[L0] Support updating kernel commands in command buffers #1353
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
againull
force-pushed
the
againull/l0_adapter_update_cmd_buffer
branch
2 times, most recently
from
7c0b644
to
ddc3675
Compare
againull
force-pushed
the
againull/l0_adapter_update_cmd_buffer
branch
from
ddc3675
to
ac72e87
Compare
EwanC
reviewed
Mar 4, 2024
EwanC
approved these changes
Mar 5, 2024
There was a problem hiding this comment.
LGTM - Ideally we could test this with the DPC++ PR in intel/llvm#12840 on a setup with this driver feature and see if the new tests pass (apart from range update) when enabled for L0. But with the CTS tests I wouldn't say this is required and we could fix up any issues in a follow-up PR.
kbenzie
changed the title
againull
force-pushed
the
againull/l0_adapter_update_cmd_buffer
branch
from
3059fa1
to
5f8a62c
Compare
nrspruit
added
level-zero
|
Please pull in the main branch to have up to date testing, also update the tag in the intel/llvm PR. |
kbenzie
force-pushed
the
againull/l0_adapter_update_cmd_buffer
branch
from
5f8a62c
to
0b355e1
Compare
Codecov ReportAttention: Patch coverage is
❗ Your organization needs to install the Codecov GitHub app to enable full functionality. Additional details and impacted files@@ Coverage Diff @@
## main #1353 +/- ##
==========================================
- Coverage 14.82% 12.48% -2.34%
==========================================
Files 250 239 -11
Lines 36220 36022 -198
Branches 4094 4092 -2
==========================================
- Hits 5369 4498 -871
- Misses 30800 31520 +720
+ Partials 51 4 -47 ☔ View full report in Codecov by Sentry. |
Closed
* Execution range update is not supported by the L0 driver right now. Currently it supports only kernel arguments update. * There is a synchronization issue with immediate submission when used for command buffers. It is reproducible even without changes of this PR, so should be fixed separately. For now use batched submission for command buffer update tests.
* Remove unnecesary env variable setting * Check support of the feature on device * Check sub-feature support before usage at urCommandBufferUpdateKernelLaunchExp * Remove redundant code
againull
force-pushed
the
againull/l0_adapter_update_cmd_buffer
branch
from
0b355e1
to
c9be1e2
Compare
Done. |
EwanC
added a commit
to Bensuo/unified-runtime
that referenced
this pull request
Mar 14, 2024
Implement the API for updating the kernel commands in a command-buffer defined by oneapi-src#1089 for the OpenCL adapter. However, the following changes to the UR kernel update API have been made based on implementation experience: 1. Forbid updating the work-dim of the kernel, see KhronosGroup/OpenCL-Docs#1057 2. Remove struct fields to update exec info, after [DPC++ implementation prototype](intel/llvm#12840) shows this isn't needed. 3. Forbid changing the local work size from user to impl defined and vice-versa. See discussion in [L0 implementation PR](oneapi-src#1353 (comment)). This adapter implementation depends on support for the [cl_khr_command_buffer_mutable_dispatch](https://registry.khronos.org/OpenCL/specs/3.0-unified/html/OpenCL_Ext.html#cl_khr_command_buffer_mutable_dispatch) extension. Tested on Intel GPU/CPUs OpenCL implementations with the [command-buffer emulation layer](https://github.com/bashbaug/SimpleOpenCLSamples/tree/main/layers/10_cmdbufemu). ```bash $ OPENCL_LAYERS=<path/to/SimpleOpenCLSamples/build/layers/10_cmdbufemu/libCmdBufEmu.so> ./bin/test-exp_command_buffer --platform="Intel(R) OpenCL Graphics" ``` DPC++ PR intel/llvm#12724
EwanC
added a commit
to Bensuo/unified-runtime
that referenced
this pull request
Mar 20, 2024
Fixes Coverity defect report from L0 command-buffer update code merged in oneapi-src#1353 ``` This greater-than-or-equal-to-zero comparison of an unsigned value is always true. "CommandDesc->newWorkDim >= 0U". ```
EwanC
added a commit
to Bensuo/unified-runtime
that referenced
this pull request
Mar 21, 2024
Fixes Coverity defect report from L0 command-buffer update code merged in oneapi-src#1353 ``` This greater-than-or-equal-to-zero comparison of an unsigned value is always true. "CommandDesc->newWorkDim >= 0U". ```
EwanC
added a commit
to Bensuo/unified-runtime
that referenced
this pull request
Mar 22, 2024
Fixes Coverity defect report from L0 command-buffer update code merged in oneapi-src#1353 ``` This greater-than-or-equal-to-zero comparison of an unsigned value is always true. "CommandDesc->newWorkDim >= 0U". ```
EwanC
added a commit
to Bensuo/unified-runtime
that referenced
this pull request
Mar 22, 2024
Fixes Coverity defect report from L0 command-buffer update code merged in oneapi-src#1353 ``` This greater-than-or-equal-to-zero comparison of an unsigned value is always true. "CommandDesc->newWorkDim >= 0U". ```
EwanC
added a commit
to Bensuo/unified-runtime
that referenced
this pull request
Mar 25, 2024
Implement the API for updating the kernel commands in a command-buffer defined by oneapi-src#1089 for the OpenCL adapter. However, the following changes to the UR kernel update API have been made based on implementation experience: 1. Forbid updating the work-dim of the kernel, see KhronosGroup/OpenCL-Docs#1057 2. Remove struct fields to update exec info, after [DPC++ implementation prototype](intel/llvm#12840) shows this isn't needed. 3. Forbid changing the local work size from user to impl defined and vice-versa. See discussion in [L0 implementation PR](oneapi-src#1353 (comment)). This adapter implementation depends on support for the [cl_khr_command_buffer_mutable_dispatch](https://registry.khronos.org/OpenCL/specs/3.0-unified/html/OpenCL_Ext.html#cl_khr_command_buffer_mutable_dispatch) extension. Tested on Intel GPU/CPUs OpenCL implementations with the [command-buffer emulation layer](https://github.com/bashbaug/SimpleOpenCLSamples/tree/main/layers/10_cmdbufemu). ```bash $ OPENCL_LAYERS=<path/to/SimpleOpenCLSamples/build/layers/10_cmdbufemu/libCmdBufEmu.so> ./bin/test-exp_command_buffer --platform="Intel(R) OpenCL Graphics" ``` DPC++ PR intel/llvm#12724
EwanC
added a commit
to Bensuo/unified-runtime
that referenced
this pull request
Apr 4, 2024
Implement the API for updating the kernel commands in a command-buffer defined by oneapi-src#1089 for the OpenCL adapter. However, the following changes to the UR kernel update API have been made based on implementation experience: 1. Forbid updating the work-dim of the kernel, see KhronosGroup/OpenCL-Docs#1057 2. Remove struct fields to update exec info, after [DPC++ implementation prototype](intel/llvm#12840) shows this isn't needed. 3. Forbid changing the local work size from user to impl defined and vice-versa. See discussion in [L0 implementation PR](oneapi-src#1353 (comment)). This adapter implementation depends on support for the [cl_khr_command_buffer_mutable_dispatch](https://registry.khronos.org/OpenCL/specs/3.0-unified/html/OpenCL_Ext.html#cl_khr_command_buffer_mutable_dispatch) extension. Tested on Intel GPU/CPUs OpenCL implementations with the [command-buffer emulation layer](https://github.com/bashbaug/SimpleOpenCLSamples/tree/main/layers/10_cmdbufemu). ```bash $ OPENCL_LAYERS=<path/to/SimpleOpenCLSamples/build/layers/10_cmdbufemu/libCmdBufEmu.so> ./bin/test-exp_command_buffer --platform="Intel(R) OpenCL Graphics" ``` DPC++ PR intel/llvm#12724
EwanC
added a commit
to Bensuo/unified-runtime
that referenced
this pull request
Apr 5, 2024
Implement the API for updating the kernel commands in a command-buffer defined by oneapi-src#1089 for the OpenCL adapter. However, the following changes to the UR kernel update API have been made based on implementation experience: 1. Forbid updating the work-dim of the kernel, see KhronosGroup/OpenCL-Docs#1057 2. Remove struct fields to update exec info, after [DPC++ implementation prototype](intel/llvm#12840) shows this isn't needed. 3. Forbid changing the local work size from user to impl defined and vice-versa. See discussion in [L0 implementation PR](oneapi-src#1353 (comment)). This adapter implementation depends on support for the [cl_khr_command_buffer_mutable_dispatch](https://registry.khronos.org/OpenCL/specs/3.0-unified/html/OpenCL_Ext.html#cl_khr_command_buffer_mutable_dispatch) extension. Tested on Intel GPU/CPUs OpenCL implementations with the [command-buffer emulation layer](https://github.com/bashbaug/SimpleOpenCLSamples/tree/main/layers/10_cmdbufemu). ```bash $ OPENCL_LAYERS=<path/to/SimpleOpenCLSamples/build/layers/10_cmdbufemu/libCmdBufEmu.so> ./bin/test-exp_command_buffer --platform="Intel(R) OpenCL Graphics" ``` DPC++ PR intel/llvm#12724
EwanC
added a commit
to Bensuo/unified-runtime
that referenced
this pull request
Apr 22, 2024
Implement the API for updating the kernel commands in a command-buffer defined by oneapi-src#1089 for the OpenCL adapter. However, the following changes to the UR kernel update API have been made based on implementation experience: 1. Forbid updating the work-dim of the kernel, see KhronosGroup/OpenCL-Docs#1057 2. Remove struct fields to update exec info, after [DPC++ implementation prototype](intel/llvm#12840) shows this isn't needed. 3. Forbid changing the local work size from user to impl defined and vice-versa. See discussion in [L0 implementation PR](oneapi-src#1353 (comment)). This adapter implementation depends on support for the [cl_khr_command_buffer_mutable_dispatch](https://registry.khronos.org/OpenCL/specs/3.0-unified/html/OpenCL_Ext.html#cl_khr_command_buffer_mutable_dispatch) extension. Tested on Intel GPU/CPUs OpenCL implementations with the [command-buffer emulation layer](https://github.com/bashbaug/SimpleOpenCLSamples/tree/main/layers/10_cmdbufemu). ```bash $ OPENCL_LAYERS=<path/to/SimpleOpenCLSamples/build/layers/10_cmdbufemu/libCmdBufEmu.so> ./bin/test-exp_command_buffer --platform="Intel(R) OpenCL Graphics" ``` DPC++ PR intel/llvm#12724
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Initial support for updating command buffers.
Also applied temporary fixes for conformance tests:
Execution range update is not supported by the L0 driver right now, so skipping those tests for L0 backend.
There is a synchronization issue with immediate submission when used for command buffers. It is reproducible even without changes of this PR, so should be fixed separately. For now use batched submission for command buffer update tests.
intel/llvm PR: intel/llvm#12897