[CUDA] Added --[no-]cuda-include-ptx=sm_XX|all option.

[tra]

Summary

Currently we always include PTX into the fatbin along
with the GPU code. It about doubles the size of the GPU binary
we need to carry in the executable. These options allow control
inclusion of PTX into GPU binary.

This patch does not change the defaults, though we may consider
making no-PTX the default in the future.

[jlebar]

Where do we document the default values for these flags? Like, how is a user supposed to figure out whether the default is to include ptx or not?

[tra]

Documented new options in ClangCommandLineReference.rst

[tra]

In D45495#1063370, @jlebar wrote:

Where do we document the default values for these flags? Like, how is a user supposed to figure out whether the default is to include ptx or not?

I've updated docs/ClangCommandLineReference.rst.

[jlebar]

lgtm!

clang/include/clang/Driver/Options.td

 def cuda_device_only : Flag<["--"], "cuda-device-only">,
   HelpText<"Compile CUDA code for device only">;
 def cuda_host_only : Flag<["--"], "cuda-host-only">,
   HelpText<"Compile CUDA code for host only.  Has no effect on non-CUDA "
            "compilations.">;
 def cuda_compile_host_device : Flag<["--"], "cuda-compile-host-device">,
   HelpText<"Compile CUDA code for both host and device (default).  Has no "
            "effect on non-CUDA compilations.">;
+def cuda_include_ptx_EQ : Joined<["--"], "cuda-include-ptx=">, Flags<[DriverOption]>,
+  HelpText<"Include PTX for the follwing GPU architecture (e.g. sm_35) or 'all'. May be specified more than once.">;
+def no_cuda_include_ptx_EQ : Joined<["--"], "no-cuda-include-ptx=">, Flags<[DriverOption]>,
+  HelpText<"Do not include PTX for the follwing GPU architecture (e.g. sm_35) or 'all'. May be specified more than once.">;
 def cuda_gpu_arch_EQ : Joined<["--"], "cuda-gpu-arch=">, Flags<[DriverOption]>,
   HelpText<"CUDA GPU architecture (e.g. sm_35).  May be specified more than once.">;
 def no_cuda_gpu_arch_EQ : Joined<["--"], "no-cuda-gpu-arch=">, Flags<[DriverOption]>,
   HelpText<"Remove GPU architecture (e.g. sm_35) from the list of GPUs to compile for. "
            "'all' resets the list to its default value.">;
 def cuda_noopt_device_debug : Flag<["--"], "cuda-noopt-device-debug">,
   HelpText<"Enable device-side debug info generation. Disables ptxas optimizations.">;
 def no_cuda_version_check : Flag<["--"], "no-cuda-version-check">,

clang/lib/Driver/ToolChains/Cuda.cpp

void NVPTX::Assembler::ConstructJob(Compilation &C, const JobAction &JA,
   const char *Exec;
   if (Arg *A = Args.getLastArg(options::OPT_ptxas_path_EQ))
     Exec = A->getValue();
   else
     Exec = Args.MakeArgString(TC.GetProgramPath("ptxas"));
   C.addCommand(llvm::make_unique<Command>(JA, *this, Exec, CmdArgs, Inputs));
 }
 
+static bool shouldIncludePTX(const ArgList &Args, const char *gpu_arch) {
+  bool includePTX = true;
+  for (Arg *A : Args) {
+    if (!(A->getOption().matches(options::OPT_cuda_include_ptx_EQ) ||
+          A->getOption().matches(options::OPT_no_cuda_include_ptx_EQ)))
+      continue;
+    A->claim();
+    const StringRef ArchStr = A->getValue();
+    if (ArchStr == "all" || ArchStr == gpu_arch) {
+      includePTX = A->getOption().matches(options::OPT_cuda_include_ptx_EQ);
+      continue;
+    }
+  }
+  return includePTX;
+}
+
 // All inputs to this linker must be from CudaDeviceActions, as we need to look
 // at the Inputs' Actions in order to figure out which GPU architecture they
 // correspond to.
 void NVPTX::Linker::ConstructJob(Compilation &C, const JobAction &JA,
                                  const InputInfo &Output,
                                  const InputInfoList &Inputs,
                                  const ArgList &Args,
                                  const char *LinkingOutput) const {
for (const auto& II : Inputs) {
     auto *A = II.getAction();
     assert(A->getInputs().size() == 1 &&
            "Device offload action is expected to have a single input");
     const char *gpu_arch_str = A->getOffloadingArch();
     assert(gpu_arch_str &&
            "Device action expected to have associated a GPU architecture!");
     CudaArch gpu_arch = StringToCudaArch(gpu_arch_str);
 
+    if (II.getType() == types::TY_PP_Asm &&
+        !shouldIncludePTX(Args, gpu_arch_str))
+      continue;
     // We need to pass an Arch of the form "sm_XX" for cubin files and
     // "compute_XX" for ptx.
     const char *Arch =
         (II.getType() == types::TY_PP_Asm)
             ? CudaVirtualArchToString(VirtualArchForCudaArch(gpu_arch))
             : gpu_arch_str;
     CmdArgs.push_back(Args.MakeArgString(llvm::Twine("--image=profile=") +
                                          Arch + ",file=" + II.getFilename()));

clang/test/Driver/cuda-options.cu

 // RUN: | FileCheck -check-prefixes NOARCH-SM20,NOARCH-SM30,ARCH-SM35 %s
 
 // g) There's no --cuda-gpu-arch=all
 // RUN: %clang -### -target x86_64-linux-gnu --cuda-device-only \
 // RUN:   --cuda-gpu-arch=all \
 // RUN:   -c %s 2>&1 \
 // RUN: | FileCheck -check-prefix ARCHALLERROR %s
 
+
+// Verify that --[no-]cuda-include-ptx arguments are handled correctly.
+// a) by default we're including PTX for all GPUs.
+// RUN: %clang -### -target x86_64-linux-gnu \
+// RUN:   --cuda-gpu-arch=sm_35 --cuda-gpu-arch=sm_30 \
+// RUN:   -c %s 2>&1 \
+// RUN: | FileCheck -check-prefixes FATBIN-COMMON,PTX-SM35,PTX-SM30 %s
+
+// b) --no-cuda-include-ptx=all disables PTX inclusion for all GPUs
+// RUN: %clang -### -target x86_64-linux-gnu \
+// RUN:   --cuda-gpu-arch=sm_35 --cuda-gpu-arch=sm_30 \
+// RUN:   --no-cuda-include-ptx=all \
+// RUN:   -c %s 2>&1 \
+// RUN: | FileCheck -check-prefixes FATBIN-COMMON,NOPTX-SM35,NOPTX-SM30 %s
+
+// c) --no-cuda-include-ptx=sm_XX disables PTX inclusion for that GPU only.
+// RUN: %clang -### -target x86_64-linux-gnu \
+// RUN:   --cuda-gpu-arch=sm_35 --cuda-gpu-arch=sm_30 \
+// RUN:   --no-cuda-include-ptx=sm_35 \
+// RUN:   -c %s 2>&1 \
+// RUN: | FileCheck -check-prefixes FATBIN-COMMON,NOPTX-SM35,PTX-SM30 %s
+// RUN: %clang -### -target x86_64-linux-gnu \
+// RUN:   --cuda-gpu-arch=sm_35 --cuda-gpu-arch=sm_30 \
+// RUN:   --no-cuda-include-ptx=sm_30 \
+// RUN:   -c %s 2>&1 \
+// RUN: | FileCheck -check-prefixes FATBIN-COMMON,PTX-SM35,NOPTX-SM30 %s
+
+// d) --cuda-include-ptx=all overrides preceding --no-cuda-include-ptx=all
+// RUN: %clang -### -target x86_64-linux-gnu \
+// RUN:   --cuda-gpu-arch=sm_35 --cuda-gpu-arch=sm_30 \
+// RUN:   --no-cuda-include-ptx=all --cuda-include-ptx=all \
+// RUN:   -c %s 2>&1 \
+// RUN: | FileCheck -check-prefixes FATBIN-COMMON,PTX-SM35,PTX-SM30 %s
+
+// e) --cuda-include-ptx=all overrides preceding --no-cuda-include-ptx=sm_XX
+// RUN: %clang -### -target x86_64-linux-gnu \
+// RUN:   --cuda-gpu-arch=sm_35 --cuda-gpu-arch=sm_30 \
+// RUN:   --no-cuda-include-ptx=sm_30 --cuda-include-ptx=all \
+// RUN:   -c %s 2>&1 \
+// RUN: | FileCheck -check-prefixes FATBIN-COMMON,PTX-SM35,PTX-SM30 %s
+
+
 // ARCH-SM20: "-cc1"{{.*}}"-target-cpu" "sm_20"
 // NOARCH-SM20-NOT: "-cc1"{{.*}}"-target-cpu" "sm_20"
 // ARCH-SM30: "-cc1"{{.*}}"-target-cpu" "sm_30"
 // NOARCH-SM30-NOT: "-cc1"{{.*}}"-target-cpu" "sm_30"
 // ARCH-SM35: "-cc1"{{.*}}"-target-cpu" "sm_35"
 // NOARCH-SM35-NOT: "-cc1"{{.*}}"-target-cpu" "sm_35"
 // ARCHALLERROR: error: Unsupported CUDA gpu architecture: all
 

 // NOHOST-NOT: "-x" "cuda"
 
 // Match linker.
 // LINK: "{{.*}}{{ld|link}}{{(.exe)?}}"
 // LINK-SAME: "[[HOSTOUTPUT]]"
 
 // Match no linker.
 // NOLINK-NOT: "{{.*}}{{ld|link}}{{(.exe)?}}"
+
+// FATBIN-COMMON:fatbinary
+// FATBIN-COMMON: "--create" "[[FATBINARY:[^"]*]]"
+// FATBIN-COMMON: "--image=profile=sm_30,file=
+// PTX-SM30: "--image=profile=compute_30,file=
+// NOPTX-SM30-NOT: "--image=profile=compute_30,file=
+// FATBIN-COMMON: "--image=profile=sm_35,file=
+// PTX-SM35: "--image=profile=compute_35,file=
+// NOPTX-SM35-NOT: "--image=profile=compute_35,file=