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
SWDEV-2 - Change OpenCL version number from 1982 to 1983.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1729 edit
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
SWDEV-2 - Change OpenCL version number from 1981 to 1982.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1728 edit
SWDEV-2 - Change OpenCL version number from 1980 to 1981.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1727 edit
SWDEV-2 - Change OpenCL version number from 1979 to 1980.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1726 edit
SWDEV-2 - Change OpenCL version number from 1978 to 1979.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1725 edit
SWDEV-80173 - HSA HLC: disable liveness analysis and jump threading
After the investigation I have found liveness analysis never changed code generation in any of the benchmarks or applications.
Its only use is in the LICM and the hoisting limitation was never really triggered.
Since the analysis is very expensive I'm disabling it.
The jump threading is generally bad on the GPU because it creates unstructured control flow.
Even if hsail might become smaller and have less branches, it does not help because finalzier's structurizer will have to clone blocks.
Jump threading is disabled for GPU. This improves compilation speed and just slightly improves performance.
Testing: smoke, precheckin, vray and blender compilation
Reviewed by Daniil Fukalov
Affected files ...
... //depot/stg/opencl/drivers/opencl/compiler/lib/utils/OPTIONS.def#133 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/include/llvm/AMDLLVMContextHook.h#29 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/lib/Analysis/AMDLiveAnalysis.cpp#32 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/lib/Transforms/IPO/AMDPassManagerBuilder.cpp#61 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/tools/opt/amdopt.inc#29 edit
SWDEV-2 - Change OpenCL version number from 1977 to 1978.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1724 edit
SWDEV-2 - Change OpenCL version number from 1976 to 1977.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1723 edit
SWDEV-2 - Change OpenCL version number from 1975 to 1976.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1722 edit
SWDEV-83467 - [SPIRV] Add support of SPIRV to CPU
Modifying runtime and compile time to allow SPIRV binaries to run on CPU since it only runs on HSAIL GPU
Added changes to allow conversion of CPU's llvmBinaryIsSpir boolean into compiler library's oclElfSections enum
Cpuprogram.cpp's llvmBinaryIsSpir flag renamed to elfSectionType will now support LLVMIR, SPIR, and SPIRV
Added SPIRV to compiler lib's elf as new oclElfSections enum
cpuprogram.cpp changes also made to gpuprogram.cpp's NullProgram to allow compilation
Affected files ...
... //depot/stg/opencl/drivers/opencl/compiler/lib/loaders/elf/elf.cpp#33 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/loaders/elf/elf.hpp#22 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/cpu/cpuprogram.cpp#69 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/device.cpp#191 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/device.hpp#266 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpucompiler.cpp#152 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuprogram.cpp#217 edit
SWDEV-2 - Change OpenCL version number from 1974 to 1975.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1721 edit
SWDEV-2 - Change OpenCL version number from 1973 to 1974.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1720 edit
SWDEV-2 - Change OpenCL version number from 1972 to 1973.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1719 edit
SWDEV-2 - Change OpenCL version number from 1971 to 1972.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1718 edit
SWDEV-2 - Change OpenCL version number from 1970 to 1971.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1717 edit
SWDEV-82296 - [CQE OCL][2.0][HWSC][16.10]SDK Sample "AtomicCounters" 32/64bit failed with HWSC driver
Disabling the cl_ext_atomic_counters_32 extension since there is no support for this extension on HSAIL and HWS.
ReviewBoardURL = http://ocltc.amd.com/reviews/r/9221/
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/device.hpp#265 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpudevice.cpp#538 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpusettings.cpp#338 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/hsa_foundation/hsasettings.cpp#11 edit
SWDEV-2 - Change OpenCL version number from 1969 to 1970.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1716 edit
SWDEV-2 - Change OpenCL version number from 1968 to 1969.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1715 edit
SWDEV-2 - Change OpenCL version number from 1967 to 1968.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1714 edit
SWDEV-83242 - RT: disable ADL reload in app detect
Reload is disabled in ADL with the change list 1198904 and ticket
SWDEV-59442 - The ADL_ApplicationProfiles_System_Reload Function is not Re-entrant.
Returned value of the call is ADL_ERR_NOT_SUPPORTED on Windows.
Thus appprofiles do not currently work with the OpenCL.
This change removes reload call result check and profiles work again.
Testing: smoke, precheckin, use of customized app profile
Reviewed by German Andreev
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/appprofile.cpp#11 edit
SWDEV-2 - Change OpenCL version number from 1966 to 1967.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1713 edit
SWDEV-2 - Change OpenCL version number from 1965 to 1966.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1712 edit
SWDEV-2 - Change OpenCL version number from 1964 to 1965.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1711 edit
SWDEV-2 - Change OpenCL version number from 1963 to 1964.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1710 edit
SWDEV-2 - Change OpenCL version number from 1962 to 1963.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1709 edit
SWDEV-2 - Change OpenCL version number from 1961 to 1962.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1708 edit
SWDEV-82256 - Limit the workaround for Win 7 only because KMD has fixed TDR issue on Win 8.1/10
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpusettings.cpp#336 edit
SWDEV-2 - Change OpenCL version number from 1960 to 1961.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1707 edit
SWDEV-2 - Change OpenCL version number from 1959 to 1960.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1706 edit
SWDEV-2 - Change OpenCL version number from 1958 to 1959.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1705 edit
SWDEV-2 - Change OpenCL version number from 1957 to 1958.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1704 edit
SWDEV-82205 - Increased workloard to pass this test.
- This is workaround because KMD don't have solution to fix TDR issue yet in 15.30.
- This workaround including CL#1201765 should be reverted once KMD has a fix
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpusettings.cpp#335 edit
SWDEV-2 - Change OpenCL version number from 1956 to 1957.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1703 edit
SWDEV-2 - Change OpenCL version number from 1955 to 1956.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1702 edit
SWDEV-82596 - HSA HLC: Create AMDInline pass
The generic llvm inlining heuristcs do not work well for GPU.
In particular we have a common problem in several tests:
If we have a pointer to private array passed into a function it will not be optimized out, leaving scratch usage.
The pass increases the inline threshold to allow inliniting in this case.
Also that we can move at least some AMD inlining customizations into this file from the common code.
Inline hint threshold is moved in this change.
Performance impact on ocltst sha256, 32 bit, Fiji:
AMDIL HSAIL Diff HSAIL+Inliner Diff Diff
before to AMDIL to HSAIL to AMDIL
OCLPerfSHA256[ 0] 43.843 40.894 0.93 69.910 1.71 1.59
OCLPerfSHA256[ 1] 53.611 51.083 0.95 80.919 1.58 1.51
OCLPerfSHA256[ 2] 52.127 51.528 0.99 80.640 1.56 1.55
OCLPerfSHA256[ 3] 60.952 57.027 0.94 68.615 1.20 1.13
OCLPerfSHA256[ 4] 76.173 70.150 0.92 80.582 1.15 1.06
OCLPerfSHA256[ 5] 75.886 70.264 0.93 81.000 1.15 1.07
Testing: smoke, precheckin, ocltst sha256
Reviewed by Danill Fukalov
Affected files ...
... //depot/stg/opencl/drivers/opencl/compiler/lib/backends/common/opt_level.cpp#28 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/include/llvm/InitializePasses.h#93 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/include/llvm/LinkAllPasses.h#49 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/include/llvm/Transforms/IPO.h#32 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/lib/Transforms/IPO/AMDInline.cpp#1 add
... //depot/stg/opencl/drivers/opencl/compiler/llvm/lib/Transforms/IPO/CMakeLists.txt#24 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/lib/Transforms/IPO/IPO.cpp#32 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/lib/Transforms/IPO/Inliner.cpp#42 edit
... //depot/stg/opencl/drivers/opencl/compiler/llvm/tools/opt/amdopt.inc#28 edit
SWDEV-2 - Change OpenCL version number from 1954 to 1955.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1701 edit
SWDEV-2 - Change OpenCL version number from 1953 to 1954.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1700 edit
SWDEV-82054 - [CQE OCL][QR][LNX] RQ Conformance "Integer_Ops" test is crashing on CPU; Faulty CL#1206023.
In llvm32 by default llvm::DisablePrettyStackTrace is off, which causes a trap handler installed by default and interferes with the trap handler in runtime, causing unhandled SIGFPE exceptions when executing conformance/integer_ops on certain cpu.
To fix this, put stack trace duping under an env var AMD_DUMP_STACK_TRACE and set llvm::DisablePrettyStackTrace=true by default.
Here env var is used because there is still no elf binary at this stage to pass the compiler option to if_aclCompilerInit.
Affected files ...
... //depot/stg/opencl/drivers/opencl/compiler/legacy-lib/backends/common/v0_8/if_acl.cpp#8 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/backends/common/v0_8/if_acl.cpp#86 edit
SWDEV-2 - Change OpenCL version number from 1952 to 1953.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1699 edit
SWDEV-2 - Change OpenCL version number from 1951 to 1952.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1698 edit
foreman