SWDEV-85602 - rename hsail-64 arch to hsail64
This is to match other existing llvm targets, such as spir64 and amdil64, as well as to match behavior of open source HSAIL BE.
For legacy users there is alias "-hsail-64" provided in the aoc2 only.
Testing: smoke, precheckin
Reviewed by Matthew Arsenault, Evgeny Mankov and Nikolay Haustov
Affected files ...
... //depot/stg/opencl/drivers/opencl/compiler/legacy-lib/backends/common/codegen.cpp#2 edit
... //depot/stg/opencl/drivers/opencl/compiler/legacy-lib/backends/common/frontend.cpp#4 edit
... //depot/stg/opencl/drivers/opencl/compiler/legacy-lib/utils/v0_8/target_mappings.h#5 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/backends/common/codegen.cpp#66 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/backends/common/frontend.cpp#37 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/utils/v0_8/target_mappings.h#37 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/lib/Target/HSAIL/HSAILTargetMachine.cpp#53 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/lib/Target/HSAIL/TargetInfo/HSAILTargetInfo.cpp#6 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/tools/aacl/aa.h#2 edit
... //depot/stg/opencl/drivers/opencl/compiler/tools/aoc2/aoc2.cpp#80 edit
... //depot/stg/opencl/drivers/opencl/library/hsa/amp_libm/build/Makefile.amp_libm#4 edit
... //depot/stg/opencl/drivers/opencl/library/hsa/gcn/build/Makefile.gcn#20 edit
... //depot/stg/opencl/drivers/opencl/library/hsa/gcndev/build/Makefile.gcndev#3 edit
... //depot/stg/opencl/drivers/opencl/library/hsa/hsail/build/Makefile.hsail#44 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpucompiler.cpp#153 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuprogram.cpp#222 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/hsa_foundation/hsacompiler.cpp#5 edit
... //depot/stg/opencl/drivers/opencl/tests/hsa/bin/test_driver.pl#25 edit
... //depot/stg/opencl/drivers/opencl/tests/hsa/tlst/complib.tlst#21 edit
... //depot/stg/opencl/drivers/opencl/tests/hsa/tlst/ocl_debug.tlst#9 edit
... //depot/stg/opencl/drivers/opencl/tests/hsa/tlst/ocl_regression.tlst#25 edit
... //depot/stg/opencl/drivers/opencl/tests/ocltst/module/spir/SPIRBase.cpp#3 edit
... //depot/stg/opencl/drivers/opencl/tests/ocltst/module/spir/SPIRVBasic.cpp#10 edit
... //depot/stg/opencl/drivers/opencl/tests/ocltst/module/spir/SPIRVDropIn.cpp#5 edit
[ROCm/clr commit: 14179df14f]
SWDEV-83192 - [CQE OCL][2.0][QR] Few MCLSubtests are falling due to faulty cl#1208929
[Synopsis]
Failure is because of additional check on compilation options existence in binary, which are expected in a separate section in the input binary. Options are expected on HSAIL path in order to make a decision whether to recompile or not the input code. AMDIL path has another strategy for recompilation algorithm and doesn't need options to be obligatory presented in input binary.
[Solution]
Remove options existence check in the input binary when compiling from BINARY_TYPE_ELF.
P.S.
With the fix HSAIL path also doesnt require compilation options existence in input binary anymore. Recompilation will be done silently in such case.
[Testing]
Pre checkin, MCL
Reviewed by Evgeniy Mankov, Nikolay Haustov, German Andryeyev
http://ocltc.amd.com/reviews/r/9560/
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/platform/program.cpp#77 edit
[ROCm/clr commit: 91bbd98a0b]
SWDEV-2 - Change OpenCL version number from 2009 to 2010.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1756 edit
[ROCm/clr commit: 84ed48b665]
SWDEV-79308 - Resubmit of CL1228064 with restriction of mininum scratch buffer size of 64K if a scratch buffer is needed.
Reduce the total scratch buffer size by a factor of 4, which in effect reducing the max. scratch waves from 32 to 8, to avoid the required total scratch buffer size exceeds the available local memory.
Made sure the scratch buffer size is aligned with 64K boundary
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuresource.cpp#235 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gslbe/src/rt/GSLDevice.cpp#156 edit
[ROCm/clr commit: db94c3fb66]
SWDEV-2 - Change OpenCL version number from 2008 to 2009.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1755 edit
[ROCm/clr commit: 44e689cfec]
SWDEV-2 - Change OpenCL version number from 2007 to 2008.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1754 edit
[ROCm/clr commit: cdffb6c9aa]
SWDEV-2 - Change OpenCL version number from 2006 to 2007.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1753 edit
[ROCm/clr commit: 269e7c74c5]
SWDEV-77172 - Choose isa handle on CZ based on whether SVM is supported or not as SVM may need additional SC Options to be passed as default.
ReviewBoardURL = http://ocltc.amd.com/reviews/r/9531/diff/
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuprogram.cpp#220 edit
[ROCm/clr commit: 3345a97263]
SWDEV-79308 - Resubmit of CL1226881 with the fix of the SC sanity check issue. Reduce the total scratch buffer size by a factor of 4, which in effect reducing the max. scratch waves from 32 to 8, to avoid the required total scratch buffer
size exceeds the available local memory.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuresource.cpp#233 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gslbe/src/rt/GSLDevice.cpp#154 edit
[ROCm/clr commit: 685d1164ee]
SWDEV-2 - Change OpenCL version number from 2005 to 2006.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1752 edit
[ROCm/clr commit: a56e584fed]
SWDEV-2 - Change OpenCL version number from 2004 to 2005.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1751 edit
[ROCm/clr commit: cfe171ecaa]
SWDEV-77172 - Disable ThreadTrace on SVM as its causing hang until a solution can be found.
ReviewBoardURL = http://ocltc.amd.com/reviews/r/9502/diff/
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpusettings.cpp#341 edit
[ROCm/clr commit: ec51f85c8e]
SWDEV-2 - Change OpenCL version number from 2003 to 2004.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1750 edit
[ROCm/clr commit: d9e28537b3]
SWDEV-79308 - Reduce the total scratch buffer size by a factor of 4, which in effect reducing the max. scratch waves from 32 to 8, to avoid the required total scratch buffer
size exceeds the available local memory.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gslbe/src/rt/GSLDevice.cpp#152 edit
[ROCm/clr commit: a319c0e35c]
SWDEV-2 - Change OpenCL version number from 2002 to 2003.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1749 edit
[ROCm/clr commit: 85f5db0446]
SWDEV-2 - Change OpenCL version number from 2001 to 2002.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1748 edit
[ROCm/clr commit: 8108ce0727]
SWDEV-2 - Change OpenCL version number from 2000 to 2001.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1747 edit
[ROCm/clr commit: d74f5269e6]
SWDEV-2 - Change OpenCL version number from 1999 to 2000.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1746 edit
[ROCm/clr commit: c6ee75d740]
SWDEV-84309 - Using agpMemAvailableCacheableBytes instead of agpMemAvailableBytes when calculating free memory for viPlus_ apu.
When memory allocation is in system memory, only agpMemAvailableCacheableBytes is changed
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpudevice.cpp#539 edit
[ROCm/clr commit: bc822d60d1]
SWDEV-2 - Change OpenCL version number from 1998 to 1999.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1745 edit
[ROCm/clr commit: e84d9750d1]
SWDEV-80364 - [Project Brahma] Disable OCL2.0
Disable OCL2.0 on Brahma because base driver doesn't have enough resource to support it.
Note: the implementation for OCL2.0 on Brahma in OCL runtime has been done.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/flags.hpp#251 edit
[ROCm/clr commit: 8a320baf8c]
SWDEV-2 - Change OpenCL version number from 1997 to 1998.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1744 edit
[ROCm/clr commit: a679cea1c3]
SWDEV-2 - Change OpenCL version number from 1996 to 1997.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1743 edit
[ROCm/clr commit: 0cf937dfc1]
SWDEV-2 - Change OpenCL version number from 1995 to 1996.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1742 edit
[ROCm/clr commit: 6245297804]
SWDEV-78467 - Adding 'CL_FILE_SIZE_AMD' queriable property for file objects.
While validating the LF extension we noticed a common use case in which the programmer has to query the size of the file being processed.
This enhancement will save the programming overhead of opening and closing files using some other API just in order to query their size.
This feature is being integrated from LF dev branch 15.30.1023.
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/opencl/amdocl/cl_lqdflash_amd.cpp#12 edit
... //depot/stg/opencl/drivers/opencl/api/opencl/khronos/headers/opencl2.0/CL/cl_ext.h#23 edit
... //depot/stg/opencl/drivers/opencl/runtime/platform/memory.hpp#95 integrate
[ROCm/clr commit: ab17489071]
SWDEV-80174 - changed the flag settting, so that the env variable OCL_SET_SVM_SIZE can be used on release version of opencl, not just for debug version.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/flags.hpp#250 edit
[ROCm/clr commit: 232af87dbf]
SWDEV-2 - Change OpenCL version number from 1994 to 1995.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1741 edit
[ROCm/clr commit: 110e10c928]
SWDEV-2 - Change OpenCL version number from 1993 to 1994.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1740 edit
[ROCm/clr commit: 2192a538fa]
SWDEV-2 - Change OpenCL version number from 1992 to 1993.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1739 edit
[ROCm/clr commit: e5c218c895]
SWDEV-2 - Change OpenCL version number from 1991 to 1992.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1738 edit
[ROCm/clr commit: 9f8760555a]
SWDEV-2 - Change OpenCL version number from 1990 to 1991.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1737 edit
[ROCm/clr commit: cd4fcd1afe]
SWDEV-2 - Change OpenCL version number from 1989 to 1990.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1736 edit
[ROCm/clr commit: 5ccd02e65b]
SWDEV-2 - Change OpenCL version number from 1988 to 1989.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1735 edit
[ROCm/clr commit: bf0f6b71bf]
SWDEV-77584 - Fix ocl_conformance compiler failures.
1. If compiling in debug mode, linkImpl wasn't called and kernelNames wasn't set
which led to CL_INVALID_KERNEL_NAME errors in debug configs in TeamCity. Looking at AMDIL
code, there is no reason to skip linkImpl in debug mode.
2. Set types to TYPE_LIBRARY/TYPE_EXECUTABLE. This fixes ocl_conformance compiler program_binary_type.
Reviewed by: Evgeniy Mankov
Testing: smoke, pre-checkin, OCLSeparateCompile.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuprogram.cpp#218 edit
[ROCm/clr commit: b4f5366fe4]
SWDEV-2 - Change OpenCL version number from 1987 to 1988.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1734 edit
[ROCm/clr commit: cfc2b489df]
SWDEV-2 - Change OpenCL version number from 1986 to 1987.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1733 edit
[ROCm/clr commit: 422318b9aa]
SWDEV-2 - Change OpenCL version number from 1985 to 1986.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1732 edit
[ROCm/clr commit: 32568eec9b]
SWDEV-2 - Change OpenCL version number from 1984 to 1985.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1731 edit
[ROCm/clr commit: 4e074c4044]
SWDEV-2 - Change OpenCL version number from 1983 to 1984.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1730 edit
[ROCm/clr commit: fe72e17898]
SWDEV-66693 - OpenCL Runtime HW Debug support development - use flag, instead of getenv() call, in IOL to indicate the enablement of HW Debug.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gslbe/src/rt/GSLDevice.cpp#151 edit
[ROCm/clr commit: f15514da95]
SWDEV-80173 - Switching HSAIL on OpenCL stack by default.
1. HSAIL becomes default backend for OpenCL (all versions) on all GFX7+ (CI+) devices.
2. AMDIL stays default on pre CI+ devices.
3. AMDIL is forced implicitly:
a) for pre CI+ devices;
b) if clc++ extension is used: "-x -clc++";
c) if input IL is SPIR: "-x -spir";
4. To force AMDIL explicitly "-legacy" option should be added to AMD_OCL_BUILD_OPTIONS or AMD_OCL_BUILD_OPTIONS_APPEND environment variable.
5. SPIRV always goes through HSAIL and couldn't be forced to AMDIL explicitly or implicitly.
6. Compilation of Blit kernels couldn't be forced to AMDIL too, they are compiled via HSAIL on GFX7+, otherwise - via AMDIL. The logic is almost the same as it is currently.
7. test_driver.pl is switched to OpenCL 1.2/HSAIL by default. Previously OpenCL 2.0/HSAIL was by default and all OpenCL 1.2 tests went through AMDIL.
[Testing]
offline: weekly, smoke
online: bootleg, ocltst, ocl conformance 1.2/2.0, AMD SDK 3.0, Blender 2.76b, performance testing, manual testing on CI+ devices with forcing AMDIL "-legacy".
pre-checkin:
http://ocltc.amd.com:8111/viewModification.html?modId=64664&personal=true&buildTypeId=&tab=vcsModificationBuilds&show_all_builds=true
[Reviewers]
Brian Sumner, Stanislav Mekhanoshin, German Andryeyev, Nikolay Haustov
http://ocltc.amd.com/reviews/r/8850
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/flags.hpp#248 edit
... //depot/stg/opencl/drivers/opencl/tests/hsa/bin/test_driver.pl#24 edit
[ROCm/clr commit: b01869b54b]
SWDEV-2 - Change OpenCL version number from 1982 to 1983.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1729 edit
[ROCm/clr commit: 719b49cde5]
SWDEV-80173 - remove the opt level adjustment based on the function size
Since it only limits the llvm backend, but not optimizer or SC it really hurts the compile time.
A huge unoptimized hsail after inlining and no backend optimizations goes to the SC with overwhelming results.
A proper way of addressing these issues is to keep inlining reasonable.
Testing: smoke, precheckin
Reviewed by Evgeny Mankov
Affected files ...
... //depot/stg/opencl/drivers/opencl/compiler/lib/backends/common/codegen.cpp#65 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/lib/Transforms/IPO/AMDInline.cpp#3 edit
[ROCm/clr commit: 13f385908c]
foreman