maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/clang/test/Driver/cuda-external-tools.cu

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

148 lines
6.3 KiB
Plaintext
Raw Normal View History

[CUDA] Driver changes to support CUDA compilation on MacOS. Summary: Compiling CUDA device code requires us to know the host toolchain, because CUDA device-side compiles pull in e.g. host headers. When we only supported Linux compilation, this worked because CudaToolChain, which is responsible for device-side CUDA compilation, inherited from the Linux toolchain. But in order to support MacOS, CudaToolChain needs to take a HostToolChain pointer. Because a CUDA toolchain now requires a host TC, we no longer will create a CUDA toolchain from Driver::getToolChain -- you have to go through CreateOffloadingDeviceToolChains. I am *pretty* sure this is correct, and that previously any attempt to create a CUDA toolchain through getToolChain() would eventually have resulted in us throwing "error: unsupported use of NVPTX for host compilation". In any case hacking getToolChain to create a CUDA+host toolchain would be wrong, because a Driver can be reused for multiple compilations, potentially with different host TCs, and getToolChain will cache the result, causing us to potentially use a stale host TC. So that's the main change in this patch. In addition, we have to pull CudaInstallationDetector out of Generic_GCC and into a top-level class. It's now used by the Generic_GCC and MachO toolchains. Reviewers: tra Subscribers: rryan, hfinkel, sfantao Differential Revision: https://reviews.llvm.org/D26774 llvm-svn: 287285
2016-11-18 08:41:22 +08:00
// Tests that ptxas and fatbinary are invoked correctly during CUDA
// compilation.
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
//
// REQUIRES: clang-driver
// REQUIRES: x86-registered-target
// REQUIRES: nvptx-registered-target
[CUDA] Handle -O options (more) correctly. Summary: 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. Reviewers: tra Subscribers: cfe-commits, echristo, jhen Differential Revision: http://reviews.llvm.org/D16307 llvm-svn: 258174
2016-01-20 03:52:21 +08:00
// 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 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,OPT0 %s
[CUDA] Handle -O options (more) correctly. Summary: 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. Reviewers: tra Subscribers: cfe-commits, echristo, jhen Differential Revision: http://reviews.llvm.org/D16307 llvm-svn: 258174
2016-01-20 03:52:21 +08:00
// RUN: %clang -### -target x86_64-linux-gnu -O1 -c %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,OPT1 %s
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
// RUN: %clang -### -target x86_64-linux-gnu -O2 -c %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,OPT2 %s
[CUDA] Handle -O options (more) correctly. Summary: 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. Reviewers: tra Subscribers: cfe-commits, echristo, jhen Differential Revision: http://reviews.llvm.org/D16307 llvm-svn: 258174
2016-01-20 03:52:21 +08:00
// RUN: %clang -### -target x86_64-linux-gnu -O3 -c %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,OPT3 %s
[CUDA] Handle -O options (more) correctly. Summary: 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. Reviewers: tra Subscribers: cfe-commits, echristo, jhen Differential Revision: http://reviews.llvm.org/D16307 llvm-svn: 258174
2016-01-20 03:52:21 +08:00
// RUN: %clang -### -target x86_64-linux-gnu -O4 -c %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,OPT3 %s
[CUDA] Handle -O options (more) correctly. Summary: 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. Reviewers: tra Subscribers: cfe-commits, echristo, jhen Differential Revision: http://reviews.llvm.org/D16307 llvm-svn: 258174
2016-01-20 03:52:21 +08:00
// RUN: %clang -### -target x86_64-linux-gnu -Ofast -c %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,OPT3 %s
[CUDA] Add option to generate relocatable device code As a first step, pass '-c/--compile-only' to ptxas so that it doesn't complain about references to external function. This will successfully generate object files, but they won't work at runtime because the registration routines need to adapted. Differential Revision: https://reviews.llvm.org/D42921 llvm-svn: 324878
2018-02-12 18:46:45 +08:00
// Generating relocatable device code
[HIP] Support early finalization of device code for -fno-gpu-rdc This patch renames -f{no-}cuda-rdc to -f{no-}gpu-rdc and keeps the original options as aliases. When -fgpu-rdc is off, clang will assume the device code in each translation unit does not call external functions except those in the device library, therefore it is possible to compile the device code in each translation unit to self-contained kernels and embed them in the host object, so that the host object behaves like usual host object which can be linked by lld. The benefits of this feature is: 1. allow users to create static libraries which can be linked by host linker; 2. amortized device code linking time. This patch modifies HIP action builder to insert actions for linking device code and generating HIP fatbin, and pass HIP fatbin to host backend action. It extracts code for constructing command for generating HIP fatbin as a function so that it can be reused by early finalization. It also modifies codegen of HIP host constructor functions to embed the device fatbin when it is available. Differential Revision: https://reviews.llvm.org/D52377 llvm-svn: 343611
2018-10-03 01:48:54 +08:00
// RUN: %clang -### -target x86_64-linux-gnu -fgpu-rdc -c %s 2>&1 \
[CUDA] Add option to generate relocatable device code As a first step, pass '-c/--compile-only' to ptxas so that it doesn't complain about references to external function. This will successfully generate object files, but they won't work at runtime because the registration routines need to adapted. Differential Revision: https://reviews.llvm.org/D42921 llvm-svn: 324878
2018-02-12 18:46:45 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,RDC %s
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
[CUDA] pass debug options to ptxas. ptxas optimizations are disabled if we need to generate debug info as ptxas does not accept '-g' otherwise. Differential Revision: http://reviews.llvm.org/D17111 llvm-svn: 261018
2016-02-17 06:03:20 +08:00
// With debugging enabled, ptxas should be run with with no ptxas optimizations.
[NVPTX] Emit debug info in DWARF-2 by default for Cuda devices. Summary: NVPTX target supports debug info in DWARF-2 format. Patch adds emission of debug info in DWARF-2 by default. Reviewers: tra, jlebar Subscribers: aprantl, JDevlieghere, cfe-commits Differential Revision: https://reviews.llvm.org/D42581 llvm-svn: 330272
2018-04-19 00:31:09 +08:00
// RUN: %clang -### -target x86_64-linux-gnu --cuda-noopt-device-debug -O2 -g -c %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,DBG %s
[CUDA] pass debug options to ptxas. ptxas optimizations are disabled if we need to generate debug info as ptxas does not accept '-g' otherwise. Differential Revision: http://reviews.llvm.org/D17111 llvm-svn: 261018
2016-02-17 06:03:20 +08:00
[CUDA] Add --no-cuda-noopt-debug, which disables --cuda-noopt-debug. Reviewers: tra Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D19251 llvm-svn: 266708
2016-04-19 10:27:11 +08:00
// --no-cuda-noopt-device-debug overrides --cuda-noopt-device-debug.
Driver must return non-zero code on errors in command line This is recommit of r302775, reverted in r302777 due to a fail in clang-tidy. Original mesage is below. Now if clang driver is given wrong arguments, in some cases it continues execution and returns zero code. This change fixes this behavior. The fix revealed some errors in clang test set. File test/Driver/gfortran.f90 added in r118203 checks forwarding gfortran flags to GCC. Now driver reports error on this file, because the option -working-directory implemented in clang differs from the option with the same name implemented in gfortran, in clang the option requires argument, in gfortran does not. In the file test/Driver/arm-darwin-builtin.c clang is called with options -fbuiltin-strcat and -fbuiltin-strcpy. These option were removed in r191435 and now clang reports error on this test. File arm-default-build-attributes.s uses option -verify, which is not supported by driver, it is cc1 option. Similarly, the file split-debug.h uses options -fmodules-embed-all-files and -fmodule-format=obj, which are not supported by driver. Other revealed errors are mainly mistypes. Differential Revision: https://reviews.llvm.org/D33013 llvm-svn: 303756
2017-05-24 22:57:17 +08:00
// RUN: %clang -### -target x86_64-linux-gnu --cuda-noopt-device-debug \
// RUN: --no-cuda-noopt-device-debug -O2 -c %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,OPT2 %s
[CUDA] Add --no-cuda-noopt-debug, which disables --cuda-noopt-debug. Reviewers: tra Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D19251 llvm-svn: 266708
2016-04-19 10:27:11 +08:00
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
// Regular compile without -O. This should result in us passing -O0 to ptxas.
// RUN: %clang -### -target x86_64-linux-gnu -c %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,OPT0 %s
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
[CUDA] Handle -O options (more) correctly. Summary: 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. Reviewers: tra Subscribers: cfe-commits, echristo, jhen Differential Revision: http://reviews.llvm.org/D16307 llvm-svn: 258174
2016-01-20 03:52:21 +08:00
// 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 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,OPT2 %s
[CUDA] Handle -O options (more) correctly. Summary: 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. Reviewers: tra Subscribers: cfe-commits, echristo, jhen Differential Revision: http://reviews.llvm.org/D16307 llvm-svn: 258174
2016-01-20 03:52:21 +08:00
// RUN: %clang -### -target x86_64-linux-gnu -Oz -c %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,OPT2 %s
[CUDA] Handle -O options (more) correctly. Summary: 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. Reviewers: tra Subscribers: cfe-commits, echristo, jhen Differential Revision: http://reviews.llvm.org/D16307 llvm-svn: 258174
2016-01-20 03:52:21 +08:00
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
// Regular compile targeting sm_35.
// RUN: %clang -### -target x86_64-linux-gnu --cuda-gpu-arch=sm_35 -c %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM35 %s
[CUDA] Add option to generate relocatable device code As a first step, pass '-c/--compile-only' to ptxas so that it doesn't complain about references to external function. This will successfully generate object files, but they won't work at runtime because the registration routines need to adapted. Differential Revision: https://reviews.llvm.org/D42921 llvm-svn: 324878
2018-02-12 18:46:45 +08:00
// Separate compilation targeting sm_35.
[HIP] Support early finalization of device code for -fno-gpu-rdc This patch renames -f{no-}cuda-rdc to -f{no-}gpu-rdc and keeps the original options as aliases. When -fgpu-rdc is off, clang will assume the device code in each translation unit does not call external functions except those in the device library, therefore it is possible to compile the device code in each translation unit to self-contained kernels and embed them in the host object, so that the host object behaves like usual host object which can be linked by lld. The benefits of this feature is: 1. allow users to create static libraries which can be linked by host linker; 2. amortized device code linking time. This patch modifies HIP action builder to insert actions for linking device code and generating HIP fatbin, and pass HIP fatbin to host backend action. It extracts code for constructing command for generating HIP fatbin as a function so that it can be reused by early finalization. It also modifies codegen of HIP host constructor functions to embed the device fatbin when it is available. Differential Revision: https://reviews.llvm.org/D52377 llvm-svn: 343611
2018-10-03 01:48:54 +08:00
// RUN: %clang -### -target x86_64-linux-gnu --cuda-gpu-arch=sm_35 -fgpu-rdc -c %s 2>&1 \
[CUDA] Add option to generate relocatable device code As a first step, pass '-c/--compile-only' to ptxas so that it doesn't complain about references to external function. This will successfully generate object files, but they won't work at runtime because the registration routines need to adapted. Differential Revision: https://reviews.llvm.org/D42921 llvm-svn: 324878
2018-02-12 18:46:45 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM35,RDC %s
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
// 32-bit compile.
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: %clang -### -target i386-linux-gnu -c %s 2>&1 \
// RUN: | FileCheck -check-prefixes=CHECK,ARCH32,SM20 %s
[CUDA] Add option to generate relocatable device code As a first step, pass '-c/--compile-only' to ptxas so that it doesn't complain about references to external function. This will successfully generate object files, but they won't work at runtime because the registration routines need to adapted. Differential Revision: https://reviews.llvm.org/D42921 llvm-svn: 324878
2018-02-12 18:46:45 +08:00
// 32-bit compile when generating relocatable device code.
[HIP] Support early finalization of device code for -fno-gpu-rdc This patch renames -f{no-}cuda-rdc to -f{no-}gpu-rdc and keeps the original options as aliases. When -fgpu-rdc is off, clang will assume the device code in each translation unit does not call external functions except those in the device library, therefore it is possible to compile the device code in each translation unit to self-contained kernels and embed them in the host object, so that the host object behaves like usual host object which can be linked by lld. The benefits of this feature is: 1. allow users to create static libraries which can be linked by host linker; 2. amortized device code linking time. This patch modifies HIP action builder to insert actions for linking device code and generating HIP fatbin, and pass HIP fatbin to host backend action. It extracts code for constructing command for generating HIP fatbin as a function so that it can be reused by early finalization. It also modifies codegen of HIP host constructor functions to embed the device fatbin when it is available. Differential Revision: https://reviews.llvm.org/D52377 llvm-svn: 343611
2018-10-03 01:48:54 +08:00
// RUN: %clang -### -target i386-linux-gnu -fgpu-rdc -c %s 2>&1 \
[CUDA] Add option to generate relocatable device code As a first step, pass '-c/--compile-only' to ptxas so that it doesn't complain about references to external function. This will successfully generate object files, but they won't work at runtime because the registration routines need to adapted. Differential Revision: https://reviews.llvm.org/D42921 llvm-svn: 324878
2018-02-12 18:46:45 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH32,SM20,RDC %s
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
// Compile with -fintegrated-as. This should still cause us to invoke ptxas.
// RUN: %clang -### -target x86_64-linux-gnu -fintegrated-as -c %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,OPT0 %s
[CUDA] Add option to generate relocatable device code As a first step, pass '-c/--compile-only' to ptxas so that it doesn't complain about references to external function. This will successfully generate object files, but they won't work at runtime because the registration routines need to adapted. Differential Revision: https://reviews.llvm.org/D42921 llvm-svn: 324878
2018-02-12 18:46:45 +08:00
// Check that we still pass -c when generating relocatable device code.
[HIP] Support early finalization of device code for -fno-gpu-rdc This patch renames -f{no-}cuda-rdc to -f{no-}gpu-rdc and keeps the original options as aliases. When -fgpu-rdc is off, clang will assume the device code in each translation unit does not call external functions except those in the device library, therefore it is possible to compile the device code in each translation unit to self-contained kernels and embed them in the host object, so that the host object behaves like usual host object which can be linked by lld. The benefits of this feature is: 1. allow users to create static libraries which can be linked by host linker; 2. amortized device code linking time. This patch modifies HIP action builder to insert actions for linking device code and generating HIP fatbin, and pass HIP fatbin to host backend action. It extracts code for constructing command for generating HIP fatbin as a function so that it can be reused by early finalization. It also modifies codegen of HIP host constructor functions to embed the device fatbin when it is available. Differential Revision: https://reviews.llvm.org/D52377 llvm-svn: 343611
2018-10-03 01:48:54 +08:00
// RUN: %clang -### -target x86_64-linux-gnu -fintegrated-as -fgpu-rdc -c %s 2>&1 \
[CUDA] Add option to generate relocatable device code As a first step, pass '-c/--compile-only' to ptxas so that it doesn't complain about references to external function. This will successfully generate object files, but they won't work at runtime because the registration routines need to adapted. Differential Revision: https://reviews.llvm.org/D42921 llvm-svn: 324878
2018-02-12 18:46:45 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,RDC %s
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
// Check -Xcuda-ptxas and -Xcuda-fatbinary
// RUN: %clang -### -target x86_64-linux-gnu -c -Xcuda-ptxas -foo1 \
// RUN: -Xcuda-fatbinary -bar1 -Xcuda-ptxas -foo2 -Xcuda-fatbinary -bar2 %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,SM20,PTXAS-EXTRA,FATBINARY-EXTRA %s
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
[CUDA] Driver changes to support CUDA compilation on MacOS. Summary: Compiling CUDA device code requires us to know the host toolchain, because CUDA device-side compiles pull in e.g. host headers. When we only supported Linux compilation, this worked because CudaToolChain, which is responsible for device-side CUDA compilation, inherited from the Linux toolchain. But in order to support MacOS, CudaToolChain needs to take a HostToolChain pointer. Because a CUDA toolchain now requires a host TC, we no longer will create a CUDA toolchain from Driver::getToolChain -- you have to go through CreateOffloadingDeviceToolChains. I am *pretty* sure this is correct, and that previously any attempt to create a CUDA toolchain through getToolChain() would eventually have resulted in us throwing "error: unsupported use of NVPTX for host compilation". In any case hacking getToolChain to create a CUDA+host toolchain would be wrong, because a Driver can be reused for multiple compilations, potentially with different host TCs, and getToolChain will cache the result, causing us to potentially use a stale host TC. So that's the main change in this patch. In addition, we have to pull CudaInstallationDetector out of Generic_GCC and into a top-level class. It's now used by the Generic_GCC and MachO toolchains. Reviewers: tra Subscribers: rryan, hfinkel, sfantao Differential Revision: https://reviews.llvm.org/D26774 llvm-svn: 287285
2016-11-18 08:41:22 +08:00
// MacOS spot-checks
// RUN: %clang -### -target x86_64-apple-macosx -O0 -c %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,OPT0 %s
[CUDA] Driver changes to support CUDA compilation on MacOS. Summary: Compiling CUDA device code requires us to know the host toolchain, because CUDA device-side compiles pull in e.g. host headers. When we only supported Linux compilation, this worked because CudaToolChain, which is responsible for device-side CUDA compilation, inherited from the Linux toolchain. But in order to support MacOS, CudaToolChain needs to take a HostToolChain pointer. Because a CUDA toolchain now requires a host TC, we no longer will create a CUDA toolchain from Driver::getToolChain -- you have to go through CreateOffloadingDeviceToolChains. I am *pretty* sure this is correct, and that previously any attempt to create a CUDA toolchain through getToolChain() would eventually have resulted in us throwing "error: unsupported use of NVPTX for host compilation". In any case hacking getToolChain to create a CUDA+host toolchain would be wrong, because a Driver can be reused for multiple compilations, potentially with different host TCs, and getToolChain will cache the result, causing us to potentially use a stale host TC. So that's the main change in this patch. In addition, we have to pull CudaInstallationDetector out of Generic_GCC and into a top-level class. It's now used by the Generic_GCC and MachO toolchains. Reviewers: tra Subscribers: rryan, hfinkel, sfantao Differential Revision: https://reviews.llvm.org/D26774 llvm-svn: 287285
2016-11-18 08:41:22 +08:00
// RUN: %clang -### -target x86_64-apple-macosx --cuda-gpu-arch=sm_35 -c %s 2>&1 \
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM35 %s
// RUN: %clang -### -target i386-apple-macosx -c %s 2>&1 \
// RUN: | FileCheck -check-prefixes=CHECK,ARCH32,SM20 %s
[CUDA] Driver changes to support CUDA compilation on MacOS. Summary: Compiling CUDA device code requires us to know the host toolchain, because CUDA device-side compiles pull in e.g. host headers. When we only supported Linux compilation, this worked because CudaToolChain, which is responsible for device-side CUDA compilation, inherited from the Linux toolchain. But in order to support MacOS, CudaToolChain needs to take a HostToolChain pointer. Because a CUDA toolchain now requires a host TC, we no longer will create a CUDA toolchain from Driver::getToolChain -- you have to go through CreateOffloadingDeviceToolChains. I am *pretty* sure this is correct, and that previously any attempt to create a CUDA toolchain through getToolChain() would eventually have resulted in us throwing "error: unsupported use of NVPTX for host compilation". In any case hacking getToolChain to create a CUDA+host toolchain would be wrong, because a Driver can be reused for multiple compilations, potentially with different host TCs, and getToolChain will cache the result, causing us to potentially use a stale host TC. So that's the main change in this patch. In addition, we have to pull CudaInstallationDetector out of Generic_GCC and into a top-level class. It's now used by the Generic_GCC and MachO toolchains. Reviewers: tra Subscribers: rryan, hfinkel, sfantao Differential Revision: https://reviews.llvm.org/D26774 llvm-svn: 287285
2016-11-18 08:41:22 +08:00
[CUDA] Add option to generate relocatable device code As a first step, pass '-c/--compile-only' to ptxas so that it doesn't complain about references to external function. This will successfully generate object files, but they won't work at runtime because the registration routines need to adapted. Differential Revision: https://reviews.llvm.org/D42921 llvm-svn: 324878
2018-02-12 18:46:45 +08:00
// Check relocatable device code generation on MacOS.
[HIP] Support early finalization of device code for -fno-gpu-rdc This patch renames -f{no-}cuda-rdc to -f{no-}gpu-rdc and keeps the original options as aliases. When -fgpu-rdc is off, clang will assume the device code in each translation unit does not call external functions except those in the device library, therefore it is possible to compile the device code in each translation unit to self-contained kernels and embed them in the host object, so that the host object behaves like usual host object which can be linked by lld. The benefits of this feature is: 1. allow users to create static libraries which can be linked by host linker; 2. amortized device code linking time. This patch modifies HIP action builder to insert actions for linking device code and generating HIP fatbin, and pass HIP fatbin to host backend action. It extracts code for constructing command for generating HIP fatbin as a function so that it can be reused by early finalization. It also modifies codegen of HIP host constructor functions to embed the device fatbin when it is available. Differential Revision: https://reviews.llvm.org/D52377 llvm-svn: 343611
2018-10-03 01:48:54 +08:00
// RUN: %clang -### -target x86_64-apple-macosx -O0 -fgpu-rdc -c %s 2>&1 \
[CUDA] Add option to generate relocatable device code As a first step, pass '-c/--compile-only' to ptxas so that it doesn't complain about references to external function. This will successfully generate object files, but they won't work at runtime because the registration routines need to adapted. Differential Revision: https://reviews.llvm.org/D42921 llvm-svn: 324878
2018-02-12 18:46:45 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM20,RDC %s
[HIP] Support early finalization of device code for -fno-gpu-rdc This patch renames -f{no-}cuda-rdc to -f{no-}gpu-rdc and keeps the original options as aliases. When -fgpu-rdc is off, clang will assume the device code in each translation unit does not call external functions except those in the device library, therefore it is possible to compile the device code in each translation unit to self-contained kernels and embed them in the host object, so that the host object behaves like usual host object which can be linked by lld. The benefits of this feature is: 1. allow users to create static libraries which can be linked by host linker; 2. amortized device code linking time. This patch modifies HIP action builder to insert actions for linking device code and generating HIP fatbin, and pass HIP fatbin to host backend action. It extracts code for constructing command for generating HIP fatbin as a function so that it can be reused by early finalization. It also modifies codegen of HIP host constructor functions to embed the device fatbin when it is available. Differential Revision: https://reviews.llvm.org/D52377 llvm-svn: 343611
2018-10-03 01:48:54 +08:00
// RUN: %clang -### -target x86_64-apple-macosx --cuda-gpu-arch=sm_35 -fgpu-rdc -c %s 2>&1 \
[CUDA] Add option to generate relocatable device code As a first step, pass '-c/--compile-only' to ptxas so that it doesn't complain about references to external function. This will successfully generate object files, but they won't work at runtime because the registration routines need to adapted. Differential Revision: https://reviews.llvm.org/D42921 llvm-svn: 324878
2018-02-12 18:46:45 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH64,SM35,RDC %s
[HIP] Support early finalization of device code for -fno-gpu-rdc This patch renames -f{no-}cuda-rdc to -f{no-}gpu-rdc and keeps the original options as aliases. When -fgpu-rdc is off, clang will assume the device code in each translation unit does not call external functions except those in the device library, therefore it is possible to compile the device code in each translation unit to self-contained kernels and embed them in the host object, so that the host object behaves like usual host object which can be linked by lld. The benefits of this feature is: 1. allow users to create static libraries which can be linked by host linker; 2. amortized device code linking time. This patch modifies HIP action builder to insert actions for linking device code and generating HIP fatbin, and pass HIP fatbin to host backend action. It extracts code for constructing command for generating HIP fatbin as a function so that it can be reused by early finalization. It also modifies codegen of HIP host constructor functions to embed the device fatbin when it is available. Differential Revision: https://reviews.llvm.org/D52377 llvm-svn: 343611
2018-10-03 01:48:54 +08:00
// RUN: %clang -### -target i386-apple-macosx -fgpu-rdc -c %s 2>&1 \
[CUDA] Add option to generate relocatable device code As a first step, pass '-c/--compile-only' to ptxas so that it doesn't complain about references to external function. This will successfully generate object files, but they won't work at runtime because the registration routines need to adapted. Differential Revision: https://reviews.llvm.org/D42921 llvm-svn: 324878
2018-02-12 18:46:45 +08:00
// RUN: | FileCheck -check-prefixes=CHECK,ARCH32,SM20,RDC %s
[OpenMP] Pass -v to PTXAS if it was passed to the driver. Summary: When compiling code being offloaded by OpenMP to an NVIDIA GPU, pass the -v to PTXAS if it was passed to the CLANG driver. Reviewers: arpith-jacob, caomhin, carlo.bertolli, ABataev, jlebar, hfinkel, tstellar Reviewed By: jlebar Subscribers: Hahnfeld, rengolin, cfe-commits Differential Revision: https://reviews.llvm.org/D29644 llvm-svn: 310295
2017-08-08 04:19:23 +08:00
// Check that CLANG forwards the -v flag to PTXAS.
// RUN: %clang -### -save-temps -no-canonical-prefixes -v %s 2>&1 \
// RUN: | FileCheck -check-prefix=CHK-PTXAS-VERBOSE %s
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
// Match clang job that produces PTX assembly.
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// CHECK: "-cc1"
// ARCH64-SAME: "-triple" "nvptx64-nvidia-cuda"
// ARCH32-SAME: "-triple" "nvptx-nvidia-cuda"
[HIP] Support early finalization of device code for -fno-gpu-rdc This patch renames -f{no-}cuda-rdc to -f{no-}gpu-rdc and keeps the original options as aliases. When -fgpu-rdc is off, clang will assume the device code in each translation unit does not call external functions except those in the device library, therefore it is possible to compile the device code in each translation unit to self-contained kernels and embed them in the host object, so that the host object behaves like usual host object which can be linked by lld. The benefits of this feature is: 1. allow users to create static libraries which can be linked by host linker; 2. amortized device code linking time. This patch modifies HIP action builder to insert actions for linking device code and generating HIP fatbin, and pass HIP fatbin to host backend action. It extracts code for constructing command for generating HIP fatbin as a function so that it can be reused by early finalization. It also modifies codegen of HIP host constructor functions to embed the device fatbin when it is available. Differential Revision: https://reviews.llvm.org/D52377 llvm-svn: 343611
2018-10-03 01:48:54 +08:00
// RDC-SAME: "-fgpu-rdc"
// CHECK-NOT: "-fgpu-rdc"
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// SM20-SAME: "-target-cpu" "sm_20"
// SM35-SAME: "-target-cpu" "sm_35"
// SM20-SAME: "-o" "[[PTXFILE:[^"]*]]"
// SM35-SAME: "-o" "[[PTXFILE:[^"]*]]"
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
// Match the call to ptxas (which assembles PTX to SASS).
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// CHECK: ptxas
// ARCH64-SAME: "-m64"
// ARCH32-SAME: "-m32"
// OPT0-SAME: "-O0"
[CUDA] pass debug options to ptxas. ptxas optimizations are disabled if we need to generate debug info as ptxas does not accept '-g' otherwise. Differential Revision: http://reviews.llvm.org/D17111 llvm-svn: 261018
2016-02-17 06:03:20 +08:00
// OPT0-NOT: "-g"
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// OPT1-SAME: "-O1"
[CUDA] pass debug options to ptxas. ptxas optimizations are disabled if we need to generate debug info as ptxas does not accept '-g' otherwise. Differential Revision: http://reviews.llvm.org/D17111 llvm-svn: 261018
2016-02-17 06:03:20 +08:00
// OPT1-NOT: "-g"
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// OPT2-SAME: "-O2"
[CUDA] pass debug options to ptxas. ptxas optimizations are disabled if we need to generate debug info as ptxas does not accept '-g' otherwise. Differential Revision: http://reviews.llvm.org/D17111 llvm-svn: 261018
2016-02-17 06:03:20 +08:00
// OPT2-NOT: "-g"
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// OPT3-SAME: "-O3"
[CUDA] pass debug options to ptxas. ptxas optimizations are disabled if we need to generate debug info as ptxas does not accept '-g' otherwise. Differential Revision: http://reviews.llvm.org/D17111 llvm-svn: 261018
2016-02-17 06:03:20 +08:00
// OPT3-NOT: "-g"
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// DBG-SAME: "-g" "--dont-merge-basicblocks" "--return-at-end"
// SM20-SAME: "--gpu-name" "sm_20"
// SM35-SAME: "--gpu-name" "sm_35"
// SM20-SAME: "--output-file" "[[CUBINFILE:[^"]*]]"
// SM35-SAME: "--output-file" "[[CUBINFILE:[^"]*]]"
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
// CHECK-SAME: "[[PTXFILE]]"
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// PTXAS-EXTRA-SAME: "-foo1"
// PTXAS-EXTRA-SAME: "-foo2"
[CUDA] Add option to generate relocatable device code As a first step, pass '-c/--compile-only' to ptxas so that it doesn't complain about references to external function. This will successfully generate object files, but they won't work at runtime because the registration routines need to adapted. Differential Revision: https://reviews.llvm.org/D42921 llvm-svn: 324878
2018-02-12 18:46:45 +08:00
// RDC-SAME: "-c"
// CHECK-NOT: "-c"
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
// Match the call to fatbinary (which combines all our PTX and SASS into one
// blob).
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// CHECK: fatbinary
// CHECK-SAME-DAG: "--cuda"
// ARCH64-SAME-DAG: "-64"
// ARCH32-SAME-DAG: "-32"
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
// CHECK-DAG: "--create" "[[FATBINARY:[^"]*]]"
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// SM20-SAME-DAG: "--image=profile=compute_20,file=[[PTXFILE]]"
// SM35-SAME-DAG: "--image=profile=compute_35,file=[[PTXFILE]]"
// SM20-SAME-DAG: "--image=profile=sm_20,file=[[CUBINFILE]]"
// SM35-SAME-DAG: "--image=profile=sm_35,file=[[CUBINFILE]]"
// FATBINARY-EXTRA-SAME: "-bar1"
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
// FATBINARY-EXTRA-SAME: "-bar2"
// Match the clang job for host compilation.
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// CHECK: "-cc1"
// ARCH64-SAME: "-triple" "x86_64-
// ARCH32-SAME: "-triple" "i386-
[CUDA] Invoke ptxas and fatbinary during compilation. Summary: Previously we compiled CUDA device code to PTX assembly and embedded that asm as text in our host binary. Now we compile to PTX assembly and then invoke ptxas to assemble the PTX into a cubin file. We gather the ptx and cubin files for each of our --cuda-gpu-archs and combine them using fatbinary, and then embed that into the host binary. Adds two new command-line flags, -Xcuda_ptxas and -Xcuda_fatbinary, which pass args down to the external tools. Reviewers: tra, echristo Subscribers: cfe-commits, jhen Differential Revision: http://reviews.llvm.org/D16082 llvm-svn: 257809
2016-01-15 05:41:27 +08:00
// CHECK-SAME: "-fcuda-include-gpubinary" "[[FATBINARY]]"
[HIP] Support early finalization of device code for -fno-gpu-rdc This patch renames -f{no-}cuda-rdc to -f{no-}gpu-rdc and keeps the original options as aliases. When -fgpu-rdc is off, clang will assume the device code in each translation unit does not call external functions except those in the device library, therefore it is possible to compile the device code in each translation unit to self-contained kernels and embed them in the host object, so that the host object behaves like usual host object which can be linked by lld. The benefits of this feature is: 1. allow users to create static libraries which can be linked by host linker; 2. amortized device code linking time. This patch modifies HIP action builder to insert actions for linking device code and generating HIP fatbin, and pass HIP fatbin to host backend action. It extracts code for constructing command for generating HIP fatbin as a function so that it can be reused by early finalization. It also modifies codegen of HIP host constructor functions to embed the device fatbin when it is available. Differential Revision: https://reviews.llvm.org/D52377 llvm-svn: 343611
2018-10-03 01:48:54 +08:00
// RDC-SAME: "-fgpu-rdc"
// CHECK-NOT: "-fgpu-rdc"
[OpenMP] Pass -v to PTXAS if it was passed to the driver. Summary: When compiling code being offloaded by OpenMP to an NVIDIA GPU, pass the -v to PTXAS if it was passed to the CLANG driver. Reviewers: arpith-jacob, caomhin, carlo.bertolli, ABataev, jlebar, hfinkel, tstellar Reviewed By: jlebar Subscribers: Hahnfeld, rengolin, cfe-commits Differential Revision: https://reviews.llvm.org/D29644 llvm-svn: 310295
2017-08-08 04:19:23 +08:00
[CUDA] Fix test cuda-external-tools.cu This didn't verify the CHECK prefix before! Differential Revision: https://reviews.llvm.org/D42920 llvm-svn: 324877
2018-02-12 18:46:34 +08:00
// CHK-PTXAS-VERBOSE: ptxas{{.*}}" "-v"