP)
# LOE rbx r15 r12d r13d
END(_ZGVdN4v_log2_avx2)
@ -770,7 +769,7 @@ __svml_dlog2_data_internal:
.quad 0xc08ff7f7fb62e230, 0x3d26abb2c547789a
.quad 0xc08ff7fbfe63b010, 0xbd59d383d543b3f5
.quad 0xc08ff80000000000, 0x8000000000000000
/*== Log_LA_table ==*/
/* Log_LA_table */
.align 32
.quad 0x0000000000000000
.quad 0xbf670f83ff0a7565
@ -1285,40 +1284,40 @@ __svml_dlog2_data_internal:
.quad 0x3f671b0ea42e5fda
.quad 0x3f57182a894b69c6
.quad 0x8000000000000000
/*== poly_coeff[5] ==*/
/* poly_coeff[5] */
.align 32
.quad 0x3fd2776E996DA1D2, 0x3fd2776E996DA1D2, 0x3fd2776E996DA1D2, 0x3fd2776E996DA1D2 /* coeff5 */
.quad 0xbfd715494C3E7C9B, 0xbfd715494C3E7C9B, 0xbfd715494C3E7C9B, 0xbfd715494C3E7C9B /* coeff4 */
.quad 0x3fdEC709DC39E926, 0x3fdEC709DC39E926, 0x3fdEC709DC39E926, 0x3fdEC709DC39E926 /* coeff3 */
.quad 0xbfe71547652B7CF8, 0xbfe71547652B7CF8, 0xbfe71547652B7CF8, 0xbfe71547652B7CF8 /* coeff2 */
.quad 0x3ff71547652B82FE, 0x3ff71547652B82FE, 0x3ff71547652B82FE, 0x3ff71547652B82FE /* coeff1 */
/*== ExpMask ==*/
/* ExpMask */
.align 32
.quad 0x000fffffffffffff, 0x000fffffffffffff, 0x000fffffffffffff, 0x000fffffffffffff
/*== Two10 ==*/
/* Two10 */
.align 32
.quad 0x3f50000000000000, 0x3f50000000000000, 0x3f50000000000000, 0x3f50000000000000
/*== MinNorm ==*/
/* MinNorm */
.align 32
.quad 0x0010000000000000, 0x0010000000000000, 0x0010000000000000, 0x0010000000000000
/*== MaxNorm ==*/
/* MaxNorm */
.align 32
.quad 0x7fefffffffffffff, 0x7fefffffffffffff, 0x7fefffffffffffff, 0x7fefffffffffffff
/*== HalfMask ==*/
/* HalfMask */
.align 32
.quad 0xfffffffffc000000, 0xfffffffffc000000, 0xfffffffffc000000, 0xfffffffffc000000
/*== One ==*/
/* One */
.align 32
.quad 0x3ff0000000000000, 0x3ff0000000000000, 0x3ff0000000000000, 0x3ff0000000000000
/*== Threshold ==*/
/* Threshold */
.align 32
.quad 0x4086a00000000000, 0x4086a00000000000, 0x4086a00000000000, 0x4086a00000000000
/*== Bias ==*/
/* Bias */
.align 32
.quad 0x408ff80000000000, 0x408ff80000000000, 0x408ff80000000000, 0x408ff80000000000
/*== Bias1 ==*/
/* Bias1 */
.align 32
.quad 0x408ff00000000000, 0x408ff00000000000, 0x408ff00000000000, 0x408ff00000000000
.align 32
.type __svml_dlog2_data_internal,@object
.size __svml_dlog2_data_internal,.-__svml_dlog2_data_internal
.type __svml_dlog2_data_internal, @object
.size __svml_dlog2_data_internal, .-__svml_dlog2_data_internal