[PATCH] D16307: [CUDA] Handle -O options (more) correctly.

[Eric Christopher via cfe-commits]

At a quick glance Os and Oz should probably map to O2.

On Mon, Jan 18, 2016, 4:19 PM Justin Lebar <jlebar at google.com> wrote:

> jlebar created this revision.
> jlebar added a reviewer: tra.
> jlebar added subscribers: jhen, echristo, cfe-commits.
>
> Previously we'd crash the driver if you passed -O0.  Now we try to
> handle all of clang's various optimization flags in a sane way.
>
> http://reviews.llvm.org/D16307
>
> Files:
>   lib/Driver/Tools.cpp
>   test/Driver/cuda-external-tools.cu
>
> Index: test/Driver/cuda-external-tools.cu
> ===================================================================
> --- test/Driver/cuda-external-tools.cu
> +++ test/Driver/cuda-external-tools.cu
> @@ -4,14 +4,31 @@
>  // REQUIRES: x86-registered-target
>  // REQUIRES: nvptx-registered-target
>
> -// Regular compile with -O2.
> +// Regular compiles with -O{0,1,2,3,4,fast}.  -O4 and -Ofast map to ptxas
> O3.
> +// RUN: %clang -### -target x86_64-linux-gnu -O0 -c %s 2>&1 \
> +// RUN: | FileCheck -check-prefix ARCH64 -check-prefix SM20 -check-prefix
> OPT0 %s
> +// RUN: %clang -### -target x86_64-linux-gnu -O1 -c %s 2>&1 \
> +// RUN: | FileCheck -check-prefix ARCH64 -check-prefix SM20 -check-prefix
> OPT1 %s
>  // RUN: %clang -### -target x86_64-linux-gnu -O2 -c %s 2>&1 \
>  // RUN: | FileCheck -check-prefix ARCH64 -check-prefix SM20 -check-prefix
> OPT2 %s
> +// RUN: %clang -### -target x86_64-linux-gnu -O3 -c %s 2>&1 \
> +// RUN: | FileCheck -check-prefix ARCH64 -check-prefix SM20 -check-prefix
> OPT3 %s
> +// RUN: %clang -### -target x86_64-linux-gnu -O4 -c %s 2>&1 \
> +// RUN: | FileCheck -check-prefix ARCH64 -check-prefix SM20 -check-prefix
> OPT3 %s
> +// RUN: %clang -### -target x86_64-linux-gnu -Ofast -c %s 2>&1 \
> +// RUN: | FileCheck -check-prefix ARCH64 -check-prefix SM20 -check-prefix
> OPT3 %s
>
>  // Regular compile without -O.  This should result in us passing -O0 to
> ptxas.
>  // RUN: %clang -### -target x86_64-linux-gnu -c %s 2>&1 \
>  // RUN: | FileCheck -check-prefix ARCH64 -check-prefix SM20 -check-prefix
> OPT0 %s
>
> +// Regular compiles with -Os and -Oz.  For lack of a better option, we map
> +// these to ptxas -O3.
> +// RUN: %clang -### -target x86_64-linux-gnu -Os -c %s 2>&1 \
> +// RUN: | FileCheck -check-prefix ARCH64 -check-prefix SM20 -check-prefix
> OPT3 %s
> +// RUN: %clang -### -target x86_64-linux-gnu -Oz -c %s 2>&1 \
> +// RUN: | FileCheck -check-prefix ARCH64 -check-prefix SM20 -check-prefix
> OPT3 %s
> +
>  // Regular compile targeting sm_35.
>  // RUN: %clang -### -target x86_64-linux-gnu --cuda-gpu-arch=sm_35 -c %s
> 2>&1 \
>  // RUN: | FileCheck -check-prefix ARCH64 -check-prefix SM35 %s
> @@ -42,7 +59,9 @@
>  // ARCH64: "-m64"
>  // ARCH32: "-m32"
>  // OPT0: "-O0"
> +// OPT1: "-O1"
>  // OPT2: "-O2"
> +// OPT3: "-O3"
>  // SM20: "--gpu-name" "sm_20"
>  // SM35: "--gpu-name" "sm_35"
>  // SM20: "--output-file" "[[CUBINFILE:[^"]*]]"
> Index: lib/Driver/Tools.cpp
> ===================================================================
> --- lib/Driver/Tools.cpp
> +++ lib/Driver/Tools.cpp
> @@ -10645,10 +10645,34 @@
>    ArgStringList CmdArgs;
>    CmdArgs.push_back(TC.getTriple().isArch64Bit() ? "-m64" : "-m32");
>
> -  // Clang's default optimization level is -O0, but ptxas's default is
> -O3.
> -  CmdArgs.push_back(Args.MakeArgString(
> -      llvm::Twine("-O") +
> -      Args.getLastArgValue(options::OPT_O_Group, "0").data()));
> +  // Map the -O we received to -O{0,1,2,3}.
> +  //
> +  // TODO: Perhaps we should map host -O2 to ptxas -O3. -O3 is ptxas's
> default,
> +  // so it may correspond more closely to the spirit of clang -O2.
> +  if (Arg *A = Args.getLastArg(options::OPT_O_Group)) {
> +    // -O3 seems like the least-bad option when -Osomething is specified
> to
> +    // clang but it isn't handled above.
> +    StringRef OOpt = "3";
> +    if (A->getOption().matches(options::OPT_O4) ||
> +        A->getOption().matches(options::OPT_Ofast))
> +      OOpt = "3";
> +    else if (A->getOption().matches(options::OPT_O0))
> +      OOpt = "0";
> +    else if (A->getOption().matches(options::OPT_O)) {
> +      // -Os, -Oz, and -O(anything else) map to -O3, for lack of better
> options.
> +      OOpt = llvm::StringSwitch<const char *>(A->getValue())
> +                 .Case("1", "1")
> +                 .Case("2", "2")
> +                 .Case("s", "3")
> +                 .Case("z", "3")
> +                 .Default("3");
> +    }
> +    CmdArgs.push_back(Args.MakeArgString(llvm::Twine("-O") + OOpt));
> +  } else {
> +    // If no -O was passed, pass -O0 to ptxas -- no opt flag should
> correspond
> +    // to no optimizations, but ptxas's default is -O3.
> +    CmdArgs.push_back("-O0");
> +  }
>
>    // Don't bother passing -g to ptxas: It's enabled by default at -O0, and
>    // not supported at other optimization levels.
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20160119/72d096b1/attachment-0001.html>