[SYCL][CUDA] Fixes for multiple backends in the same program #1252
Conversation
This file contains hidden or 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
|
Yes, it works: Thank you for the fixes ! |
fwyzard
added a commit
to cms-patatrack/patatrack-website
that referenced
this pull request
Mar 5, 2020
romanovvlad
requested changes
Mar 5, 2020
steffenlarsen
force-pushed
the
ruyman/event-callback
branch
3 times, most recently
from
3595834 to
8a642f2
Compare
|
An additional commit has been added: This ensures that CUDA operations are allowed in device-event callbacks, which could be needed if e.g. an event ends its life in the callback, by introducing a worker-thread that sleeps until callbacks are added to its work-queue. Edit: This commit has been dropped. |
steffenlarsen
force-pushed
the
ruyman/event-callback
branch
from
8a642f2 to
f6904a8
Compare
steffenlarsen
force-pushed
the
ruyman/event-callback
branch
2 times, most recently
from
968e7e9 to
5baf374
Compare
Could you please share more details?
You mean you can call cuEventDestroy in the callback? Are there other use cases than "an event ends its life in the callback"? |
Since CUDA events are explicitly destroyed, we must retain the event when we enqueue callbacks, so yes we need this for Likewise I believe for one of the current uses of the callbacks the registered callback also destroys the |
Ruyk
changed the title
fwyzard
added a commit
to fwyzard/llvm
that referenced
this pull request
Mar 14, 2020
|
@Ruyk, @steffenlarsen, could you resolve merge conflicts, please? |
steffenlarsen
force-pushed
the
ruyman/event-callback
branch
from
5baf374 to
458754f
Compare
steffenlarsen
force-pushed
the
ruyman/event-callback
branch
from
458754f to
146b163
Compare
romanovvlad
previously approved these changes
Mar 19, 2020
|
@steffenlarsen, @Ruyk, could you take a look at failed checks, please? |
|
The unittests are currently selecting the first plugin, which seems to be the OpenCL plugin. Is this failing on a machine with no available OpenCL devices? This should definitely be fixed, but we need a way of selecting the plugins correctly. |
|
Isn't that what the |
The unittests tests the PI API and therefore doesn't use the SYCL device selector. The initial way adjusting the tests to the plugin system was to select the first plugin in the collection returned by |
steffenlarsen
force-pushed
the
ruyman/event-callback
branch
from
146b163 to
63c0b40
Compare
|
The EventTest unittest has been disabled. I notice that most unittests are currently disabled, so I opened a new issue (#1362) to investigate this. |
bader
previously approved these changes
Mar 20, 2020
fwyzard
added a commit
to cms-patatrack/llvm
that referenced
this pull request
Mar 21, 2020
|
It seems to be failing with MSVC. From what I gather there is a problem with printing the parameters of |
|
Could you try to ifdef the check at line 150 i.e. sycl\unittests\pi\EventTest.cpp(150)? |
Or it has problems with printing lambda object. |
steffenlarsen
force-pushed
the
ruyman/event-callback
branch
from
63c0b40 to
585902a
Compare
That's an interesting point. I've tried to move the callback to a separate function. I can't test it, so let's see what the bot says. |
Signed-off-by: Stuart Adams <stuart.adams@codeplay.com> Signed-off-by: Steffen Larsen <steffen.larsen@codeplay.com> Signed-off-by: Ruyman Reyes <ruyman.reyes@codeplay.com>
The SYCL RT code for GlueEvent calls now the right plugin to create the event that triggers the dependency chain. Renamed variables to clarify the source code and avoid confusions between Context and Plugin Signed-off-by: Ruyman Reyes <ruyman@codeplay.com>
steffenlarsen
force-pushed
the
ruyman/event-callback
branch
from
585902a to
cd9b90e
Compare
romanovvlad
approved these changes
Mar 24, 2020
alexbatashev
added a commit
to alexbatashev/llvm
that referenced
this pull request
Mar 27, 2020
…hinx * upstream/sycl: (357 commits) [Support] Implement a simple tabular data management library (intel#1358) [Support] Implement a property set I/O library (intel#1357) [SYCL] Fix buffer constructor using iterators (intel#1386) [SYCL][FPGA] Enable a set of loop attributes (intel#1312) [Driver][SYCL][FPGA] Proper dependency output location when given /Fo<dir> (intel#1346) [SPIR-V] Enabling SPIR-V builtin lookup in device SYCL mode (intel#1384) [SYCL][NFC] Unify setting kernel arguments (intel#1379) [SYCL][Doc] First revision of standard layout relaxation extension (intel#1344) [SYCL] Fixed sub-buffer alloca search (intel#1385) [SYCL][FPGA] Emit multiple IR variants for the IVDep attribute (intel#1383) [SYCL] Add experimental flag to enable front-end optimizations (intel#1376) [SYCL] Remove unexpected double in complex SPIR-V for float support (intel#1381) [SYCL] Default work-group sizes based on max (intel#952) [SYCL][CUDA] Fix usage of multiple backends in the same program (intel#1252) [SPIR-V] Add SPIR-V builtin definitions to the builtin lookup. [SPIR-V] Add macro definition when -fdeclare-spirv-builtins is activated [SYCL] Fix sycl_generic printing [SYCL] Support intel::reqd_work_group_size (intel#1328) [SYCL][NFC] Make the RT::PiPlugin object private (intel#1375) [SPIRV] Add convergent attribute to SPIR-V built-ins (intel#1373) ...
|
Dear all,
when I executed the program i have following output: The clinfo detects following platforms on my machine. Could you plase advice on that? How can I enable both backends? Thanks. Here is the compilation log: |
Could you please open an issue? https://github.com/intel/llvm/issues |
|
Thanks, I created: #2372 |
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.
This pull request contains two commits:
This two patches should fix #1231, but should not be merged until we run some additional testing on our side.