From 2682b2720068a0f126135abd6724534ba3eca242 Mon Sep 17 00:00:00 2001 From: Sunil K Pandey Date: Mon, 7 Mar 2022 10:47:13 -0800 Subject: x86_64: Fix svml_d_hypot8_core_avx512.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 --- .../fpu/multiarch/svml_d_hypot8_core_avx512.S | 286 ++++++++++----------- 1 file changed, 142 insertions(+), 144 deletions(-) diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_hypot8_core_avx512.S b/sysdeps/x86_64/fpu/multiarch/svml_d_hypot8_core_avx512.S index 32d9ef30b4..0c404fd5ee 100644 --- a/sysdeps/x86_64/fpu/multiarch/svml_d_hypot8_core_avx512.S +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_hypot8_core_avx512.S @@ -63,173 +63,171 @@ /* Offsets for data table __svml_dhypot_data_internal */ -#define _dAbsMask 0 -#define _lExpBound_uisa 64 -#define _lExpBound 128 -#define _dHalf 192 +#define _dAbsMask 0 +#define _lExpBound_uisa 64 +#define _lExpBound 128 +#define _dHalf 192 #include - .text - .section .text.evex512,"ax",@progbits + .section .text.evex512, "ax", @progbits ENTRY(_ZGVeN8vv_hypot_skx) - pushq %rbp - cfi_def_cfa_offset(16) - movq %rsp, %rbp - cfi_def_cfa(6, 16) - cfi_offset(6, -16) - andq $-64, %rsp - subq $256, %rsp - vgetexppd {sae}, %zmm0, %zmm2 - vgetexppd {sae}, %zmm1, %zmm3 - vmovups _dHalf+__svml_dhypot_data_internal(%rip), %zmm9 - vmaxpd {sae}, %zmm3, %zmm2, %zmm4 - vmulpd {rn-sae}, %zmm0, %zmm0, %zmm2 - vandpd _dAbsMask+__svml_dhypot_data_internal(%rip), %zmm4, %zmm5 - vfmadd231pd {rn-sae}, %zmm1, %zmm1, %zmm2 - -/* Select exponent bound so that no scaling is needed */ - vpcmpq $5, _lExpBound_uisa+__svml_dhypot_data_internal(%rip), %zmm5, %k0 - vrsqrt14pd %zmm2, %zmm6 - kmovw %k0, %edx - vmulpd {rn-sae}, %zmm6, %zmm2, %zmm7 - vmulpd {rn-sae}, %zmm6, %zmm9, %zmm8 - vfnmadd231pd {rn-sae}, %zmm7, %zmm8, %zmm9 - vfmadd231pd {rn-sae}, %zmm9, %zmm8, %zmm8 - vfmadd213pd {rn-sae}, %zmm7, %zmm7, %zmm9 - vfnmadd231pd {rn-sae}, %zmm9, %zmm9, %zmm2 - vfmadd213pd {rn-sae}, %zmm9, %zmm8, %zmm2 - -/* The end of implementation */ - testl %edx, %edx - -/* Go to special inputs processing branch */ - jne L(SPECIAL_VALUES_BRANCH) - # LOE rbx r12 r13 r14 r15 edx zmm0 zmm1 zmm2 - -/* Restore registers - * and exit the function - */ + pushq %rbp + cfi_def_cfa_offset(16) + movq %rsp, %rbp + cfi_def_cfa(6, 16) + cfi_offset(6, -16) + andq $-64, %rsp + subq $256, %rsp + vgetexppd {sae}, %zmm0, %zmm2 + vgetexppd {sae}, %zmm1, %zmm3 + vmovups _dHalf+__svml_dhypot_data_internal(%rip), %zmm9 + vmaxpd {sae}, %zmm3, %zmm2, %zmm4 + vmulpd {rn-sae}, %zmm0, %zmm0, %zmm2 + vandpd _dAbsMask+__svml_dhypot_data_internal(%rip), %zmm4, %zmm5 + vfmadd231pd {rn-sae}, %zmm1, %zmm1, %zmm2 + + /* Select exponent bound so that no scaling is needed */ + vpcmpq $5, _lExpBound_uisa+__svml_dhypot_data_internal(%rip), %zmm5, %k0 + vrsqrt14pd %zmm2, %zmm6 + kmovw %k0, %edx + vmulpd {rn-sae}, %zmm6, %zmm2, %zmm7 + vmulpd {rn-sae}, %zmm6, %zmm9, %zmm8 + vfnmadd231pd {rn-sae}, %zmm7, %zmm8, %zmm9 + vfmadd231pd {rn-sae}, %zmm9, %zmm8, %zmm8 + vfmadd213pd {rn-sae}, %zmm7, %zmm7, %zmm9 + vfnmadd231pd {rn-sae}, %zmm9, %zmm9, %zmm2 + vfmadd213pd {rn-sae}, %zmm9, %zmm8, %zmm2 + + /* The end of implementation */ + testl %edx, %edx + + /* Go to special inputs processing branch */ + jne L(SPECIAL_VALUES_BRANCH) + # LOE rbx r12 r13 r14 r15 edx zmm0 zmm1 zmm2 + + /* Restore registers + * and exit the function + */ L(EXIT): - vmovaps %zmm2, %zmm0 - movq %rbp, %rsp - popq %rbp - cfi_def_cfa(7, 8) - cfi_restore(6) - ret - cfi_def_cfa(6, 16) - cfi_offset(6, -16) - -/* Branch to process - * special inputs - */ + vmovaps %zmm2, %zmm0 + movq %rbp, %rsp + popq %rbp + cfi_def_cfa(7, 8) + cfi_restore(6) + ret + cfi_def_cfa(6, 16) + cfi_offset(6, -16) + + /* Branch to process + * special inputs + */ L(SPECIAL_VALUES_BRANCH): - vmovups %zmm0, 64(%rsp) - vmovups %zmm1, 128(%rsp) - vmovups %zmm2, 192(%rsp) - # LOE rbx r12 r13 r14 r15 edx zmm2 - - xorl %eax, %eax - # LOE rbx r12 r13 r14 r15 eax edx - - vzeroupper - movq %r12, 16(%rsp) - /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */ - .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22 - movl %eax, %r12d - movq %r13, 8(%rsp) - /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */ - .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22 - movl %edx, %r13d - movq %r14, (%rsp) - /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */ - .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22 - # LOE rbx r15 r12d r13d - -/* Range mask - * bits check - */ + vmovups %zmm0, 64(%rsp) + vmovups %zmm1, 128(%rsp) + vmovups %zmm2, 192(%rsp) + # LOE rbx r12 r13 r14 r15 edx zmm2 + + xorl %eax, %eax + # LOE rbx r12 r13 r14 r15 eax edx + + vzeroupper + movq %r12, 16(%rsp) + /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22 + movl %eax, %r12d + movq %r13, 8(%rsp) + /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22 + movl %edx, %r13d + movq %r14, (%rsp) + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r15 r12d r13d + + /* Range mask + * bits check + */ L(RANGEMASK_CHECK): - btl %r12d, %r13d + btl %r12d, %r13d -/* Call scalar math function */ - jc L(SCALAR_MATH_CALL) - # LOE rbx r15 r12d r13d + /* Call scalar math function */ + jc L(SCALAR_MATH_CALL) + # LOE rbx r15 r12d r13d -/* Special inputs - * processing loop - */ + /* Special inputs + * processing loop + */ L(SPECIAL_VALUES_LOOP): - incl %r12d - cmpl $8, %r12d - -/* Check bits in range mask */ - jl L(RANGEMASK_CHECK) - # LOE rbx r15 r12d r13d - - movq 16(%rsp), %r12 - cfi_restore(12) - movq 8(%rsp), %r13 - cfi_restore(13) - movq (%rsp), %r14 - cfi_restore(14) - vmovups 192(%rsp), %zmm2 - -/* Go to exit */ - jmp L(EXIT) - /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */ - .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22 - /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */ - .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22 - /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */ - .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22 - # LOE rbx r12 r13 r14 r15 zmm2 - -/* Scalar math fucntion call - * to process special input - */ + incl %r12d + cmpl $8, %r12d + + /* Check bits in range mask */ + jl L(RANGEMASK_CHECK) + # LOE rbx r15 r12d r13d + + movq 16(%rsp), %r12 + cfi_restore(12) + movq 8(%rsp), %r13 + cfi_restore(13) + movq (%rsp), %r14 + cfi_restore(14) + vmovups 192(%rsp), %zmm2 + + /* Go to exit */ + jmp L(EXIT) + /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22 + /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22 + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r12 r13 r14 r15 zmm2 + + /* Scalar math fucntion call + * to process special input + */ L(SCALAR_MATH_CALL): - movl %r12d, %r14d - movsd 64(%rsp,%r14,8), %xmm0 - movsd 128(%rsp,%r14,8), %xmm1 - call hypot@PLT - # LOE rbx r14 r15 r12d r13d xmm0 + movl %r12d, %r14d + movsd 64(%rsp, %r14, 8), %xmm0 + movsd 128(%rsp, %r14, 8), %xmm1 + call hypot@PLT + # LOE rbx r14 r15 r12d r13d xmm0 - movsd %xmm0, 192(%rsp,%r14,8) + movsd %xmm0, 192(%rsp, %r14, 8) -/* Process special inputs in loop */ - jmp L(SPECIAL_VALUES_LOOP) - # LOE rbx r15 r12d r13d + /* Process special inputs in loop */ + jmp L(SPECIAL_VALUES_LOOP) + # LOE rbx r15 r12d r13d END(_ZGVeN8vv_hypot_skx) - .section .rodata, "a" - .align 64 + .section .rodata, "a" + .align 64 #ifdef __svml_dhypot_data_internal_typedef typedef unsigned int VUINT32; -typedef struct -{ - __declspec(align(64)) VUINT32 _dAbsMask[8][2]; - __declspec(align(64)) VUINT32 _lExpBound_uisa[8][2]; - __declspec(align(64)) VUINT32 _lExpBound[8][2]; - __declspec(align(64)) VUINT32 _dHalf[8][2]; +typedef struct { + __declspec(align(64)) VUINT32 _dAbsMask[8][2]; + __declspec(align(64)) VUINT32 _lExpBound_uisa[8][2]; + __declspec(align(64)) VUINT32 _lExpBound[8][2]; + __declspec(align(64)) VUINT32 _dHalf[8][2]; } __svml_dhypot_data_internal; #endif __svml_dhypot_data_internal: - /* legacy algorithm */ - .quad 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff /* _dAbsMask */ - /* fma based algorithm*/ - .align 64 - .quad 0x407ff00000000000, 0x407ff00000000000, 0x407ff00000000000, 0x407ff00000000000, 0x407ff00000000000, 0x407ff00000000000, 0x407ff00000000000, 0x407ff00000000000 /* _lExpBound_uisa */ - .align 64 - .quad 0x404f800000000000, 0x404f800000000000, 0x404f800000000000, 0x404f800000000000, 0x404f800000000000, 0x404f800000000000, 0x404f800000000000, 0x404f800000000000 /* _lExpBound */ - .align 64 - .quad 0x3FE0000000000000, 0x3FE0000000000000, 0x3FE0000000000000, 0x3FE0000000000000, 0x3FE0000000000000, 0x3FE0000000000000, 0x3FE0000000000000, 0x3FE0000000000000 /* _dHalf */ - .align 64 - .type __svml_dhypot_data_internal,@object - .size __svml_dhypot_data_internal,.-__svml_dhypot_data_internal + /* legacy algorithm */ + .quad 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff /* _dAbsMask */ + /* fma based algorithm*/ + .align 64 + .quad 0x407ff00000000000, 0x407ff00000000000, 0x407ff00000000000, 0x407ff00000000000, 0x407ff00000000000, 0x407ff00000000000, 0x407ff00000000000, 0x407ff00000000000 /* _lExpBound_uisa */ + .align 64 + .quad 0x404f800000000000, 0x404f800000000000, 0x404f800000000000, 0x404f800000000000, 0x404f800000000000, 0x404f800000000000, 0x404f800000000000, 0x404f800000000000 /* _lExpBound */ + .align 64 + .quad 0x3FE0000000000000, 0x3FE0000000000000, 0x3FE0000000000000, 0x3FE0000000000000, 0x3FE0000000000000, 0x3FE0000000000000, 0x3FE0000000000000, 0x3FE0000000000000 /* _dHalf */ + .align 64 + .type __svml_dhypot_data_internal, @object + .size __svml_dhypot_data_internal, .-__svml_dhypot_data_internal -- cgit v1.2.1