Author: yaxunl
Date: Fri Aug 17 10:47:31 2018
New Revision: 340056
URL: http://llvm.org/viewvc/llvm-project?rev=340056&view=rev
Log:
[HIP] Make __hip_gpubin_handle hidden to avoid being merged across different
shared libraries
Different shared libraries contain different fat binary, which is sto
Author: yaxunl
Date: Tue May 8 14:02:12 2018
New Revision: 331811
URL: http://llvm.org/viewvc/llvm-project?rev=331811&view=rev
Log:
[HIP] Add hip offload kind
There are quite differences in HIP action builder and action job creation,
which justifies to define a separate offload kind.
Differenti
Author: yaxunl
Date: Wed May 9 10:07:06 2018
New Revision: 331895
URL: http://llvm.org/viewvc/llvm-project?rev=331895&view=rev
Log:
[OpenCL] Fix typos in emitted enqueue kernel function names
Two typos:
vaarg => vararg
get_kernel_preferred_work_group_multiple =>
get_kernel_preferred_work_group
Author: yaxunl
Date: Fri May 11 12:02:18 2018
New Revision: 332121
URL: http://llvm.org/viewvc/llvm-project?rev=332121&view=rev
Log:
[HIP] Let clang-offload-bundler support HIP
When bundle/unbundle intermediate files for HIP, there may be multiple
sub archs, therefore BoundArch needs to be includ
Author: yaxunl
Date: Fri May 11 12:14:34 2018
New Revision: 332122
URL: http://llvm.org/viewvc/llvm-project?rev=332122&view=rev
Log:
[HIP] Diagnose unsupported host triple
Differential Revision: https://reviews.llvm.org/D46487
Modified:
cfe/trunk/include/clang/Basic/DiagnosticDriverKinds.td
Author: yaxunl
Date: Fri May 11 12:21:39 2018
New Revision: 332123
URL: http://llvm.org/viewvc/llvm-project?rev=332123&view=rev
Log:
[HIP] Set proper triple and offload kind for the toolchain
Also introduce --hip-link option to indicate HIP for linking.
Differential Revision: https://reviews.llv
Author: yaxunl
Date: Mon May 14 12:20:12 2018
New Revision: 332279
URL: http://llvm.org/viewvc/llvm-project?rev=332279&view=rev
Log:
CodeGen: Emit string literal in constant address space
Some targets have constant address space (e.g. amdgcn). For them string literal
should be
emitted in constan
Author: yaxunl
Date: Thu May 17 04:16:35 2018
New Revision: 332593
URL: http://llvm.org/viewvc/llvm-project?rev=332593&view=rev
Log:
CodeGen: Fix invalid bitcast for lifetime.start/end
lifetime.start/end expects pointer argument in alloca address space.
However in C++ a temporary variable is in d
Author: yaxunl
Date: Fri May 18 08:07:56 2018
New Revision: 332724
URL: http://llvm.org/viewvc/llvm-project?rev=332724&view=rev
Log:
[HIP] Support offloading by linker script
To support linking device code in different source files, it is necessary to
embed fat binary at host linking stage.
This
Author: yaxunl
Date: Tue May 22 07:36:26 2018
New Revision: 332982
URL: http://llvm.org/viewvc/llvm-project?rev=332982&view=rev
Log:
Call CreateTempMemWithoutCast for ActiveFlag
Introduced CreateMemTempWithoutCast and CreateTemporaryAllocaWithoutCast to
emit alloca
without casting to default add
Author: yaxunl
Date: Tue May 22 09:13:07 2018
New Revision: 332991
URL: http://llvm.org/viewvc/llvm-project?rev=332991&view=rev
Log:
Revert r332982 Call CreateTempMemWithoutCast for ActiveFlag
Due to regression on arm.
Modified:
cfe/trunk/lib/CodeGen/CGCall.cpp
cfe/trunk/lib/CodeGen/CGCl
Author: yaxunl
Date: Tue May 29 17:49:10 2018
New Revision: 333483
URL: http://llvm.org/viewvc/llvm-project?rev=333483&view=rev
Log:
Add action builder for HIP
To support separate compile/link and linking across device IR in different
source files,
a new HIP action builder is introduced. Basical
Author: yaxunl
Date: Tue May 29 17:53:50 2018
New Revision: 333484
URL: http://llvm.org/viewvc/llvm-project?rev=333484&view=rev
Log:
Add HIP toolchain
This patch adds HIP toolchain to support HIP language mode. It includes:
Create specific compiler jobs for HIP.
Choose specific libraries for HI
Author: yaxunl
Date: Thu Sep 28 12:07:59 2017
New Revision: 314452
URL: http://llvm.org/viewvc/llvm-project?rev=314452&view=rev
Log:
[AMDGPU] Allow flexible register names in inline asm constraints
Currently AMDGPU inline asm only allow v and s as register names in constraints.
This patch allows
Author: yaxunl
Date: Tue Oct 3 07:34:29 2017
New Revision: 314802
URL: http://llvm.org/viewvc/llvm-project?rev=314802&view=rev
Log:
[OpenCL] Fix checking of vector type casting
Currently clang allows the following code
int a;
int b = (const int) a;
However it does not the following code
int4 a
Author: yaxunl
Date: Wed Oct 4 13:32:17 2017
New Revision: 314932
URL: http://llvm.org/viewvc/llvm-project?rev=314932&view=rev
Log:
[OpenCL] Clean up and add missing fields for block struct
Currently block is translated to a structure equivalent to
struct Block {
void *isa;
int flags;
int
Author: yaxunl
Date: Thu Oct 12 20:37:48 2017
New Revision: 315668
URL: http://llvm.org/viewvc/llvm-project?rev=315668&view=rev
Log:
[OpenCL] Add LangAS::opencl_private to represent private address space in AST
Currently Clang uses default address space (0) to represent private address
space for
Author: yaxunl
Date: Fri Oct 13 06:53:06 2017
New Revision: 315678
URL: http://llvm.org/viewvc/llvm-project?rev=315678&view=rev
Log:
Fix regression of test/CodeGenOpenCL/address-spaces.cl on ppc
Modified:
cfe/trunk/test/CodeGenOpenCL/address-spaces.cl
Modified: cfe/trunk/test/CodeGenOpenCL/a
Author: yaxunl
Date: Sat Oct 14 05:23:50 2017
New Revision: 315804
URL: http://llvm.org/viewvc/llvm-project?rev=315804&view=rev
Log:
[OpenCL] Emit enqueued block as kernel
In OpenCL the kernel function and non-kernel function has different calling
conventions.
For certain targets they have diffe
Author: yaxunl
Date: Sat Oct 14 05:51:52 2017
New Revision: 315805
URL: http://llvm.org/viewvc/llvm-project?rev=315805&view=rev
Log:
Fix build failure on android due to missing std::to_string()
Modified:
cfe/trunk/lib/CodeGen/TargetInfo.cpp
Modified: cfe/trunk/lib/CodeGen/TargetInfo.cpp
URL:
Author: yaxunl
Date: Tue Oct 17 07:19:29 2017
New Revision: 316000
URL: http://llvm.org/viewvc/llvm-project?rev=316000&view=rev
Log:
CodeGen: Fix invalid bitcasts for atomic builtins
Currently clang assumes the temporary variables emitted during
codegen of atomic builtins have address space 0, wh
Author: yaxunl
Date: Thu Oct 19 08:56:13 2017
New Revision: 316165
URL: http://llvm.org/viewvc/llvm-project?rev=316165&view=rev
Log:
[AMDGPU] Fix bug in enqueued block codegen due to an extra line
Modified:
cfe/trunk/lib/CodeGen/TargetInfo.cpp
cfe/trunk/test/CodeGenOpenCL/amdgpu-enqueue-k
Author: yaxunl
Date: Mon Oct 23 10:49:26 2017
New Revision: 316353
URL: http://llvm.org/viewvc/llvm-project?rev=316353&view=rev
Log:
CodeGen: Fix invalid bitcast in partial initialization of automatic arrary
variable
Differential Revision: https://reviews.llvm.org/D39184
Modified:
cfe/trunk
Author: yaxunl
Date: Tue Oct 24 12:14:43 2017
New Revision: 316484
URL: http://llvm.org/viewvc/llvm-project?rev=316484&view=rev
Log:
CodeGen: Fix missing debug loc due to alloca
Builder save/restores insertion pointer when emitting addr space cast
for alloca, but does not save/restore debug loc,
Author: yaxunl
Date: Tue Aug 28 14:09:09 2018
New Revision: 340873
URL: http://llvm.org/viewvc/llvm-project?rev=340873&view=rev
Log:
[HIP] Fix output file extension
OffloadBundlingJobAction constructor accepts a list of JobAction as inputs.
The host JobAction is the last one. The file type of Off
Author: yaxunl
Date: Wed Aug 29 13:39:22 2018
New Revision: 340967
URL: http://llvm.org/viewvc/llvm-project?rev=340967&view=rev
Log:
Add predefined macro __gnu_linux__ for proper aux-triple
Clang predefine macro __linx__ for aux-triple with Linux OS
but does not predefine macro __gnu_linux__. Thi
Author: yaxunl
Date: Thu Aug 30 08:10:20 2018
New Revision: 341077
URL: http://llvm.org/viewvc/llvm-project?rev=341077&view=rev
Log:
[HIP] Add -fvisibility hidden option to clang
AMDGPU target need -fvisibility hidden option for clang to
work around a limitation of no PLT support, otherwise there
Author: yaxunl
Date: Fri Jul 20 04:32:51 2018
New Revision: 337540
URL: http://llvm.org/viewvc/llvm-project?rev=337540&view=rev
Log:
Sema: Fix explicit address space cast in C++
Currently clang does not allow implicit cast of a pointer to a pointer type
in different address space but allows C-sty
Author: yaxunl
Date: Fri Jul 20 15:45:24 2018
New Revision: 337631
URL: http://llvm.org/viewvc/llvm-project?rev=337631&view=rev
Log:
[HIP] Register/unregister device fat binary only once
HIP generates one fat binary for all devices after linking. However, for each
compilation
unit a ctor functio
Author: yaxunl
Date: Fri Jul 20 19:02:22 2018
New Revision: 337639
URL: http://llvm.org/viewvc/llvm-project?rev=337639&view=rev
Log:
[HIP] Support -fcuda-flush-denormals-to-zero for amdgcn
Differential Revision: https://reviews.llvm.org/D48287
Modified:
cfe/trunk/include/clang/Basic/LangOpti
Author: yaxunl
Date: Mon Jul 23 18:03:44 2018
New Revision: 337791
URL: http://llvm.org/viewvc/llvm-project?rev=337791&view=rev
Log:
Enable .hip files for test/Driver
Partially revert r334128 due to regressions.
Modified:
cfe/trunk/lib/Driver/Driver.cpp
cfe/trunk/test/Driver/lit.local.cf
Author: yaxunl
Date: Mon Jul 23 18:40:44 2018
New Revision: 337793
URL: http://llvm.org/viewvc/llvm-project?rev=337793&view=rev
Log:
[HIP] pass -target-cpu when running the device-mode compiler
Differential Revision: https://reviews.llvm.org/D49643
Modified:
cfe/trunk/lib/Driver/ToolChains/H
Author: yaxunl
Date: Mon Jul 23 19:12:24 2018
New Revision: 337797
URL: http://llvm.org/viewvc/llvm-project?rev=337797&view=rev
Log:
Attempt to fix regression due to r337791
Modified:
cfe/trunk/test/Driver/hip-toolchain.hip
Modified: cfe/trunk/test/Driver/hip-toolchain.hip
URL:
http://llvm.
Author: yaxunl
Date: Fri Jul 27 20:05:25 2018
New Revision: 338188
URL: http://llvm.org/viewvc/llvm-project?rev=338188&view=rev
Log:
[CUDA][HIP] Allow function-scope static const variable
CUDA 8.0 E.3.9.4 says: Within the body of a __device__ or __global__
function, only __shared__ variables or v
Author: yaxunl
Date: Thu Aug 2 20:18:56 2018
New Revision: 338805
URL: http://llvm.org/viewvc/llvm-project?rev=338805&view=rev
Log:
Sema: Fix explicit address space cast involving void pointers
Explicit cast of a void pointer to a pointer type in different address space is
incorrectly classified
Author: yaxunl
Date: Tue Feb 13 10:01:21 2018
New Revision: 325031
URL: http://llvm.org/viewvc/llvm-project?rev=325031&view=rev
Log:
[AMDGPU] Change constant addr space to 4
Differential Revision: https://reviews.llvm.org/D43171
Added:
cfe/trunk/test/CodeGenOpenCL/amdgpu-env-amdgcn.cl
Remove
Author: yaxunl
Date: Thu Feb 15 08:39:19 2018
New Revision: 325264
URL: http://llvm.org/viewvc/llvm-project?rev=325264&view=rev
Log:
[OpenCL] Fix __enqueue_block for block with captures
The following test case causes issue with codegen of __enqueue_block
void (^block)(void) = ^{ callee(id, out);
Author: yaxunl
Date: Thu Feb 15 11:12:41 2018
New Revision: 325279
URL: http://llvm.org/viewvc/llvm-project?rev=325279&view=rev
Log:
Clean up AMDGCN tests
Differential Revision: https://reviews.llvm.org/D43340
Modified:
cfe/trunk/test/CodeGen/address-space.c
cfe/trunk/test/CodeGenCXX/cxx
Author: yaxunl
Date: Mon Jun 19 12:03:41 2017
New Revision: 305711
URL: http://llvm.org/viewvc/llvm-project?rev=305711&view=rev
Log:
CodeGen: Cast temporary variable to proper address space
In C++ all variables are in default address space. Previously change has been
made to cast automatic variab
Author: yaxunl
Date: Mon Mar 5 09:50:10 2018
New Revision: 326725
URL: http://llvm.org/viewvc/llvm-project?rev=326725&view=rev
Log:
[AMDGPU] Clean up old address space mapping and fix constant address space value
Differential Revision: https://reviews.llvm.org/D43911
Modified:
cfe/trunk/lib
Author: yaxunl
Date: Wed Mar 7 11:32:58 2018
New Revision: 326937
URL: http://llvm.org/viewvc/llvm-project?rev=326937&view=rev
Log:
[OpenCL] Remove block invoke function from emitted block literal struct
OpenCL runtime tracks the invoke function emitted for
any block expression. Due to restricti
Author: yaxunl
Date: Wed Mar 7 13:45:40 2018
New Revision: 326946
URL: http://llvm.org/viewvc/llvm-project?rev=326946&view=rev
Log:
CodeGen: Fix address space of indirect function argument
The indirect function argument is in alloca address space in LLVM IR. However,
during Clang codegen for C++
Author: yaxunl
Date: Wed Mar 14 08:02:28 2018
New Revision: 327515
URL: http://llvm.org/viewvc/llvm-project?rev=327515&view=rev
Log:
CodeGen: Reduce LValue and CallArgList memory footprint before recommitting
r326946
Recent change r326946 (https://reviews.llvm.org/D34367) causes regression in
E
Author: yaxunl
Date: Wed Mar 14 08:03:31 2018
New Revision: 327516
URL: http://llvm.org/viewvc/llvm-project?rev=327516&view=rev
Log:
Add deep AST tree test for r327515
Added:
cfe/trunk/test/CodeGenCXX/deep-ast-tree.cpp
Added: cfe/trunk/test/CodeGenCXX/deep-ast-tree.cpp
URL:
http://llvm.org/
Author: yaxunl
Date: Wed Mar 14 09:47:49 2018
New Revision: 327529
URL: http://llvm.org/viewvc/llvm-project?rev=327529&view=rev
Log:
Attempt to fix failure of deep-ast-tree.cpp on ppc64 and atom
Modified:
cfe/trunk/test/CodeGenCXX/deep-ast-tree.cpp
Modified: cfe/trunk/test/CodeGenCXX/deep-as
Author: yaxunl
Date: Wed Mar 14 11:24:38 2018
New Revision: 327547
URL: http://llvm.org/viewvc/llvm-project?rev=327547&view=rev
Log:
Attempt to fix failure of deep-ast-tree.cpp on atom and s390
Modified:
cfe/trunk/test/CodeGenCXX/deep-ast-tree.cpp
Modified: cfe/trunk/test/CodeGenCXX/deep-ast
Author: yaxunl
Date: Wed Mar 14 13:41:05 2018
New Revision: 327570
URL: http://llvm.org/viewvc/llvm-project?rev=327570&view=rev
Log:
Reduce AST depth for test deep-ast-tree.cpp for atom
Modified:
cfe/trunk/test/CodeGenCXX/deep-ast-tree.cpp
Modified: cfe/trunk/test/CodeGenCXX/deep-ast-tree.cp
Author: yaxunl
Date: Wed Mar 14 14:40:55 2018
New Revision: 327578
URL: http://llvm.org/viewvc/llvm-project?rev=327578&view=rev
Log:
Remove test deep-ast-tree.cpp
Since there is no reliable way to change the AST depth of this test by
supported stack size
of the test environment, remove this test
Author: yaxunl
Date: Thu Mar 15 08:25:19 2018
New Revision: 327634
URL: http://llvm.org/viewvc/llvm-project?rev=327634&view=rev
Log:
Recommit r326946 after reducing CallArgList memory footprint
Added:
cfe/trunk/test/CodeGenCXX/amdgcn-func-arg.cpp
Modified:
cfe/trunk/lib/CodeGen/CGAtomic.c
Author: yaxunl
Date: Tue Jun 5 08:11:02 2018
New Revision: 334021
URL: http://llvm.org/viewvc/llvm-project?rev=334021&view=rev
Log:
[CUDA][HIP] Do not emit type info when compiling for device
CUDA/HIP does not support RTTI on device side, therefore there
is no point of emitting type info when co
Author: yaxunl
Date: Wed Jun 6 12:44:10 2018
New Revision: 334128
URL: http://llvm.org/viewvc/llvm-project?rev=334128&view=rev
Log:
[HIP] Fix unbundling
HIP uses clang-offload-bundler to bundle intermediate files for host
and different gpu archs together. When a file is unbundled,
clang-offload-
Author: yaxunl
Date: Mon Jun 11 17:16:33 2018
New Revision: 334457
URL: http://llvm.org/viewvc/llvm-project?rev=334457&view=rev
Log:
[CUDA][HIP] Set kernel calling convention before arrange function
Currently clang set kernel calling convention for CUDA/HIP after
arranging function, which causes
Author: yaxunl
Date: Tue Jun 12 16:58:59 2018
New Revision: 334561
URL: http://llvm.org/viewvc/llvm-project?rev=334561&view=rev
Log:
[CUDA][HIP] Allow CUDA __global__ functions to have amdgpu kernel attributes
There are HIP applications e.g. Tensorflow 1.3 using amdgpu kernel attributes,
however
Author: yaxunl
Date: Fri Jun 15 08:33:22 2018
New Revision: 334837
URL: http://llvm.org/viewvc/llvm-project?rev=334837&view=rev
Log:
[NFC] Add CreateMemTempWithoutCast and CreateTempAllocaWithoutCast
This is partial re-commit of r332982
Modified:
cfe/trunk/lib/CodeGen/CGCall.cpp
cfe/trun
Author: yaxunl
Date: Fri Jun 15 18:20:52 2018
New Revision: 334879
URL: http://llvm.org/viewvc/llvm-project?rev=334879&view=rev
Log:
Call CreateTempAllocaWithoutCast for ActiveFlag
This is partial re-commit of r332982.
Modified:
cfe/trunk/lib/CodeGen/CGCleanup.cpp
Modified: cfe/trunk/lib/Co
Author: yaxunl
Date: Sat Jun 16 05:28:51 2018
New Revision: 334886
URL: http://llvm.org/viewvc/llvm-project?rev=334886&view=rev
Log:
Add RUN line for amdgcn to lit test conditional-temporaries.cpp
This is partial re-commit of r332982.
Modified:
cfe/trunk/test/CodeGenCXX/conditional-temporari
Author: yaxunl
Date: Tue Oct 9 08:53:14 2018
New Revision: 344057
URL: http://llvm.org/viewvc/llvm-project?rev=344057&view=rev
Log:
[CUDA][HIP] Fix ShouldDeleteSpecialMember for inherited constructors
ShouldDeleteSpecialMember is called upon inherited constructors.
It calls inferCUDATargetForImp
Author: yaxunl
Date: Tue Oct 16 10:36:23 2018
New Revision: 344630
URL: http://llvm.org/viewvc/llvm-project?rev=344630&view=rev
Log:
Disable code object version 3 for HIP toolchain
AMDGPU backend will switch to code object version 3 by default.
Since HIP runtime is not ready, disable it until the
Author: yaxunl
Date: Tue Oct 16 19:32:26 2018
New Revision: 344665
URL: http://llvm.org/viewvc/llvm-project?rev=344665&view=rev
Log:
AMDGPU: add __builtin_amdgcn_update_dpp
Emit llvm.amdgcn.update.dpp for both __builtin_amdgcn_mov_dpp and
__builtin_amdgcn_update_dpp. The first argument to
llvm.am
Author: yaxunl
Date: Mon Oct 22 19:05:31 2018
New Revision: 344996
URL: http://llvm.org/viewvc/llvm-project?rev=344996&view=rev
Log:
Add gfx904 and gfx906 to GPU Arch
Differential Revision: https://reviews.llvm.org/D53472
Modified:
cfe/trunk/include/clang/Basic/Cuda.h
cfe/trunk/lib/Basic
Author: yaxunl
Date: Tue Jan 29 05:20:23 2019
New Revision: 352488
URL: http://llvm.org/viewvc/llvm-project?rev=352488&view=rev
Log:
[CUDA][HIP] Do not diagnose use of _Float16
r352221 caused regressions in CUDA/HIP since device function may use _Float16
whereas host does not support it.
In this
Author: yaxunl
Date: Wed Jan 30 04:26:54 2019
New Revision: 352620
URL: http://llvm.org/viewvc/llvm-project?rev=352620&view=rev
Log:
[HIP] Fix size_t for MSVC environment
In 64 bit MSVC environment size_t is defined as unsigned long long.
In single source language like HIP, data layout should be
Author: yaxunl
Date: Thu Nov 8 08:55:46 2018
New Revision: 346413
URL: http://llvm.org/viewvc/llvm-project?rev=346413&view=rev
Log:
Fix bitcast to address space cast for coerced load/stores
Coerced load/stores through memory do not take into account potential
address space differences when it c
Author: yaxunl
Date: Fri Nov 9 10:52:05 2018
New Revision: 346536
URL: http://llvm.org/viewvc/llvm-project?rev=346536&view=rev
Log:
[HIP] Remove useless sections in linked files
clang-offload-bundler creates __CLANG_OFFLOAD_BUNDLE__* sections in the bundles,
which get into the linked files. Thes
Author: yaxunl
Date: Tue Nov 13 20:47:31 2018
New Revision: 346828
URL: http://llvm.org/viewvc/llvm-project?rev=346828&view=rev
Log:
[HIP] Fix device only compilation
Fix a bug causing host code being compiled when --cude-device-only is set.
Differential Revision: https://reviews.llvm.org/D54496
Author: yaxunl
Date: Tue Feb 26 08:20:41 2019
New Revision: 354893
URL: http://llvm.org/viewvc/llvm-project?rev=354893&view=rev
Log:
[OpenCL] Fix assertion due to blocks
A recent change caused assertion in CodeGenFunction::EmitBlockCallExpr when a
block is called.
There is code
Func = CGM.ge
Author: yaxunl
Date: Tue Feb 26 14:24:49 2019
New Revision: 354929
URL: http://llvm.org/viewvc/llvm-project?rev=354929&view=rev
Log:
[CUDA][HIP] Check calling convention based on function target
MSVC header files using vectorcall to differentiate overloaded functions, which
causes failure for AMD
Author: yaxunl
Date: Tue Feb 26 18:02:52 2019
New Revision: 354948
URL: http://llvm.org/viewvc/llvm-project?rev=354948&view=rev
Log:
[HIP] change kernel stub name
Add .stub to kernel stub function name so that it is different from kernel
name in device code. This is necessary to let debugger find
Author: yaxunl
Date: Wed Feb 27 07:46:29 2019
New Revision: 354990
URL: http://llvm.org/viewvc/llvm-project?rev=354990&view=rev
Log:
[NFC] minor revision of r354929 [CUDA][HIP] Check calling convention based on
function target
Add comments and move a variable to if block.
Differential Revision:
Author: yaxunl
Date: Thu Feb 28 09:08:26 2019
New Revision: 355106
URL: http://llvm.org/viewvc/llvm-project?rev=355106&view=rev
Log:
Partial revert of r353952: [HIP] Handle compile -m options and propagate into
LLC
Remove comments and tests about passing -mcode-object-v3 to driver since it does
Author: yaxunl
Date: Tue Mar 5 08:07:56 2019
New Revision: 355410
URL: http://llvm.org/viewvc/llvm-project?rev=355410&view=rev
Log:
[HIP] Do not unbundle object files for -fno-gpu-rdc
When -fno-gpu-rdc is set, device code is compiled, linked, and assembled into
fat binary
and embedded as string
Author: yaxunl
Date: Tue Mar 5 09:52:32 2019
New Revision: 355419
URL: http://llvm.org/viewvc/llvm-project?rev=355419&view=rev
Log:
Allow bundle size to be 0 in clang-offload-bundler
HIP uses clang-offload-bundler to create fat binary. The bundle for host is
empty.
Currently clang-offload-bundl
Author: yaxunl
Date: Tue Mar 5 10:19:35 2019
New Revision: 355421
URL: http://llvm.org/viewvc/llvm-project?rev=355421&view=rev
Log:
[CUDA][HIP][Sema] Fix template kernel with function as template parameter
If a kernel template has a function as its template parameter, a device
function should b
Author: yaxunl
Date: Fri Apr 12 09:23:31 2019
New Revision: 358290
URL: http://llvm.org/viewvc/llvm-project?rev=358290&view=rev
Log:
[HIP] Use -mlink-builtin-bitcode to link device library
Use -mlink-builtin-bitcode instead of llvm-link to link
device library so that device library bitcode and us
Author: yaxunl
Date: Tue Apr 30 11:35:37 2019
New Revision: 359594
URL: http://llvm.org/viewvc/llvm-project?rev=359594&view=rev
Log:
AMDGPU: Enable _Float16
Added:
cfe/trunk/test/CodeGenCXX/amdgpu-float16.cpp
Modified:
cfe/trunk/lib/Basic/Targets/AMDGPU.cpp
Modified: cfe/trunk/lib/Basic/
Author: yaxunl
Date: Tue Apr 30 12:06:15 2019
New Revision: 359598
URL: http://llvm.org/viewvc/llvm-project?rev=359598&view=rev
Log:
Add requires amdgpu-registered-target for amdgpu-float16.cpp
Modified:
cfe/trunk/test/CodeGenCXX/amdgpu-float16.cpp
Modified: cfe/trunk/test/CodeGenCXX/amdgpu-
Author: yaxunl
Date: Tue Oct 2 10:48:54 2018
New Revision: 343611
URL: http://llvm.org/viewvc/llvm-project?rev=343611&view=rev
Log:
[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. Whe
Author: yaxunl
Date: Thu Jan 31 13:57:51 2019
New Revision: 352801
URL: http://llvm.org/viewvc/llvm-project?rev=352801&view=rev
Log:
Do not copy long double and 128-bit fp format from aux target for AMDGPU
rC352620 caused regressions because it copied floating point format from
aux target.
float
Author: yaxunl
Date: Wed Feb 13 18:00:09 2019
New Revision: 354004
URL: http://llvm.org/viewvc/llvm-project?rev=354004&view=rev
Log:
[CUDA][HIP] Use device side kernel and variable names when registering them
__hipRegisterFunction and __hipRegisterVar need to accept device side kernel
and variab
Author: yaxunl
Date: Thu Feb 21 12:12:16 2019
New Revision: 354615
URL: http://llvm.org/viewvc/llvm-project?rev=354615&view=rev
Log:
[HIP] change kernel stub name
Add .stub to kernel stub function name so that it is different from kernel
name in device code. This is necessary to let debugger find
Author: yaxunl
Date: Thu Feb 21 20:20:12 2019
New Revision: 354651
URL: http://llvm.org/viewvc/llvm-project?rev=354651&view=rev
Log:
revert r354615: [HIP] change kernel stub name
It caused regressions.
Differential Revision: https://reviews.llvm.org/D58518
Modified:
cfe/trunk/lib/CodeGen/CG
Author: yaxunl
Date: Sun Sep 3 06:52:24 2017
New Revision: 312441
URL: http://llvm.org/viewvc/llvm-project?rev=312441&view=rev
Log:
[OpenCL] Do not use vararg in emitted functions for enqueue_kernel
Not all targets support vararg (e.g. amdgpu). Instead of using vararg in the
emitted functions f
Author: yaxunl
Date: Wed Sep 13 11:50:42 2017
New Revision: 313171
URL: http://llvm.org/viewvc/llvm-project?rev=313171&view=rev
Log:
[AMDGPU] Change addr space of clk_event_t, queue_t and reserve_id_t to global
Differential Revision: https://reviews.llvm.org/D37703
Modified:
cfe/trunk/lib/Ba
Author: yaxunl
Date: Wed Sep 13 11:56:25 2017
New Revision: 313172
URL: http://llvm.org/viewvc/llvm-project?rev=313172&view=rev
Log:
Add more tests for OpenCL atomic builtin functions
Add tests for different address spaces and insert some blank lines to make them
more readable.
Differential Rev
Author: yaxunl
Date: Tue Jun 11 08:05:11 2019
New Revision: 363076
URL: http://llvm.org/viewvc/llvm-project?rev=363076&view=rev
Log:
Revert r344630 Disable code object version 3 for HIP toolchain.
Remove the workaround so that by default code object v3 is enabled.
Modified:
cfe/trunk/lib/Dri
Author: yaxunl
Date: Fri Jun 14 08:54:47 2019
New Revision: 363414
URL: http://llvm.org/viewvc/llvm-project?rev=363414&view=rev
Log:
[AMDGPU] Enable the implicit arguments for HIP (CLANG)
Enable 48-bytes of implicit arguments for HIP as well. Earlier it was enabled
for OpenCL. This code is speci
Author: yaxunl
Date: Tue Jun 25 20:33:03 2019
New Revision: 364380
URL: http://llvm.org/viewvc/llvm-project?rev=364380&view=rev
Log:
Fix build failure due to missing break
Modified:
cfe/trunk/lib/Basic/Targets/ARM.cpp
Modified: cfe/trunk/lib/Basic/Targets/ARM.cpp
URL:
http://llvm.org/viewvc
Author: yaxunl
Date: Tue Jun 25 20:47:37 2019
New Revision: 364381
URL: http://llvm.org/viewvc/llvm-project?rev=364381&view=rev
Log:
[HIP] Support attribute hip_pinned_shadow
This patch introduces support of hip_pinned_shadow variable for HIP.
A hip_pinned_shadow variable is a global variable wi
Author: yaxunl
Date: Thu Jul 11 10:50:09 2019
New Revision: 365799
URL: http://llvm.org/viewvc/llvm-project?rev=365799&view=rev
Log:
[HIP] Add GPU arch gfx1010, gfx1011, and gfx1012
Differential Revision: https://reviews.llvm.org/D64364
Modified:
cfe/trunk/include/clang/Basic/Cuda.h
cfe/
Author: yaxunl
Date: Mon May 27 04:19:07 2019
New Revision: 361757
URL: http://llvm.org/viewvc/llvm-project?rev=361757&view=rev
Log:
[OpenCL] Fix file-scope const sampler variable for 2.0
OpenCL spec v2.0 s6.13.14:
Samplers can also be declared as global constants in the program
source using the
Author: yaxunl
Date: Tue May 28 14:18:59 2019
New Revision: 361880
URL: http://llvm.org/viewvc/llvm-project?rev=361880&view=rev
Log:
[CUDA][HIP] Emit dependent libs for host only
Recently D60274 was introduced to allow lld to handle dependent libs. However
current
usage of dependent libs (e.g. p
Author: yaxunl
Date: Tue May 28 18:34:44 2019
New Revision: 361905
URL: http://llvm.org/viewvc/llvm-project?rev=361905&view=rev
Log:
Fix failure of lit test dependent-libs.cu
Modified:
cfe/trunk/test/CodeGenCUDA/dependent-libs.cu
Modified: cfe/trunk/test/CodeGenCUDA/dependent-libs.cu
URL:
h
Author: yaxunl
Date: Thu Jan 10 12:09:52 2019
New Revision: 350885
URL: http://llvm.org/viewvc/llvm-project?rev=350885&view=rev
Log:
[HIP] Use nul instead of /dev/null when running on windows
When clang is running on windows, /dev/null is not available. Use nul as empty
input file instead.
Diff
Author: yaxunl
Date: Fri Sep 20 07:28:09 2019
New Revision: 372394
URL: http://llvm.org/viewvc/llvm-project?rev=372394&view=rev
Log:
[CUDA][HIP] Fix hostness of defaulted constructor
Clang does not respect the explicit device host attributes of defaulted special
members.
Also clang does not respe
Author: yaxunl
Date: Fri Sep 20 19:51:44 2019
New Revision: 372452
URL: http://llvm.org/viewvc/llvm-project?rev=372452&view=rev
Log:
Revert assertion added by r372394
The assertion added by r372394 causes CUDA test in test-suite to assert.
The assertion was not there originally, so revert it.
Author: yaxunl
Date: Tue Sep 24 12:16:40 2019
New Revision: 372773
URL: http://llvm.org/viewvc/llvm-project?rev=372773&view=rev
Log:
[HIP] Support new kernel launching API
Differential Revision: https://reviews.llvm.org/D67947
Modified:
cfe/trunk/include/clang/Basic/LangOptions.def
cfe/t
Author: yaxunl
Date: Wed Oct 2 20:27:43 2019
New Revision: 373561
URL: http://llvm.org/viewvc/llvm-project?rev=373561&view=rev
Log:
[HIP] Support -emit-llvm for device compilation
Sometimes it is useful to compile HIP device code to LLVM BC. It is not
convenient to use clang -cc1 since
there ar
Author: yaxunl
Date: Thu Oct 3 11:59:56 2019
New Revision: 373649
URL: http://llvm.org/viewvc/llvm-project?rev=373649&view=rev
Log:
[HIP] Use option -nogpulib to disable linking device lib
Differential Revision: https://reviews.llvm.org/D68300
Added:
cfe/trunk/test/Driver/hip-no-device-libs
Author: yaxunl
Date: Wed Apr 18 11:25:03 2018
New Revision: 330279
URL: http://llvm.org/viewvc/llvm-project?rev=330279&view=rev
Log:
[HIP] Add driver input type for HIP
Patch by Greg Rodgers.
Revised by Yaxun Liu.
Differential Revision: https://reviews.llvm.org/D45489
Modified:
cfe/trunk/in
Author: yaxunl
Date: Fri Apr 20 10:01:03 2018
New Revision: 330447
URL: http://llvm.org/viewvc/llvm-project?rev=330447&view=rev
Log:
[CUDA] Set LLVM calling convention for CUDA kernel
Some targets need special LLVM calling convention for CUDA kernel.
This patch does that through a TargetCodeGenIn
1 - 100 of 1625 matches
Mail list logo
