{"id":2225440,"url":"http://patchwork.ozlabs.org/api/patches/2225440/?format=json","web_url":"http://patchwork.ozlabs.org/project/gcc/patch/orzf2wuczy.fsf_-_@lxoliva.fsfla.org/","project":{"id":17,"url":"http://patchwork.ozlabs.org/api/projects/17/?format=json","name":"GNU Compiler Collection","link_name":"gcc","list_id":"gcc-patches.gcc.gnu.org","list_email":"gcc-patches@gcc.gnu.org","web_url":null,"scm_url":null,"webscm_url":null,"list_archive_url":"","list_archive_url_format":"","commit_url_format":""},"msgid":"","list_archive_url":null,"date":"2026-04-21T04:57:21","name":"libstdc++: simd: x86: accept 64-bit long double as double [PR124657]","commit_ref":null,"pull_url":null,"state":"new","archived":false,"hash":"a897165457dc4a05d1243881119020b4604eb5d7","submitter":{"id":74937,"url":"http://patchwork.ozlabs.org/api/people/74937/?format=json","name":"Alexandre Oliva","email":"oliva@adacore.com"},"delegate":null,"mbox":"http://patchwork.ozlabs.org/project/gcc/patch/orzf2wuczy.fsf_-_@lxoliva.fsfla.org/mbox/","series":[{"id":500728,"url":"http://patchwork.ozlabs.org/api/series/500728/?format=json","web_url":"http://patchwork.ozlabs.org/project/gcc/list/?series=500728","date":"2026-04-21T04:57:21","name":"libstdc++: simd: x86: accept 64-bit long double as double [PR124657]","version":1,"mbox":"http://patchwork.ozlabs.org/series/500728/mbox/"}],"comments":"http://patchwork.ozlabs.org/api/patches/2225440/comments/","check":"pending","checks":"http://patchwork.ozlabs.org/api/patches/2225440/checks/","tags":{},"related":[],"headers":{"Return-Path":"","X-Original-To":["incoming@patchwork.ozlabs.org","gcc-patches@gcc.gnu.org"],"Delivered-To":["patchwork-incoming@legolas.ozlabs.org","gcc-patches@gcc.gnu.org"],"Authentication-Results":["legolas.ozlabs.org;\n\tdkim=pass (2048-bit key;\n secure) header.d=adacore.com header.i=@adacore.com header.a=rsa-sha256\n header.s=google header.b=Z1J2k3Ag;\n\tdkim-atps=neutral","legolas.ozlabs.org;\n spf=pass (sender SPF authorized) smtp.mailfrom=gcc.gnu.org\n (client-ip=2620:52:6:3111::32; helo=vm01.sourceware.org;\n envelope-from=gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org;\n receiver=patchwork.ozlabs.org)","sourceware.org;\n\tdkim=pass (2048-bit key,\n secure) header.d=adacore.com header.i=@adacore.com header.a=rsa-sha256\n header.s=google header.b=Z1J2k3Ag","sourceware.org; dmarc=pass (p=quarantine dis=none)\n header.from=adacore.com","sourceware.org; spf=pass smtp.mailfrom=adacore.com","server2.sourceware.org;\n arc=none smtp.remote-ip=74.125.82.53"],"Received":["from vm01.sourceware.org (vm01.sourceware.org\n [IPv6:2620:52:6:3111::32])\n\t(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)\n\t key-exchange x25519 server-signature ECDSA (secp384r1) server-digest SHA384)\n\t(No client certificate requested)\n\tby legolas.ozlabs.org (Postfix) with ESMTPS id 4g099l4YjGz1yGt\n\tfor ; Tue, 21 Apr 2026 14:58:50 +1000 (AEST)","from vm01.sourceware.org (localhost [127.0.0.1])\n\tby sourceware.org (Postfix) with ESMTP id 39CED4BA9006\n\tfor ; Tue, 21 Apr 2026 04:58:48 +0000 (GMT)","from mail-dl1-f53.google.com (mail-dl1-f53.google.com\n [74.125.82.53])\n by sourceware.org (Postfix) with ESMTPS id C14274BA23F1\n for ; Tue, 21 Apr 2026 04:57:42 +0000 (GMT)","by mail-dl1-f53.google.com with SMTP id\n a92af1059eb24-12c7212836bso10021508c88.0\n for ; Mon, 20 Apr 2026 21:57:42 -0700 (PDT)","from free.home ([2804:14c:4d1:41a6::2000])\n by smtp.gmail.com with ESMTPSA id\n a92af1059eb24-12c74a185a8sm18047833c88.9.2026.04.20.21.57.39\n (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256);\n Mon, 20 Apr 2026 21:57:40 -0700 (PDT)","from livre (livre.home [172.31.160.2])\n by free.home (8.15.2/8.15.2) with ESMTPS id 63L4vLTO003675\n (version=TLSv1.3 cipher=TLS_AES_256_GCM_SHA384 bits=256 verify=NOT);\n Tue, 21 Apr 2026 01:57:22 -0300"],"DKIM-Filter":["OpenDKIM Filter v2.11.0 sourceware.org 39CED4BA9006","OpenDKIM Filter v2.11.0 sourceware.org C14274BA23F1"],"DMARC-Filter":"OpenDMARC Filter v1.4.2 sourceware.org C14274BA23F1","ARC-Filter":"OpenARC Filter v1.0.0 sourceware.org C14274BA23F1","ARC-Seal":"i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1776747463; cv=none;\n b=BNwjLQMS/w9GKg9PxgB/H8HezLZ3Gb2Db3LmqQyEb75AOmrALaQkPHYhRuiIdx1dpLkVQ/KOrSeg5sNncR3XCUoV+n+zNIOeBpDsy1FyMf9n0EN9V+ZVo/Q+B7C84o1q+Zco+HGuyIx5OhlUM7xHwsYyvjuAIA1eegXnE6eyjAI=","ARC-Message-Signature":"i=1; a=rsa-sha256; d=sourceware.org; s=key;\n t=1776747463; c=relaxed/simple;\n bh=Dapm5X37p1I8m5TXyXZnQ42cIOoiZtSavo1pUswboPw=;\n h=DKIM-Signature:From:To:Subject:Date:Message-ID:MIME-Version;\n b=AFrAyyxM8nwthqU6PrefWFTrDYJUYV2u0ZWonRUIq/SaGBdOm2Hd/eZ4xiT1Eqxpx7HwEORvR62mWdlZ52no5Lvjmfg0YEkvqv+lWGFhWSdSJCk2aeboyTDkkFLfzMx+2YyluUGLcfg5GnUDrCTKk7ZmigzXxl0M7UOQKPLDx/s=","ARC-Authentication-Results":"i=1; server2.sourceware.org","DKIM-Signature":"v=1; a=rsa-sha256; c=relaxed/relaxed;\n d=adacore.com; s=google; t=1776747462; x=1777352262; darn=gcc.gnu.org;\n h=mime-version:user-agent:message-id:in-reply-to:date:references\n :organization:subject:cc:to:from:from:to:cc:subject:date:message-id\n :reply-to; bh=qtIaVxFscPqs4JWuGK2q89uRssUGIYkuvwsCioS1MNQ=;\n b=Z1J2k3Ag/xuLTlhonMRc9NbNo84OcqS5OL7e0/+N29jnb/EDaEMtebyI72/OEr8i4K\n 0kw37T/GvQZZKhmGLvFkTQJfNtsiE0L8u4GefyPCfcZ2Rz8sj5RvTUYTRcn23pVuPITK\n V4zqsFqgjjxfIxkm2zUn/1rKG8BEFq3BCBhbTneAzxMxdFQ7uw06T+oTOOFk0lAyKtSq\n rWf0B3iHX4DBuV3ZVnavyZEf3KPasuhssXqU3PP0zGcOZ5USTo4dXGXGbnehFCOB7kTX\n qT0FZu92QbLC63XEdycr8nyU4mTq5n8UQoVGK+oI2I3JjmryfoGROiwfhP9YcHE/FOxk\n jJ1A==","X-Google-DKIM-Signature":"v=1; a=rsa-sha256; c=relaxed/relaxed;\n d=1e100.net; s=20251104; t=1776747462; x=1777352262;\n h=mime-version:user-agent:message-id:in-reply-to:date:references\n :organization:subject:cc:to:from:x-gm-gg:x-gm-message-state:from:to\n :cc:subject:date:message-id:reply-to;\n bh=qtIaVxFscPqs4JWuGK2q89uRssUGIYkuvwsCioS1MNQ=;\n b=NOFmxCMuA0q6ZElc1lkl1AoliumRyvJnkWcyhEbCj2gJebfEw6uIaMCJSA3GsuC6wg\n 9JLaBykMWGRr0Q5ddjioGjVarMjHJDMls9ZR5cF4DeNRAeE6u7+0tQqahb64Rf10vO2e\n cDAJaSvjd/wCsWc94guYNcMRZhPsc5Nrplq+mr5NJuZnbigTDTCzrpUceiiATHPyRpd7\n FYMirE/kxOC/+35X2AGQYE0r+PxYs0mefK/9OLBml38gdkbhv9RVT5nYNUqSRPm95feE\n sEArYJLSkM8px8xjpdbulqxQoyCeWY7PyVHxRCediV7YZR0i8Nh5jxijw8XPcqSJrozx\n rRMw==","X-Forwarded-Encrypted":"i=1;\n AFNElJ8VU7X9whXEVR0nzpeKGDtAlBqsXtZEQ0DWDM9unoqEwjMXaLfQA455UhkhChtJLjgOSc4kQh/aLsjGVw==@gcc.gnu.org","X-Gm-Message-State":"AOJu0YwjwubmrHKF8yqKdUGKbnmKs4QAthFiDLK+/oKffG1UgDrZ45ie\n mNo8/igmIHhNkJp/vX96m4BDDmCiamYbqOS2z08lyJZyjJif7+7gW8n8iFm0dCuzlQ==","X-Gm-Gg":"AeBDiet+C6JZiyy/5vyMYxH4AcpLEUd7dJjnjKWTSG+UO9VLiP+nnmGE/LaIG53+hD3\n v97kbUlOJ1oCKeoaK/gKPoNnV4hYn48H1ycqbJyCEBY0YBWLYJwZM5mH+qqvEj9bycDbHTZhAxG\n rfeGOp3g5PBvKMF+KTQsq0EicHaeq74ZUpCDd3zNcM+Jovo+i7bnrkHk+t0Avc4LLDQcu01gZnq\n gB3HvxM7/+eG80S7rQXB48RvL5Ibr8yDRTTlSSxL4QZCBRaPlKNn1ndMuOkK6eZ9v1rzwzTsBm7\n qWDKtowqtWYSc9UsvV4NqA2dzzsWXfzblqUyJ5YNjaZy26iuUVt4k/RioGkBP/9U8P1D+i+7E3g\n /4oE7rgwjSEvGDobD3CJdIhVvN0vkj+suNIuI+EaVArAZnTzUxlNoYJlLIhgBDS/s7pTFVPWKQT\n enIaTK7jk5cQ==","X-Received":"by 2002:a05:7022:69a:b0:12c:4928:e57f with SMTP id\n a92af1059eb24-12c73fa3e3cmr7904647c88.25.1776747461444;\n Mon, 20 Apr 2026 21:57:41 -0700 (PDT)","From":"Alexandre Oliva ","To":"Matthias Kretz ","Cc":"libstdc++@gcc.gnu.org, Jonathan Wakely ,\n gcc-patches ","Subject":"[PATCH] libstdc++: simd: x86: accept 64-bit long double as double\n [PR124657]","Organization":"Free thinker, does not speak for AdaCore","References":"\n \n \n <47992676.fMDQidcC6G@vir-laptop> ","Date":"Tue, 21 Apr 2026 01:57:21 -0300","In-Reply-To":" (Alexandre Oliva's message\n of \"Thu, 09 Apr 2026 19:14:35 -0300\")","Message-ID":"","User-Agent":"Gnus/5.13 (Gnus v5.13) Emacs/27.1 (gnu/linux)","MIME-Version":"1.0","Content-Type":"text/plain","X-Scanned-By":"MIMEDefang 2.84","X-BeenThere":"gcc-patches@gcc.gnu.org","X-Mailman-Version":"2.1.30","Precedence":"list","List-Id":"Gcc-patches mailing list ","List-Unsubscribe":",\n ","List-Archive":"","List-Post":"","List-Help":"","List-Subscribe":",\n ","Errors-To":"gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org"},"content":"On Apr 9, 2026, Alexandre Oliva wrote:\n\n> We probably need a better abstraction.\n\nHere's one:\n\nVarious simd_x86 functions that handle double need to be adjusted to\nmatch 64-bit long double as well.\n\nIntroduce __is_x86_ps<_Tp>() and __is_x86_pd<_Tp>() and use them\ninstead of is_same_v<_Tp, float> and is_same_v<_Tp, double>,\nrespectively.\n\nRegstrapped on x86_64-linux-gnu. Also tested pr109261_constexpr_simd.cc\nwith -mlong-double-64. Ok to install? (presumably for stage1)\n\n\nfor libstdc++-v3/ChangeLog\n\n\tPR libstdc++/124657\n\t* include/experimental/bits/simd_x86.h\n\t(__is_x86_ps<_Tp>): New. Replace is_same_v<_Tp, float> with it.\n\t(__is_x86_pd<_Tp>): New. Replace is_same_v<_Tp, double> with it.\n---\n libstdc++-v3/include/experimental/bits/simd_x86.h | 79 ++++++++++++---------\n 1 file changed, 47 insertions(+), 32 deletions(-)","diff":"diff --git a/libstdc++-v3/include/experimental/bits/simd_x86.h b/libstdc++-v3/include/experimental/bits/simd_x86.h\nindex 74c7a61998e0d..73c10c1dcbbe7 100644\n--- a/libstdc++-v3/include/experimental/bits/simd_x86.h\n+++ b/libstdc++-v3/include/experimental/bits/simd_x86.h\n@@ -415,13 +415,28 @@ template \n #endif\n \n // ISA & type detection {{{\n+template \n+ constexpr bool\n+ __is_x86_ps()\n+ {\n+ return is_same_v<_Tp, float>;\n+ }\n+\n+template \n+ constexpr bool\n+ __is_x86_pd()\n+ {\n+ return is_same_v<_Tp, double>\n+ || (sizeof (long double) == sizeof (double) && is_same_v<_Tp, long double>);\n+ }\n+\n template \n constexpr bool\n __is_sse_ps()\n {\n return __have_sse\n-\t && is_same_v<_Tp,\n-\t\t\tfloat> && sizeof(__intrinsic_type_t<_Tp, _Np>) == 16;\n+\t && __is_x86_ps<_Tp>()\n+\t && sizeof(__intrinsic_type_t<_Tp, _Np>) == 16;\n }\n \n template \n@@ -429,8 +444,8 @@ template \n __is_sse_pd()\n {\n return __have_sse2\n-\t && is_same_v<_Tp,\n-\t\t\tdouble> && sizeof(__intrinsic_type_t<_Tp, _Np>) == 16;\n+\t && __is_x86_pd<_Tp>()\n+\t && sizeof(__intrinsic_type_t<_Tp, _Np>) == 16;\n }\n \n template \n@@ -438,8 +453,8 @@ template \n __is_avx_ps()\n {\n return __have_avx\n-\t && is_same_v<_Tp,\n-\t\t\tfloat> && sizeof(__intrinsic_type_t<_Tp, _Np>) == 32;\n+\t && __is_x86_ps<_Tp>()\n+\t && sizeof(__intrinsic_type_t<_Tp, _Np>) == 32;\n }\n \n template \n@@ -447,8 +462,8 @@ template \n __is_avx_pd()\n {\n return __have_avx\n-\t && is_same_v<_Tp,\n-\t\t\tdouble> && sizeof(__intrinsic_type_t<_Tp, _Np>) == 32;\n+\t && __is_x86_pd<_Tp>()\n+\t && sizeof(__intrinsic_type_t<_Tp, _Np>) == 32;\n }\n \n template \n@@ -456,8 +471,8 @@ template \n __is_avx512_ps()\n {\n return __have_avx512f\n-\t && is_same_v<_Tp,\n-\t\t\tfloat> && sizeof(__intrinsic_type_t<_Tp, _Np>) == 64;\n+\t && __is_x86_ps<_Tp>()\n+\t && sizeof(__intrinsic_type_t<_Tp, _Np>) == 64;\n }\n \n template \n@@ -465,8 +480,8 @@ template \n __is_avx512_pd()\n {\n return __have_avx512f\n-\t && is_same_v<_Tp,\n-\t\t\tdouble> && sizeof(__intrinsic_type_t<_Tp, _Np>) == 64;\n+\t && __is_x86_pd<_Tp>()\n+\t && sizeof(__intrinsic_type_t<_Tp, _Np>) == 64;\n }\n \n // }}}\n@@ -2397,9 +2412,9 @@ template \n \t [[maybe_unused]] const auto __yi = __to_intrin(__y);\n \t if constexpr (sizeof(__xi) == 64)\n \t {\n-\t\tif constexpr (is_same_v<_Tp, float>)\n+\t\tif constexpr (__is_x86_ps<_Tp> ())\n \t\t return _mm512_mask_cmp_ps_mask(__k1, __xi, __yi, _CMP_LT_OS);\n-\t\telse if constexpr (is_same_v<_Tp, double>)\n+\t\telse if constexpr (__is_x86_pd<_Tp> ())\n \t\t return _mm512_mask_cmp_pd_mask(__k1, __xi, __yi, _CMP_LT_OS);\n \t\telse if constexpr (is_signed_v<_Tp> && sizeof(_Tp) == 1)\n \t\t return _mm512_mask_cmplt_epi8_mask(__k1, __xi, __yi);\n@@ -2422,9 +2437,9 @@ template \n \t }\n \t else if constexpr (sizeof(__xi) == 32)\n \t {\n-\t\tif constexpr (is_same_v<_Tp, float>)\n+\t\tif constexpr (__is_x86_ps<_Tp> ())\n \t\t return _mm256_mask_cmp_ps_mask(__k1, __xi, __yi, _CMP_LT_OS);\n-\t\telse if constexpr (is_same_v<_Tp, double>)\n+\t\telse if constexpr (__is_x86_pd<_Tp> ())\n \t\t return _mm256_mask_cmp_pd_mask(__k1, __xi, __yi, _CMP_LT_OS);\n \t\telse if constexpr (is_signed_v<_Tp> && sizeof(_Tp) == 1)\n \t\t return _mm256_mask_cmplt_epi8_mask(__k1, __xi, __yi);\n@@ -2447,9 +2462,9 @@ template \n \t }\n \t else if constexpr (sizeof(__xi) == 16)\n \t {\n-\t\tif constexpr (is_same_v<_Tp, float>)\n+\t\tif constexpr (__is_x86_ps<_Tp> ())\n \t\t return _mm_mask_cmp_ps_mask(__k1, __xi, __yi, _CMP_LT_OS);\n-\t\telse if constexpr (is_same_v<_Tp, double>)\n+\t\telse if constexpr (__is_x86_pd<_Tp> ())\n \t\t return _mm_mask_cmp_pd_mask(__k1, __xi, __yi, _CMP_LT_OS);\n \t\telse if constexpr (is_signed_v<_Tp> && sizeof(_Tp) == 1)\n \t\t return _mm_mask_cmplt_epi8_mask(__k1, __xi, __yi);\n@@ -2505,9 +2520,9 @@ template \n \t [[maybe_unused]] const auto __yi = __to_intrin(__y);\n \t if constexpr (sizeof(__xi) == 64)\n \t {\n-\t\tif constexpr (is_same_v<_Tp, float>)\n+\t\tif constexpr (__is_x86_ps<_Tp> ())\n \t\t return _mm512_mask_cmp_ps_mask(__k1, __xi, __yi, _CMP_LE_OS);\n-\t\telse if constexpr (is_same_v<_Tp, double>)\n+\t\telse if constexpr (__is_x86_pd<_Tp> ())\n \t\t return _mm512_mask_cmp_pd_mask(__k1, __xi, __yi, _CMP_LE_OS);\n \t\telse if constexpr (is_signed_v<_Tp> && sizeof(_Tp) == 1)\n \t\t return _mm512_mask_cmple_epi8_mask(__k1, __xi, __yi);\n@@ -2530,9 +2545,9 @@ template \n \t }\n \t else if constexpr (sizeof(__xi) == 32)\n \t {\n-\t\tif constexpr (is_same_v<_Tp, float>)\n+\t\tif constexpr (__is_x86_ps<_Tp> ())\n \t\t return _mm256_mask_cmp_ps_mask(__k1, __xi, __yi, _CMP_LE_OS);\n-\t\telse if constexpr (is_same_v<_Tp, double>)\n+\t\telse if constexpr (__is_x86_pd<_Tp> ())\n \t\t return _mm256_mask_cmp_pd_mask(__k1, __xi, __yi, _CMP_LE_OS);\n \t\telse if constexpr (is_signed_v<_Tp> && sizeof(_Tp) == 1)\n \t\t return _mm256_mask_cmple_epi8_mask(__k1, __xi, __yi);\n@@ -2555,9 +2570,9 @@ template \n \t }\n \t else if constexpr (sizeof(__xi) == 16)\n \t {\n-\t\tif constexpr (is_same_v<_Tp, float>)\n+\t\tif constexpr (__is_x86_ps<_Tp> ())\n \t\t return _mm_mask_cmp_ps_mask(__k1, __xi, __yi, _CMP_LE_OS);\n-\t\telse if constexpr (is_same_v<_Tp, double>)\n+\t\telse if constexpr (__is_x86_pd<_Tp> ())\n \t\t return _mm_mask_cmp_pd_mask(__k1, __xi, __yi, _CMP_LE_OS);\n \t\telse if constexpr (is_signed_v<_Tp> && sizeof(_Tp) == 1)\n \t\t return _mm_mask_cmple_epi8_mask(__k1, __xi, __yi);\n@@ -5021,10 +5036,10 @@ template \n \t\t = _Abi::template _S_implicit_mask_intrin<_Tp>();\n \t\treturn 0 != __testc(__a, __b);\n \t }\n-\t else if constexpr (is_same_v<_Tp, float>)\n+\t else if constexpr (__is_x86_ps<_Tp> ())\n \t return (_mm_movemask_ps(__a) & ((1 << _Np) - 1))\n \t\t == (1 << _Np) - 1;\n-\t else if constexpr (is_same_v<_Tp, double>)\n+\t else if constexpr (__is_x86_pd<_Tp> ())\n \t return (_mm_movemask_pd(__a) & ((1 << _Np) - 1))\n \t\t == (1 << _Np) - 1;\n \t else\n@@ -5084,9 +5099,9 @@ template \n \t\telse\n \t\t return 0 == __testz(__a, __a);\n \t }\n-\t else if constexpr (is_same_v<_Tp, float>)\n+\t else if constexpr (__is_x86_ps<_Tp> ())\n \t return (_mm_movemask_ps(__a) & ((1 << _Np) - 1)) != 0;\n-\t else if constexpr (is_same_v<_Tp, double>)\n+\t else if constexpr (__is_x86_pd<_Tp> ())\n \t return (_mm_movemask_pd(__a) & ((1 << _Np) - 1)) != 0;\n \t else\n \t return (_mm_movemask_epi8(__a) & ((1 << (_Np * sizeof(_Tp))) - 1))\n@@ -5120,9 +5135,9 @@ template \n \t\telse\n \t\t return 0 != __testz(__a, __a);\n \t }\n-\t else if constexpr (is_same_v<_Tp, float>)\n+\t else if constexpr (__is_x86_ps<_Tp> ())\n \t return (__movemask(__a) & ((1 << _Np) - 1)) == 0;\n-\t else if constexpr (is_same_v<_Tp, double>)\n+\t else if constexpr (__is_x86_pd<_Tp> ())\n \t return (__movemask(__a) & ((1 << _Np) - 1)) == 0;\n \t else\n \t return (__movemask(__a) & int((1ull << (_Np * sizeof(_Tp))) - 1))\n@@ -5150,13 +5165,13 @@ template \n \t\t = _Abi::template _S_implicit_mask_intrin<_Tp>();\n \t\treturn 0 != __testnzc(__a, __b);\n \t }\n-\t else if constexpr (is_same_v<_Tp, float>)\n+\t else if constexpr (__is_x86_ps<_Tp> ())\n \t {\n \t\tconstexpr int __allbits = (1 << _Np) - 1;\n \t\tconst auto __tmp = _mm_movemask_ps(__a) & __allbits;\n \t\treturn __tmp > 0 && __tmp < __allbits;\n \t }\n-\t else if constexpr (is_same_v<_Tp, double>)\n+\t else if constexpr (__is_x86_pd<_Tp> ())\n \t {\n \t\tconstexpr int __allbits = (1 << _Np) - 1;\n \t\tconst auto __tmp = _mm_movemask_pd(__a) & __allbits;\n","prefixes":[]}