Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions clang/docs/ReleaseNotes.rst
Original file line number Diff line number Diff line change
Expand Up @@ -580,6 +580,8 @@ Android Support

Windows Support
^^^^^^^^^^^^^^^
- clang-cl now supports /arch:AVX10.1 and /arch:AVX10.2.
- clang-cl now supports /vlen, /vlen=256 and /vlen=512.

LoongArch Support
^^^^^^^^^^^^^^^^^
Expand Down
2 changes: 2 additions & 0 deletions clang/lib/Driver/ToolChains/Arch/X86.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -42,6 +42,8 @@ std::string x86::getX86TargetCPU(const Driver &D, const ArgList &Args,
{"AVX2", "haswell"},
{"AVX512F", "knl"},
{"AVX512", "skylake-avx512"},
{"AVX10.1", "sapphirerapids"},
{"AVX10.2", "diamondrapids"},
});
if (Triple.getArch() == llvm::Triple::x86) {
// 32-bit-only /arch: flags.
Expand Down
17 changes: 12 additions & 5 deletions clang/lib/Driver/ToolChains/Clang.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -8272,22 +8272,29 @@ void Clang::AddClangCLArgs(const ArgList &Args, types::ID InputType,
llvm::Triple::ArchType AT = getToolChain().getArch();
StringRef Default = AT == llvm::Triple::x86 ? "IA32" : "SSE2";
StringRef Arch = Args.getLastArgValue(options::OPT__SLASH_arch, Default);
llvm::SmallSet<StringRef, 4> Arch512 = {"AVX512F", "AVX512", "AVX10.1",
"AVX10.2"};

if (A->getOption().matches(options::OPT__SLASH_vlen_EQ_512)) {
if (Arch == "AVX512F" || Arch == "AVX512")
if (Arch512.contains(Arch))
CmdArgs.push_back("-mprefer-vector-width=512");
else
D.Diag(diag::warn_drv_argument_not_allowed_with)
<< "/vlen=512" << std::string("/arch:").append(Arch);
}

if (A->getOption().matches(options::OPT__SLASH_vlen_EQ_256)) {
if (Arch == "AVX512F" || Arch == "AVX512")
} else if (A->getOption().matches(options::OPT__SLASH_vlen_EQ_256)) {
if (Arch512.contains(Arch))
CmdArgs.push_back("-mprefer-vector-width=256");
else if (Arch != "AVX" && Arch != "AVX2")
D.Diag(diag::warn_drv_argument_not_allowed_with)
<< "/vlen=256" << std::string("/arch:").append(Arch);
} else {
if (Arch == "AVX10.1" || Arch == "AVX10.2")
CmdArgs.push_back("-mprefer-vector-width=256");
}
} else {
StringRef Arch = Args.getLastArgValue(options::OPT__SLASH_arch);
if (Arch == "AVX10.1" || Arch == "AVX10.2")
CmdArgs.push_back("-mprefer-vector-width=256");
}

Arg *MostGeneralArg = Args.getLastArg(options::OPT__SLASH_vmg);
Expand Down
60 changes: 57 additions & 3 deletions clang/test/Driver/cl-x86-flags.c
Original file line number Diff line number Diff line change
Expand Up @@ -18,10 +18,10 @@

// arch: args are case-sensitive.
// RUN: %clang_cl -m32 -arch:ia32 --target=i386-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=ia32 %s
// ia32: invalid /arch: argument 'ia32'; for 32-bit expected one of AVX, AVX2, AVX512, AVX512F, IA32, SSE, SSE2
// ia32: invalid /arch: argument 'ia32'; for 32-bit expected one of AVX, AVX10.1, AVX10.2, AVX2, AVX512, AVX512F, IA32, SSE, SSE2

// RUN: %clang_cl -m64 -arch:IA32 --target=x86_64-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=IA3264 %s
// IA3264: invalid /arch: argument 'IA32'; for 64-bit expected one of AVX, AVX2, AVX512, AVX512F
// IA3264: invalid /arch: argument 'IA32'; for 64-bit expected one of AVX, AVX10.1, AVX10.2, AVX2, AVX512, AVX512F

// RUN: %clang_cl -m32 -arch:SSE --target=i386-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_32_ARCH_SSE -- %s
#if defined(TEST_32_ARCH_SSE)
Expand All @@ -44,7 +44,7 @@
// sse2: invalid /arch: argument

// RUN: %clang_cl -m64 -arch:SSE --target=x86_64-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=SSE64 %s
// SSE64: invalid /arch: argument 'SSE'; for 64-bit expected one of AVX, AVX2, AVX512, AVX512F
// SSE64: invalid /arch: argument 'SSE'; for 64-bit expected one of AVX, AVX10.1, AVX10.2, AVX2, AVX512, AVX512F

// RUN: %clang_cl -m64 -arch:SSE2 --target=x86_64-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=SSE264 %s
// SSE264: invalid /arch: argument
Expand Down Expand Up @@ -80,6 +80,8 @@
// avx512f: invalid /arch: argument

// RUN: %clang_cl -m32 -arch:AVX512 --target=i386-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_32_ARCH_AVX512 -- %s
// RUN: %clang_cl -m32 -arch:AVX10.1 --target=i386-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_32_ARCH_AVX512 -- %s
// RUN: %clang_cl -m32 -arch:AVX10.2 --target=i386-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_32_ARCH_AVX512 -- %s
#if defined(TEST_32_ARCH_AVX512)
#if _M_IX86_FP != 2 || !__AVX__ || !__AVX2__ || !__AVX512F__ || !__AVX512BW__
#error fail
Expand All @@ -89,6 +91,28 @@
// RUN: %clang_cl -m32 -arch:avx512 --target=i386-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=avx512 %s
// avx512: invalid /arch: argument

// RUN: %clang_cl -m32 -arch:AVX10.1 --target=i386-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_32_ARCH_AVX10_1_ADD -- %s
// RUN: %clang_cl -m32 -arch:AVX10.2 --target=i386-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_32_ARCH_AVX10_1_ADD -- %s
#if defined(TEST_32_ARCH_AVX10_1_ADD)
#if !__AVX512VL__ || !__AVX512CD__ || !__AVX512DQ__ || !__AVX512VBMI__ || !__AVX512IFMA__ || !__AVX512VNNI__ ||\
!__AVX512BF16__ || !__AVX512VPOPCNTDQ__ || !__AVX512VBMI2__ || !__AVX512VPOPCNTDQ__ || !__AVX512BITALG__ || !__AVX512FP16__
#error fail
#endif
#endif

// RUN: %clang_cl -m32 -arch:avx10.1 --target=i386-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=avx10_1 %s
// avx10_1: invalid /arch: argument

// RUN: %clang_cl -m32 -arch:AVX10.2 --target=i386-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_32_ARCH_AVX10_2_ADD -- %s
#if defined(TEST_32_ARCH_AVX10_2_ADD)
#if !__AVX10_2__
#error fail
#endif
#endif

// RUN: %clang_cl -m32 -arch:avx10.2 --target=i386-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=avx10_2 %s
// avx10_2: invalid /arch: argument

// RUN: %clang_cl -m64 -arch:AVX --target=x86_64-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_64_ARCH_AVX -- %s
#if defined(TEST_64_ARCH_AVX)
#if _M_IX86_FP || !__AVX__ || __AVX2__ || __AVX512F__ || __AVX512BW__
Expand Down Expand Up @@ -120,6 +144,8 @@
// avx512f64: invalid /arch: argument

// RUN: %clang_cl -m64 -arch:AVX512 --target=i386-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_64_ARCH_AVX512 -- %s
// RUN: %clang_cl -m64 -arch:AVX10.1 --target=i386-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_64_ARCH_AVX512 -- %s
// RUN: %clang_cl -m64 -arch:AVX10.2 --target=i386-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_64_ARCH_AVX512 -- %s
#if defined(TEST_64_ARCH_AVX512)
#if _M_IX86_FP || !__AVX__ || !__AVX2__ || !__AVX512F__ || !__AVX512BW__
#error fail
Expand All @@ -129,14 +155,42 @@
// RUN: %clang_cl -m64 -arch:avx512 --target=i386-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=avx51264 %s
// avx51264: invalid /arch: argument

// RUN: %clang_cl -m64 -arch:AVX10.1 --target=i386-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_64_ARCH_AVX10_1_ADD -- %s
// RUN: %clang_cl -m64 -arch:AVX10.2 --target=i386-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_64_ARCH_AVX10_1_ADD -- %s
#if defined(TEST_64_ARCH_AVX10_1_ADD)
#if !__AVX512VL__ || !__AVX512CD__ || !__AVX512DQ__ || !__AVX512VBMI__ || !__AVX512IFMA__ || !__AVX512VNNI__ ||\
!__AVX512BF16__ || !__AVX512VPOPCNTDQ__ || !__AVX512VBMI2__ || !__AVX512VPOPCNTDQ__ || !__AVX512BITALG__ || !__AVX512FP16__
#error fail
#endif
#endif

// RUN: %clang_cl -m64 -arch:avx10.1 --target=i386-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=avx10_1_64 %s
// avx10_1_64: invalid /arch: argument

// RUN: %clang_cl -m64 -arch:AVX10.2 --target=i386-pc-windows /c /Fo%t.obj -Xclang -verify -DTEST_64_ARCH_AVX10_2_ADD -- %s
#if defined(TEST_64_ARCH_AVX10_2_ADD)
#if !__AVX10_2__
#error fail
#endif
#endif

// RUN: %clang_cl -m64 -arch:avx10.2 --target=i386-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=avx10_2_64 %s
// avx10_2_64: invalid /arch: argument

// RUN: %clang_cl -m64 -arch:AVX -tune:haswell --target=x86_64-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=tune %s
// tune: "-target-cpu" "sandybridge"
// tune-SAME: "-tune-cpu" "haswell"

// RUN: %clang_cl -m64 -arch:AVX512 -vlen=512 --target=x86_64-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=vlen512 %s
// RUN: %clang_cl -m64 -arch:AVX10.1 -vlen=512 --target=x86_64-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=vlen512 %s
// RUN: %clang_cl -m64 -arch:AVX10.2 -vlen=512 --target=x86_64-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=vlen512 %s
// vlen512: "-mprefer-vector-width=512"

// RUN: %clang_cl -m64 -arch:AVX512 -vlen=256 --target=x86_64-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=vlen256 %s
// RUN: %clang_cl -m64 -arch:AVX10.1 --target=x86_64-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=vlen256 %s
// RUN: %clang_cl -m64 -arch:AVX10.2 --target=x86_64-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=vlen256 %s
// RUN: %clang_cl -m64 -arch:AVX10.1 -vlen=256 --target=x86_64-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=vlen256 %s
// RUN: %clang_cl -m64 -arch:AVX10.2 -vlen=256 --target=x86_64-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=vlen256 %s
// vlen256: "-mprefer-vector-width=256"

// RUN: %clang_cl -m64 -arch:AVX512 -vlen=512 -vlen --target=x86_64-pc-windows -### -- 2>&1 %s | FileCheck -check-prefix=novlen %s
Expand Down
Loading