From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 1039) id 07FA2385840F; Wed, 5 Apr 2023 22:45:25 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 07FA2385840F DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1680734725; bh=Uh9kRBW4bn7N5Jt94PY7mxCdzE5AEcPXOP/xPrgITHk=; h=From:To:Subject:Date:From; b=MDT61C7CKlfP1jhw0AmK2ACJhwg33rJk+56B7R2V034/p6TbGVgb1rCtODzPn3fiU m11cEqb9yEhyw40nHUczCHWO8nKeK9+WZk5C4Ih8kM9eNOokWpDOVJioNDiryPXyFt XDl+hs0rKOLXCRPvQDBLkqpt/N14q/8W2ClWZapM= Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit From: H.J. Lu To: glibc-cvs@sourceware.org Subject: [glibc] : Add AMX-COMPLEX support X-Act-Checkin: glibc X-Git-Author: H.J. Lu X-Git-Refname: refs/heads/master X-Git-Oldrev: 609b7b2d3c0c44f4e42a235f8d7c207c903a6bb1 X-Git-Newrev: b05521c916790dbaeb953fca7b979a889d868db2 Message-Id: <20230405224525.07FA2385840F@sourceware.org> Date: Wed, 5 Apr 2023 22:45:25 +0000 (GMT) List-Id: https://sourceware.org/git/gitweb.cgi?p=glibc.git;h=b05521c916790dbaeb953fca7b979a889d868db2 commit b05521c916790dbaeb953fca7b979a889d868db2 Author: H.J. Lu Date: Wed Apr 5 09:21:43 2023 -0700 : Add AMX-COMPLEX support Add AMX-COMPLEX support to . Reviewed-by: Noah Goldstein Diff: --- manual/platform.texi | 3 +++ sysdeps/x86/bits/platform/x86.h | 1 + sysdeps/x86/cpu-features.c | 2 ++ sysdeps/x86/include/cpu-features.h | 3 +++ sysdeps/x86/tst-get-cpu-features.c | 2 ++ 5 files changed, 11 insertions(+) diff --git a/manual/platform.texi b/manual/platform.texi index 1e120993d7..e7448ffc1a 100644 --- a/manual/platform.texi +++ b/manual/platform.texi @@ -197,6 +197,9 @@ The supported processor features are: @item @code{AMX_BF16} -- Tile computational operations on bfloat16 numbers. +@item +@code{AMX_COMPLEX} -- Tile computational operations on complex FP16 numbers. + @item @code{AMX_INT8} -- Tile computational operations on 8-bit numbers. diff --git a/sysdeps/x86/bits/platform/x86.h b/sysdeps/x86/bits/platform/x86.h index d8ba33bd42..96eb4c070d 100644 --- a/sysdeps/x86/bits/platform/x86.h +++ b/sysdeps/x86/bits/platform/x86.h @@ -310,6 +310,7 @@ enum x86_cpu_AVX_VNNI_INT8 = x86_cpu_index_7_ecx_1_edx + 4, x86_cpu_AVX_NE_CONVERT = x86_cpu_index_7_ecx_1_edx + 5, + x86_cpu_AMX_COMPLEX = x86_cpu_index_7_ecx_1_edx + 8, x86_cpu_index_19_ebx = (CPUID_INDEX_19 * 8 * 4 * sizeof (unsigned int) diff --git a/sysdeps/x86/cpu-features.c b/sysdeps/x86/cpu-features.c index dfd1b85dce..c2bea6a32d 100644 --- a/sysdeps/x86/cpu-features.c +++ b/sysdeps/x86/cpu-features.c @@ -221,6 +221,8 @@ update_active (struct cpu_features *cpu_features) CPU_FEATURE_SET_ACTIVE (cpu_features, AMX_INT8); /* Determine if AMX_FP16 is usable. */ CPU_FEATURE_SET_ACTIVE (cpu_features, AMX_FP16); + /* Determine if AMX_COMPLEX is usable. */ + CPU_FEATURE_SET_ACTIVE (cpu_features, AMX_COMPLEX); } /* These features are usable only when OSXSAVE is enabled. */ diff --git a/sysdeps/x86/include/cpu-features.h b/sysdeps/x86/include/cpu-features.h index 673cf8ca92..f14c1078d5 100644 --- a/sysdeps/x86/include/cpu-features.h +++ b/sysdeps/x86/include/cpu-features.h @@ -317,6 +317,7 @@ enum /* EDX. */ #define bit_cpu_AVX_VNNI_INT8 (1u << 4) #define bit_cpu_AVX_NE_CONVERT (1u << 5) +#define bit_cpu_AMX_COMPLEX (1u << 8) /* CPUID_INDEX_19. */ @@ -558,6 +559,7 @@ enum #define index_cpu_LAM CPUID_INDEX_7_ECX_1 #define index_cpu_AVX_VNNI_INT8 CPUID_INDEX_7_ECX_1 #define index_cpu_AVX_NE_CONVERT CPUID_INDEX_7_ECX_1 +#define index_cpu_AMX_COMPLEX CPUID_INDEX_7_ECX_1 /* CPUID_INDEX_19. */ @@ -801,6 +803,7 @@ enum /* EDX. */ #define reg_AVX_VNNI_INT8 edx #define reg_AVX_NE_CONVERT edx +#define reg_AMX_COMPLEX edx /* CPUID_INDEX_19. */ diff --git a/sysdeps/x86/tst-get-cpu-features.c b/sysdeps/x86/tst-get-cpu-features.c index bb1b67fd1c..87fe27340f 100644 --- a/sysdeps/x86/tst-get-cpu-features.c +++ b/sysdeps/x86/tst-get-cpu-features.c @@ -217,6 +217,7 @@ do_test (void) CHECK_CPU_FEATURE_PRESENT (MSRLIST); CHECK_CPU_FEATURE_PRESENT (AVX_VNNI_INT8); CHECK_CPU_FEATURE_PRESENT (AVX_NE_CONVERT); + CHECK_CPU_FEATURE_PRESENT (AMX_COMPLEX); CHECK_CPU_FEATURE_PRESENT (AESKLE); CHECK_CPU_FEATURE_PRESENT (WIDE_KL); CHECK_CPU_FEATURE_PRESENT (PTWRITE); @@ -386,6 +387,7 @@ do_test (void) CHECK_CPU_FEATURE_ACTIVE (AVX_IFMA); CHECK_CPU_FEATURE_ACTIVE (AVX_VNNI_INT8); CHECK_CPU_FEATURE_ACTIVE (AVX_NE_CONVERT); + CHECK_CPU_FEATURE_ACTIVE (AMX_COMPLEX); CHECK_CPU_FEATURE_ACTIVE (AESKLE); CHECK_CPU_FEATURE_ACTIVE (WIDE_KL); CHECK_CPU_FEATURE_ACTIVE (PTWRITE);