From patchwork Mon Mar 7 15:00:04 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Sunil Pandey X-Patchwork-Id: 51632 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 98E953858404 for ; Mon, 7 Mar 2022 15:11:37 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 98E953858404 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1646665897; bh=SlHkzRhgvhrGYAL1Ub4dfyrbjp4fI+qSjyt+3cCELxk=; h=To:Subject:Date:In-Reply-To:References:List-Id:List-Unsubscribe: List-Archive:List-Post:List-Help:List-Subscribe:From:Reply-To: From; b=hoNbQD02agAuKd3uQ36zhbBTp5GEcqoKT3c3P2RgOUxPVOL1VYBixp8eJu8uLPCPn igdNjHs6QZhxhFlkRiNVc6TjVrBkrS4Ne7XYZZ+wJLkgDYKHWpFlSy2fWewUFv+Cuw WWWLgN1maSLIRQ/hYyjpojtN/09iySFvOtG5vmc4= X-Original-To: libc-alpha@sourceware.org Delivered-To: libc-alpha@sourceware.org Received: from mga09.intel.com (mga09.intel.com [134.134.136.24]) by sourceware.org (Postfix) with ESMTPS id 2DC053858410 for ; Mon, 7 Mar 2022 15:02:10 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org 2DC053858410 X-IronPort-AV: E=McAfee;i="6200,9189,10278"; a="254138345" X-IronPort-AV: E=Sophos;i="5.90,162,1643702400"; d="scan'208";a="254138345" Received: from fmsmga008.fm.intel.com ([10.253.24.58]) by orsmga102.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 07 Mar 2022 07:02:04 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.90,162,1643702400"; d="scan'208";a="595526468" Received: from scymds01.sc.intel.com ([10.148.94.138]) by fmsmga008.fm.intel.com with ESMTP; 07 Mar 2022 07:02:02 -0800 Received: from gskx-1.sc.intel.com (gskx-1.sc.intel.com [172.25.149.211]) by scymds01.sc.intel.com with ESMTP id 227F21dh016772; Mon, 7 Mar 2022 07:02:02 -0800 To: libc-alpha@sourceware.org Subject: [PATCH 009/126] x86_64: Fix svml_s_acoshf8_core_avx2.S code formatting Date: Mon, 7 Mar 2022 07:00:04 -0800 Message-Id: <20220307150201.10590-10-skpgkp2@gmail.com> X-Mailer: git-send-email 2.35.1 In-Reply-To: <20220307150201.10590-1-skpgkp2@gmail.com> References: <20220307150201.10590-1-skpgkp2@gmail.com> MIME-Version: 1.0 X-Spam-Status: No, score=-7.5 required=5.0 tests=BAYES_00, DKIM_ADSP_CUSTOM_MED, FORGED_GMAIL_RCVD, FREEMAIL_ENVFROM_END_DIGIT, FREEMAIL_FROM, GIT_PATCH_0, HK_RANDOM_ENVFROM, HK_RANDOM_FROM, KAM_DMARC_NONE, KAM_DMARC_STATUS, KAM_LOTSOFHASH, NML_ADSP_CUSTOM_MED, RCVD_IN_MSPIKE_H3, RCVD_IN_MSPIKE_WL, SPF_HELO_NONE, SPF_SOFTFAIL, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.4 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on server2.sourceware.org X-BeenThere: libc-alpha@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Libc-alpha mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Sunil K Pandey via Libc-alpha From: Sunil Pandey Reply-To: Sunil K Pandey Errors-To: libc-alpha-bounces+patchwork=sourceware.org@sourceware.org Sender: "Libc-alpha" 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. --- .../fpu/multiarch/svml_s_acoshf8_core_avx2.S | 615 +++++++++--------- 1 file changed, 307 insertions(+), 308 deletions(-) diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_acoshf8_core_avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_s_acoshf8_core_avx2.S index 6746aff576..3a70fc1448 100644 --- a/sysdeps/x86_64/fpu/multiarch/svml_s_acoshf8_core_avx2.S +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_acoshf8_core_avx2.S @@ -33,338 +33,337 @@ /* Offsets for data table __svml_sacosh_data_internal */ -#define sOne 0 -#define sPoly 32 -#define iBrkValue 288 -#define iOffExpoMask 320 -#define sBigThreshold 352 -#define sC2 384 -#define sC3 416 -#define sHalf 448 -#define sLargestFinite 480 -#define sThirtyOne 512 -#define sTopMask8 544 -#define XScale 576 -#define sLn2 608 +#define sOne 0 +#define sPoly 32 +#define iBrkValue 288 +#define iOffExpoMask 320 +#define sBigThreshold 352 +#define sC2 384 +#define sC3 416 +#define sHalf 448 +#define sLargestFinite 480 +#define sThirtyOne 512 +#define sTopMask8 544 +#define XScale 576 +#define sLn2 608 #include - .text - .section .text.avx2,"ax",@progbits + .section .text.avx2, "ax", @progbits ENTRY(_ZGVdN8v_acoshf_avx2) - pushq %rbp - cfi_def_cfa_offset(16) - movq %rsp, %rbp - cfi_def_cfa(6, 16) - cfi_offset(6, -16) - andq $-32, %rsp - subq $96, %rsp - -/* Load constants, always including One = 1 */ - vmovups sOne+__svml_sacosh_data_internal(%rip), %ymm2 - -/* Finally, express Y + W = U * V accurately where Y has <= 8 bits */ - vmovups sTopMask8+__svml_sacosh_data_internal(%rip), %ymm9 - -/* - * Now 1 / (1 + d) - * = 1 / (1 + (sqrt(1 - e) - 1)) - * = 1 / sqrt(1 - e) - * = 1 + 1/2 * e + 3/8 * e^2 + 5/16 * e^3 + 35/128 * e^4 + ... - * So compute the first three nonconstant terms of that, so that - * we have a relative correction (1 + Corr) to apply to S etc. - * C1 = 1/2 - * C2 = 3/8 - * C3 = 5/16 - */ - vmovups sC3+__svml_sacosh_data_internal(%rip), %ymm14 - vmovaps %ymm0, %ymm3 - vmovaps %ymm2, %ymm7 - vfmsub231ps %ymm3, %ymm3, %ymm7 - -/* - * Check that 1 < X < +inf; otherwise go to the callout function. - * We need the callout for X = 1 to avoid division by zero below. - * This test ensures that callout handles NaN and either infinity. - */ - vcmpnle_uqps sLargestFinite+__svml_sacosh_data_internal(%rip), %ymm3, %ymm4 - vcmpngt_uqps %ymm2, %ymm3, %ymm5 - -/* - * The following computation can go wrong for very large X, e.g. - * the X^2 - 1 = U * V can overflow. But for large X we have - * acosh(X) / log(2 X) - 1 =~= 1/(4 * X^2), so for X >= 2^30 - * we can just later stick X back into the log and tweak up the exponent. - * Actually we scale X by 2^-30 and tweak the exponent up by 31, - * to stay in the safe range for the later log computation. - * Compute a flag now telling us when to do this. - */ - vcmplt_oqps sBigThreshold+__svml_sacosh_data_internal(%rip), %ymm3, %ymm1 - vandps %ymm9, %ymm7, %ymm10 - -/* - * Compute R = 1/sqrt(Y + W) * (1 + d) - * Force R to <= 8 significant bits. - * This means that R * Y and R^2 * Y are exactly representable. - */ - vrsqrtps %ymm10, %ymm8 - vsubps %ymm10, %ymm7, %ymm11 - vandps %ymm9, %ymm8, %ymm12 - -/* - * Compute S = (Y/sqrt(Y + W)) * (1 + d) - * and T = (W/sqrt(Y + W)) * (1 + d) - * so that S + T = sqrt(Y + W) * (1 + d) - * S is exact, and the rounding error in T is OK. - */ - vmulps %ymm12, %ymm10, %ymm15 - vmulps %ymm11, %ymm12, %ymm0 - -/* Now multiplex to the case X = 2^-30 * input, Xl = 0 in the "big" case. */ - vmulps XScale+__svml_sacosh_data_internal(%rip), %ymm3, %ymm11 - -/* - * Compute e = -(2 * d + d^2) - * The first FMR is exact, and the rounding error in the other is acceptable - * since d and e are ~ 2^-8 - */ - vmovaps %ymm2, %ymm13 - vfnmadd231ps %ymm15, %ymm12, %ymm13 - vfnmadd231ps %ymm0, %ymm12, %ymm13 - vfmadd213ps sC2+__svml_sacosh_data_internal(%rip), %ymm13, %ymm14 - vfmadd213ps sHalf+__svml_sacosh_data_internal(%rip), %ymm13, %ymm14 - vmulps %ymm14, %ymm13, %ymm7 - vorps %ymm5, %ymm4, %ymm6 - -/* - * For low-accuracy versions, the computation can be done - * just as U + ((S + T) + (S + T) * Corr) - */ - vaddps %ymm0, %ymm15, %ymm5 - -/* sU is needed later on */ - vsubps %ymm2, %ymm3, %ymm4 - vfmadd213ps %ymm5, %ymm7, %ymm5 - vmovmskps %ymm6, %edx - vaddps %ymm5, %ymm4, %ymm6 - -/* - * Now resume the main code. - * reduction: compute r,n - */ - vmovups iBrkValue+__svml_sacosh_data_internal(%rip), %ymm4 - -/* - * Now we feed into the log1p code, using H in place of _VARG1 and - * also adding L into Xl. - * compute 1+x as high, low parts - */ - vmaxps %ymm6, %ymm2, %ymm8 - vminps %ymm6, %ymm2, %ymm9 - vaddps %ymm9, %ymm8, %ymm12 - vblendvps %ymm1, %ymm12, %ymm11, %ymm14 - vsubps %ymm12, %ymm8, %ymm10 - vpsubd %ymm4, %ymm14, %ymm15 - vaddps %ymm10, %ymm9, %ymm13 - vpand iOffExpoMask+__svml_sacosh_data_internal(%rip), %ymm15, %ymm14 - vpsrad $23, %ymm15, %ymm15 - vpaddd %ymm4, %ymm14, %ymm8 - vpslld $23, %ymm15, %ymm5 - vmovups sPoly+224+__svml_sacosh_data_internal(%rip), %ymm4 - vcvtdq2ps %ymm15, %ymm0 - vpsubd %ymm5, %ymm2, %ymm7 - -/* polynomial evaluation */ - vsubps %ymm2, %ymm8, %ymm2 - -/* Add 31 to the exponent in the "large" case to get log(2 * input) */ - vaddps sThirtyOne+__svml_sacosh_data_internal(%rip), %ymm0, %ymm5 - vandps %ymm1, %ymm13, %ymm6 - vmulps %ymm7, %ymm6, %ymm9 - vblendvps %ymm1, %ymm0, %ymm5, %ymm0 - vaddps %ymm2, %ymm9, %ymm2 - vfmadd213ps sPoly+192+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 - vfmadd213ps sPoly+160+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 - vfmadd213ps sPoly+128+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 - vfmadd213ps sPoly+96+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 - vfmadd213ps sPoly+64+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 - vfmadd213ps sPoly+32+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 - vfmadd213ps sPoly+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 - vmulps %ymm4, %ymm2, %ymm6 - vfmadd213ps %ymm2, %ymm2, %ymm6 - -/* final reconstruction */ - vfmadd132ps sLn2+__svml_sacosh_data_internal(%rip), %ymm6, %ymm0 - testl %edx, %edx - -/* Go to special inputs processing branch */ - jne L(SPECIAL_VALUES_BRANCH) - # LOE rbx r12 r13 r14 r15 edx ymm0 ymm3 - -/* 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 $-32, %rsp + subq $96, %rsp + + /* Load constants, always including One = 1 */ + vmovups sOne+__svml_sacosh_data_internal(%rip), %ymm2 + + /* Finally, express Y + W = U * V accurately where Y has <= 8 bits */ + vmovups sTopMask8+__svml_sacosh_data_internal(%rip), %ymm9 + + /* + * Now 1 / (1 + d) + * = 1 / (1 + (sqrt(1 - e) - 1)) + * = 1 / sqrt(1 - e) + * = 1 + 1/2 * e + 3/8 * e^2 + 5/16 * e^3 + 35/128 * e^4 + ... + * So compute the first three nonconstant terms of that, so that + * we have a relative correction (1 + Corr) to apply to S etc. + * C1 = 1/2 + * C2 = 3/8 + * C3 = 5/16 + */ + vmovups sC3+__svml_sacosh_data_internal(%rip), %ymm14 + vmovaps %ymm0, %ymm3 + vmovaps %ymm2, %ymm7 + vfmsub231ps %ymm3, %ymm3, %ymm7 + + /* + * Check that 1 < X < +inf; otherwise go to the callout function. + * We need the callout for X = 1 to avoid division by zero below. + * This test ensures that callout handles NaN and either infinity. + */ + vcmpnle_uqps sLargestFinite+__svml_sacosh_data_internal(%rip), %ymm3, %ymm4 + vcmpngt_uqps %ymm2, %ymm3, %ymm5 + + /* + * The following computation can go wrong for very large X, e.g. + * the X^2 - 1 = U * V can overflow. But for large X we have + * acosh(X) / log(2 X) - 1 =~= 1/(4 * X^2), so for X >= 2^30 + * we can just later stick X back into the log and tweak up the exponent. + * Actually we scale X by 2^-30 and tweak the exponent up by 31, + * to stay in the safe range for the later log computation. + * Compute a flag now telling us when to do this. + */ + vcmplt_oqps sBigThreshold+__svml_sacosh_data_internal(%rip), %ymm3, %ymm1 + vandps %ymm9, %ymm7, %ymm10 + + /* + * Compute R = 1/sqrt(Y + W) * (1 + d) + * Force R to <= 8 significant bits. + * This means that R * Y and R^2 * Y are exactly representable. + */ + vrsqrtps %ymm10, %ymm8 + vsubps %ymm10, %ymm7, %ymm11 + vandps %ymm9, %ymm8, %ymm12 + + /* + * Compute S = (Y/sqrt(Y + W)) * (1 + d) + * and T = (W/sqrt(Y + W)) * (1 + d) + * so that S + T = sqrt(Y + W) * (1 + d) + * S is exact, and the rounding error in T is OK. + */ + vmulps %ymm12, %ymm10, %ymm15 + vmulps %ymm11, %ymm12, %ymm0 + + /* Now multiplex to the case X = 2^-30 * input, Xl = 0 in the "big" case. */ + vmulps XScale+__svml_sacosh_data_internal(%rip), %ymm3, %ymm11 + + /* + * Compute e = -(2 * d + d^2) + * The first FMR is exact, and the rounding error in the other is acceptable + * since d and e are ~ 2^-8 + */ + vmovaps %ymm2, %ymm13 + vfnmadd231ps %ymm15, %ymm12, %ymm13 + vfnmadd231ps %ymm0, %ymm12, %ymm13 + vfmadd213ps sC2+__svml_sacosh_data_internal(%rip), %ymm13, %ymm14 + vfmadd213ps sHalf+__svml_sacosh_data_internal(%rip), %ymm13, %ymm14 + vmulps %ymm14, %ymm13, %ymm7 + vorps %ymm5, %ymm4, %ymm6 + + /* + * For low-accuracy versions, the computation can be done + * just as U + ((S + T) + (S + T) * Corr) + */ + vaddps %ymm0, %ymm15, %ymm5 + + /* sU is needed later on */ + vsubps %ymm2, %ymm3, %ymm4 + vfmadd213ps %ymm5, %ymm7, %ymm5 + vmovmskps %ymm6, %edx + vaddps %ymm5, %ymm4, %ymm6 + + /* + * Now resume the main code. + * reduction: compute r, n + */ + vmovups iBrkValue+__svml_sacosh_data_internal(%rip), %ymm4 + + /* + * Now we feed into the log1p code, using H in place of _VARG1 and + * also adding L into Xl. + * compute 1+x as high, low parts + */ + vmaxps %ymm6, %ymm2, %ymm8 + vminps %ymm6, %ymm2, %ymm9 + vaddps %ymm9, %ymm8, %ymm12 + vblendvps %ymm1, %ymm12, %ymm11, %ymm14 + vsubps %ymm12, %ymm8, %ymm10 + vpsubd %ymm4, %ymm14, %ymm15 + vaddps %ymm10, %ymm9, %ymm13 + vpand iOffExpoMask+__svml_sacosh_data_internal(%rip), %ymm15, %ymm14 + vpsrad $23, %ymm15, %ymm15 + vpaddd %ymm4, %ymm14, %ymm8 + vpslld $23, %ymm15, %ymm5 + vmovups sPoly+224+__svml_sacosh_data_internal(%rip), %ymm4 + vcvtdq2ps %ymm15, %ymm0 + vpsubd %ymm5, %ymm2, %ymm7 + + /* polynomial evaluation */ + vsubps %ymm2, %ymm8, %ymm2 + + /* Add 31 to the exponent in the "large" case to get log(2 * input) */ + vaddps sThirtyOne+__svml_sacosh_data_internal(%rip), %ymm0, %ymm5 + vandps %ymm1, %ymm13, %ymm6 + vmulps %ymm7, %ymm6, %ymm9 + vblendvps %ymm1, %ymm0, %ymm5, %ymm0 + vaddps %ymm2, %ymm9, %ymm2 + vfmadd213ps sPoly+192+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 + vfmadd213ps sPoly+160+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 + vfmadd213ps sPoly+128+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 + vfmadd213ps sPoly+96+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 + vfmadd213ps sPoly+64+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 + vfmadd213ps sPoly+32+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 + vfmadd213ps sPoly+__svml_sacosh_data_internal(%rip), %ymm2, %ymm4 + vmulps %ymm4, %ymm2, %ymm6 + vfmadd213ps %ymm2, %ymm2, %ymm6 + + /* final reconstruction */ + vfmadd132ps sLn2+__svml_sacosh_data_internal(%rip), %ymm6, %ymm0 + testl %edx, %edx + + /* Go to special inputs processing branch */ + jne L(SPECIAL_VALUES_BRANCH) + # LOE rbx r12 r13 r14 r15 edx ymm0 ymm3 + + /* Restore registers + * and exit the function + */ L(EXIT): - 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 - */ + 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 %ymm3, 32(%rsp) - vmovups %ymm0, 64(%rsp) - # LOE rbx r12 r13 r14 r15 edx ymm0 - - 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: -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 - movl %eax, %r12d - movq %r13, 8(%rsp) - /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -88; DW_OP_plus) */ - .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa8, 0xff, 0xff, 0xff, 0x22 - movl %edx, %r13d - movq %r14, (%rsp) - /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -96; DW_OP_plus) */ - .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 - * bits check - */ + vmovups %ymm3, 32(%rsp) + vmovups %ymm0, 64(%rsp) + # LOE rbx r12 r13 r14 r15 edx ymm0 + + 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: -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 + movl %eax, %r12d + movq %r13, 8(%rsp) + /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -88; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa8, 0xff, 0xff, 0xff, 0x22 + movl %edx, %r13d + movq %r14, (%rsp) + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -96; DW_OP_plus) */ + .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 + * 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 64(%rsp), %ymm0 - -/* 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 - /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -88; DW_OP_plus) */ - .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa8, 0xff, 0xff, 0xff, 0x22 - /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -96; DW_OP_plus) */ - .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 - * 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 64(%rsp), %ymm0 + + /* 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 + /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -88; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa8, 0xff, 0xff, 0xff, 0x22 + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -96; DW_OP_plus) */ + .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 + * to process special input + */ L(SCALAR_MATH_CALL): - movl %r12d, %r14d - movss 32(%rsp,%r14,4), %xmm0 - call acoshf@PLT - # LOE rbx r14 r15 r12d r13d xmm0 + movl %r12d, %r14d + movss 32(%rsp, %r14, 4), %xmm0 + call acoshf@PLT + # LOE rbx r14 r15 r12d r13d xmm0 - movss %xmm0, 64(%rsp,%r14,4) + movss %xmm0, 64(%rsp, %r14, 4) -/* 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(_ZGVdN8v_acoshf_avx2) - .section .rodata, "a" - .align 32 + .section .rodata, "a" + .align 32 #ifdef __svml_sacosh_data_internal_typedef typedef unsigned int VUINT32; typedef struct { - __declspec(align(32)) VUINT32 sOne[8][1]; - __declspec(align(32)) VUINT32 sPoly[8][8][1]; - __declspec(align(32)) VUINT32 iBrkValue[8][1]; - __declspec(align(32)) VUINT32 iOffExpoMask[8][1]; - __declspec(align(32)) VUINT32 sBigThreshold[8][1]; - __declspec(align(32)) VUINT32 sC2[8][1]; - __declspec(align(32)) VUINT32 sC3[8][1]; - __declspec(align(32)) VUINT32 sHalf[8][1]; - __declspec(align(32)) VUINT32 sLargestFinite[8][1]; - __declspec(align(32)) VUINT32 sThirtyOne[8][1]; - __declspec(align(32)) VUINT32 sTopMask8[8][1]; - __declspec(align(32)) VUINT32 XScale[8][1]; - __declspec(align(32)) VUINT32 sLn2[8][1]; + __declspec(align(32)) VUINT32 sOne[8][1]; + __declspec(align(32)) VUINT32 sPoly[8][8][1]; + __declspec(align(32)) VUINT32 iBrkValue[8][1]; + __declspec(align(32)) VUINT32 iOffExpoMask[8][1]; + __declspec(align(32)) VUINT32 sBigThreshold[8][1]; + __declspec(align(32)) VUINT32 sC2[8][1]; + __declspec(align(32)) VUINT32 sC3[8][1]; + __declspec(align(32)) VUINT32 sHalf[8][1]; + __declspec(align(32)) VUINT32 sLargestFinite[8][1]; + __declspec(align(32)) VUINT32 sThirtyOne[8][1]; + __declspec(align(32)) VUINT32 sTopMask8[8][1]; + __declspec(align(32)) VUINT32 XScale[8][1]; + __declspec(align(32)) VUINT32 sLn2[8][1]; } __svml_sacosh_data_internal; #endif __svml_sacosh_data_internal: - /*== sOne = SP 1.0 ==*/ - .long 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000 - /*== sPoly[] = SP polynomial ==*/ - .align 32 - .long 0xbf000000, 0xbf000000, 0xbf000000, 0xbf000000, 0xbf000000, 0xbf000000, 0xbf000000, 0xbf000000 /* -5.0000000000000000000000000e-01 P0 */ - .long 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94 /* 3.3333265781402587890625000e-01 P1 */ - .long 0xbe80058e, 0xbe80058e, 0xbe80058e, 0xbe80058e, 0xbe80058e, 0xbe80058e, 0xbe80058e, 0xbe80058e /* -2.5004237890243530273437500e-01 P2 */ - .long 0x3e4ce190, 0x3e4ce190, 0x3e4ce190, 0x3e4ce190, 0x3e4ce190, 0x3e4ce190, 0x3e4ce190, 0x3e4ce190 /* 2.0007920265197753906250000e-01 P3 */ - .long 0xbe28ad37, 0xbe28ad37, 0xbe28ad37, 0xbe28ad37, 0xbe28ad37, 0xbe28ad37, 0xbe28ad37, 0xbe28ad37 /* -1.6472326219081878662109375e-01 P4 */ - .long 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12 /* 1.4042308926582336425781250e-01 P5 */ - .long 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3 /* -1.5122179687023162841796875e-01 P6 */ - .long 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed /* 1.3820238411426544189453125e-01 P7 */ - /*== iBrkValue = SP 2/3 ==*/ - .align 32 - .long 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab - /*== iOffExpoMask = SP significand mask ==*/ - .align 32 - .long 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff - /*== sBigThreshold ==*/ - .align 32 - .long 0x4E800000, 0x4E800000, 0x4E800000, 0x4E800000, 0x4E800000, 0x4E800000, 0x4E800000, 0x4E800000 - /*== sC2 ==*/ - .align 32 - .long 0x3EC00000, 0x3EC00000, 0x3EC00000, 0x3EC00000, 0x3EC00000, 0x3EC00000, 0x3EC00000, 0x3EC00000 - /*== sC3 ==*/ - .align 32 - .long 0x3EA00000, 0x3EA00000, 0x3EA00000, 0x3EA00000, 0x3EA00000, 0x3EA00000, 0x3EA00000, 0x3EA00000 - /*== sHalf ==*/ - .align 32 - .long 0x3F000000, 0x3F000000, 0x3F000000, 0x3F000000, 0x3F000000, 0x3F000000, 0x3F000000, 0x3F000000 - /*== sLargestFinite ==*/ - .align 32 - .long 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF - /*== sThirtyOne ==*/ - .align 32 - .long 0x41F80000, 0x41F80000, 0x41F80000, 0x41F80000, 0x41F80000, 0x41F80000, 0x41F80000, 0x41F80000 - /*== sTopMask8 ==*/ - .align 32 - .long 0xFFFF0000, 0xFFFF0000, 0xFFFF0000, 0xFFFF0000, 0xFFFF0000, 0xFFFF0000, 0xFFFF0000, 0xFFFF0000 - /*== XScale ==*/ - .align 32 - .long 0x30800000, 0x30800000, 0x30800000, 0x30800000, 0x30800000, 0x30800000, 0x30800000, 0x30800000 - /*== sLn2 = SP ln(2) ==*/ - .align 32 - .long 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218 - .align 32 - .type __svml_sacosh_data_internal,@object - .size __svml_sacosh_data_internal,.-__svml_sacosh_data_internal + /* sOne = SP 1.0 */ + .long 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000 + /* sPoly[] = SP polynomial */ + .align 32 + .long 0xbf000000, 0xbf000000, 0xbf000000, 0xbf000000, 0xbf000000, 0xbf000000, 0xbf000000, 0xbf000000 /* -5.0000000000000000000000000e-01 P0 */ + .long 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94 /* 3.3333265781402587890625000e-01 P1 */ + .long 0xbe80058e, 0xbe80058e, 0xbe80058e, 0xbe80058e, 0xbe80058e, 0xbe80058e, 0xbe80058e, 0xbe80058e /* -2.5004237890243530273437500e-01 P2 */ + .long 0x3e4ce190, 0x3e4ce190, 0x3e4ce190, 0x3e4ce190, 0x3e4ce190, 0x3e4ce190, 0x3e4ce190, 0x3e4ce190 /* 2.0007920265197753906250000e-01 P3 */ + .long 0xbe28ad37, 0xbe28ad37, 0xbe28ad37, 0xbe28ad37, 0xbe28ad37, 0xbe28ad37, 0xbe28ad37, 0xbe28ad37 /* -1.6472326219081878662109375e-01 P4 */ + .long 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12 /* 1.4042308926582336425781250e-01 P5 */ + .long 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3 /* -1.5122179687023162841796875e-01 P6 */ + .long 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed /* 1.3820238411426544189453125e-01 P7 */ + /* iBrkValue = SP 2/3 */ + .align 32 + .long 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab + /* iOffExpoMask = SP significand mask */ + .align 32 + .long 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff + /* sBigThreshold */ + .align 32 + .long 0x4E800000, 0x4E800000, 0x4E800000, 0x4E800000, 0x4E800000, 0x4E800000, 0x4E800000, 0x4E800000 + /* sC2 */ + .align 32 + .long 0x3EC00000, 0x3EC00000, 0x3EC00000, 0x3EC00000, 0x3EC00000, 0x3EC00000, 0x3EC00000, 0x3EC00000 + /* sC3 */ + .align 32 + .long 0x3EA00000, 0x3EA00000, 0x3EA00000, 0x3EA00000, 0x3EA00000, 0x3EA00000, 0x3EA00000, 0x3EA00000 + /* sHalf */ + .align 32 + .long 0x3F000000, 0x3F000000, 0x3F000000, 0x3F000000, 0x3F000000, 0x3F000000, 0x3F000000, 0x3F000000 + /* sLargestFinite */ + .align 32 + .long 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF + /* sThirtyOne */ + .align 32 + .long 0x41F80000, 0x41F80000, 0x41F80000, 0x41F80000, 0x41F80000, 0x41F80000, 0x41F80000, 0x41F80000 + /* sTopMask8 */ + .align 32 + .long 0xFFFF0000, 0xFFFF0000, 0xFFFF0000, 0xFFFF0000, 0xFFFF0000, 0xFFFF0000, 0xFFFF0000, 0xFFFF0000 + /* XScale */ + .align 32 + .long 0x30800000, 0x30800000, 0x30800000, 0x30800000, 0x30800000, 0x30800000, 0x30800000, 0x30800000 + /* sLn2 = SP ln(2) */ + .align 32 + .long 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218 + .align 32 + .type __svml_sacosh_data_internal, @object + .size __svml_sacosh_data_internal, .-__svml_sacosh_data_internal