Patch submitted by Besar Wicaksono
1. Bug on detecting local memory size interpreted as 32 bit value
instead of 64. The bug causes thunk to go into an infinite loop trying
to reserve virtual address range for dgpu system memory.
2. SIMD count in the node property is 0. Runtime use this attribute to
find a gpu device.
Regarding other attributes of intel+tonga topology, Harish started a
discussion on August iirc, could you please share an update ?
This would help me progress with more tests such as scratch memory,
which require the scratch aperture information in order to construct a
buffer srd in gpuvm space.
3. Bug on releasing memory via fmm_release, where no actual release is
being done. The vm_object can't be found because the memory size does
not match due to the allocation padded the size with 32KB.
4. Pointer arithmetic on vm_area allocation/release. The value of
vm_area_t::end seems to be interpreted inconsistently whether it is
(start + size -1) or (start + size).
One example of potential issue I see is the logic could report
larger size of the hole in the vm area list.
5. Resource cleanup on multiple library load/unload within a single
process.
- Any memory allocation on subsequent library load will result
an error "va above limit". To my understanding this is due to
the reserved memory for the system memory not being released on unload.
- The static variable events_page needs to be invalidated
appropriately on library unload so the next load could
reinitialize it.
6. Could you please update if AQL queue is ready to test with the stg
kfd/kmt ?
7. The system memory allocation with size larger than 32KB seems to be
padded by an extra 32KB. I was wondering if we could remove this
overhead.
Change-Id: I039988d36637525089c7569dc3b77e58750e2121
Patch submitted by Besar Wicaksono
1. Bug on detecting local memory size interpreted as 32 bit value
instead of 64. The bug causes thunk to go into an infinite loop trying
to reserve virtual address range for dgpu system memory.
2. SIMD count in the node property is 0. Runtime use this attribute to
find a gpu device.
Regarding other attributes of intel+tonga topology, Harish started a
discussion on August iirc, could you please share an update ?
This would help me progress with more tests such as scratch memory,
which require the scratch aperture information in order to construct a
buffer srd in gpuvm space.
3. Bug on releasing memory via fmm_release, where no actual release is
being done. The vm_object can't be found because the memory size does
not match due to the allocation padded the size with 32KB.
4. Pointer arithmetic on vm_area allocation/release. The value of
vm_area_t::end seems to be interpreted inconsistently whether it is
(start + size -1) or (start + size).
One example of potential issue I see is the logic could report
larger size of the hole in the vm area list.
5. Resource cleanup on multiple library load/unload within a single
process.
- Any memory allocation on subsequent library load will result
an error "va above limit". To my understanding this is due to
the reserved memory for the system memory not being released on unload.
- The static variable events_page needs to be invalidated
appropriately on library unload so the next load could
reinitialize it.
6. Could you please update if AQL queue is ready to test with the stg
kfd/kmt ?
7. The system memory allocation with size larger than 32KB seems to be
padded by an extra 32KB. I was wondering if we could remove this
overhead.
Change-Id: I039988d36637525089c7569dc3b77e58750e2121
[ROCm/ROCR-Runtime commit: ee08f537a7]
EPR #010002 - Change OpenCL version number from 1893 to 1894.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1640 edit
[ROCm/clr commit: 15841f57cb]
EPR #010002 - Change OpenCL version number from 1893 to 1894.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1640 edit
EPR #010002 - Change OpenCL version number from 1892 to 1893.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1639 edit
[ROCm/clr commit: 6e1eb56309]
EPR #010002 - Change OpenCL version number from 1892 to 1893.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1639 edit
EPR #010002 - Change OpenCL version number from 1891 to 1892.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1638 edit
[ROCm/clr commit: 31c862938a]
EPR #010002 - Change OpenCL version number from 1891 to 1892.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1638 edit
EPR #425801 - Only apply optimize init sequeuce for Windows 10 - to avoid issue observed in EPR#425801 for Win7.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gslbe/src/rt/GSLDevice.cpp#140 edit
[ROCm/clr commit: ef264b7508]
EPR #425801 - Only apply optimize init sequeuce for Windows 10 - to avoid issue observed in EPR#425801 for Win7.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gslbe/src/rt/GSLDevice.cpp#140 edit
ECR #333753 - ORCA RT: fix for continuing work after Program build failure.
[Synopsis]
If program build (compile or link) is failed RT continues its work and fails further (for example, in symbol.setDeviceKernel). It was hard to find out the real failure (HSAILProgram::linkImpl) due to buildLog, which is printed once in the end, and due to abscence of LogError function call.
[Misc]
+ Error handling improvement in HSAILProgram::linkImpl and some others. ToDo: Do the same for the rest.
+ LogError() is added to Program::build/compile/link
[Testing] pre check-in
Reviewer: German Andryeyev
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/device.cpp#184 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuprogram.cpp#206 edit
... //depot/stg/opencl/drivers/opencl/runtime/platform/program.cpp#65 edit
[ROCm/clr commit: 94eccbc8e5]
ECR #333753 - ORCA RT: fix for continuing work after Program build failure.
[Synopsis]
If program build (compile or link) is failed RT continues its work and fails further (for example, in symbol.setDeviceKernel). It was hard to find out the real failure (HSAILProgram::linkImpl) due to buildLog, which is printed once in the end, and due to abscence of LogError function call.
[Misc]
+ Error handling improvement in HSAILProgram::linkImpl and some others. ToDo: Do the same for the rest.
+ LogError() is added to Program::build/compile/link
[Testing] pre check-in
Reviewer: German Andryeyev
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/device.cpp#184 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuprogram.cpp#206 edit
... //depot/stg/opencl/drivers/opencl/runtime/platform/program.cpp#65 edit
EPR #010002 - Change OpenCL version number from 1890 to 1891.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1637 edit
[ROCm/clr commit: fb75127a61]
EPR #010002 - Change OpenCL version number from 1890 to 1891.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1637 edit
EPR #010002 - Change OpenCL version number from 1889 to 1890.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1636 edit
[ROCm/clr commit: 5e0e35b64e]
EPR #010002 - Change OpenCL version number from 1889 to 1890.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1636 edit
EPR #010002 - Change OpenCL version number from 1888 to 1889.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1635 edit
[ROCm/clr commit: 414042ad4b]
EPR #010002 - Change OpenCL version number from 1888 to 1889.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1635 edit
EPR #010002 - Change OpenCL version number from 1887 to 1888.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1634 edit
[ROCm/clr commit: 6f7a4636a6]
EPR #010002 - Change OpenCL version number from 1887 to 1888.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1634 edit
EPR #010002 - Change OpenCL version number from 1886 to 1887.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1633 edit
[ROCm/clr commit: 0a81c9aed9]
EPR #010002 - Change OpenCL version number from 1886 to 1887.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1633 edit
ECR #304775 - Modify caching library code to be used by HSAIL runtime
1. Remove dependencies on shader compiler version, amd::Os and amd::option
2. Let version be one the of input
3. Change all prefix "Kernel" to prefix "String", e.g., class name is modified from KernelCache to StringCache, etc.
Affected files ...
... //depot/stg/opencl/drivers/opencl/compiler/lib/backends/common/cache.cpp#5 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/backends/common/cache.hpp#5 edit
... //depot/stg/opencl/drivers/opencl/compiler/tools/caching/cache.cpp#1 branch
... //depot/stg/opencl/drivers/opencl/compiler/tools/caching/cache.hpp#1 branch
[ROCm/clr commit: 4e9e419b6b]
ECR #304775 - Modify caching library code to be used by HSAIL runtime
1. Remove dependencies on shader compiler version, amd::Os and amd::option
2. Let version be one the of input
3. Change all prefix "Kernel" to prefix "String", e.g., class name is modified from KernelCache to StringCache, etc.
Affected files ...
... //depot/stg/opencl/drivers/opencl/compiler/lib/backends/common/cache.cpp#5 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/backends/common/cache.hpp#5 edit
... //depot/stg/opencl/drivers/opencl/compiler/tools/caching/cache.cpp#1 branch
... //depot/stg/opencl/drivers/opencl/compiler/tools/caching/cache.hpp#1 branch
EPR #426143 - [AVID] clEnqueueWriteImage with a row pitch different from 0 will fail if we use the pre-pinned path
- pass pitch and slice to the copy functions
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuvirtual.cpp#377 edit
[ROCm/clr commit: 8e3e9dbae5]
EPR #426143 - [AVID] clEnqueueWriteImage with a row pitch different from 0 will fail if we use the pre-pinned path
- pass pitch and slice to the copy functions
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuvirtual.cpp#377 edit
EPR #412821 - Default AMD_THREAD_TRACE_ENABLE to true for both main and stg.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/flags.hpp#239 edit
[ROCm/clr commit: decb7a1668]
EPR #412821 - Default AMD_THREAD_TRACE_ENABLE to true for both main and stg.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/flags.hpp#239 edit
EPR #010002 - Change OpenCL version number from 1885 to 1886.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1632 edit
[ROCm/clr commit: 5fe8a883a8]
EPR #010002 - Change OpenCL version number from 1885 to 1886.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1632 edit
ECR #304775 - Add "dbg" to temp filename to prevent clashing with compiler lib temp file name for dlls.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/cpu/cpuprogram.cpp#68 edit
EPR #010002 - Change OpenCL version number from 1884 to 1885.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1631 edit
[ROCm/clr commit: 2cb1ec95ea]
EPR #010002 - Change OpenCL version number from 1884 to 1885.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1631 edit
EPR #424190 - IOMMU2/SVM on CZ Win10, Optimization_1
- Map can now return the surface address directly as we do system memory allocations that GPU can rw via IOMMUv2.
- Unpack shaders binary to the shader surface directly, skip unmap which saves a lot of CPDMA calls.
ReviewBoardURL = http://ocltc.amd.com/reviews/r/8386/diff/
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gslbe/src/rt/GSLDevice.cpp#138 edit
[ROCm/clr commit: 4d4e4ce7a2]
EPR #424190 - IOMMU2/SVM on CZ Win10, Optimization_1
- Map can now return the surface address directly as we do system memory allocations that GPU can rw via IOMMUv2.
- Unpack shaders binary to the shader surface directly, skip unmap which saves a lot of CPDMA calls.
ReviewBoardURL = http://ocltc.amd.com/reviews/r/8386/diff/
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gslbe/src/rt/GSLDevice.cpp#138 edit
Makefile currently sends build output a default location.
Allow choice of build output location if so desired
using a variable.
Signed-off-by: David Ogbeide <davidboyowa.ogbeide@amd.com>
Makefile currently sends build output a default location.
Allow choice of build output location if so desired
using a variable.
Signed-off-by: David Ogbeide <davidboyowa.ogbeide@amd.com>
[ROCm/ROCR-Runtime commit: 8a01cd1212]
EPR #010002 - Change OpenCL version number from 1883 to 1884.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1630 edit
[ROCm/clr commit: 73b4bc7b83]
EPR #010002 - Change OpenCL version number from 1883 to 1884.
Affected files ...
... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#1630 edit
foreman
Harish Kasiviswanathan
David Ogbeide
Ding, Wei (xN/A) TX