Graf commitů

9 Commity

Autor SHA1 Zpráva Datum
foreman f998e7094b P4 to Git Change 1057998 by gandryey@gera-dev-w7 on 2014/07/22 17:15:58
ECR #304775 - Device enqueuing
	- Use atomic fetch for enqueue flags
	- Switch to a multithreaded scheduler
	- Add a workaround for Linux host_multi_queue failures. Linux has only 2 queues, but the test allocates multiple host queues and the same HW ring can be used

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpublit.cpp#106 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpudevice.cpp#449 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpudevice.hpp#127 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuschedcl.cpp#22 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuvirtual.cpp#325 edit


[ROCm/clr commit: d2b905f18e]
2014-07-22 17:30:56 -04:00
foreman 1d845a69b2 P4 to Git Change 1057571 by gandryey@gera-dev-w7 on 2014/07/21 18:35:02
ECR #304775 - Device enqueuing
	- Add L2 cache flush after the scheduler execution. Although CP has to work with L2 cache, it seems some functionality relies on direct memory access and without explicit L2 flush CP can pick old values in the template.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuvirtual.cpp#324 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gslbe/src/rt/GSLContext.cpp#61 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gslbe/src/rt/GSLContext.h#35 edit


[ROCm/clr commit: 4599bd0d4a]
2014-07-21 18:43:54 -04:00
foreman 609ac948cd P4 to Git Change 1056280 by gandryey@gera-dev-w7 on 2014/07/17 11:55:30
ECR #304775 - Device enqueuing
	- Add extra sync point for the scheduler exit.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpusched.hpp#9 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuschedcl.cpp#21 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuvirtual.cpp#323 edit


[ROCm/clr commit: 6627cf5cf7]
2014-07-17 12:20:49 -04:00
foreman af0bc54257 P4 to Git Change 1053379 by xcui@merged_opencl_jxcwin on 2014/07/08 19:38:01
EPR #304775 - fixed the bug 9838. The svm pointer in the arugment list needs to be tracked to make sure all operation of resource has been finished before we dispatch kernel.

	code review:
	http://ocltc.amd.com/reviews/r/5200/
	precheckin:
	http://ocltc.amd.com:8111/viewModification.html?modId=35125&personal=true&buildTypeId=&tab=vcsModificationTests

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpukernel.cpp#257 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuvirtual.cpp#322 edit


[ROCm/clr commit: bf32cddc03]
2014-07-08 19:45:31 -04:00
foreman 1553537a82 P4 to Git Change 1053268 by gandryey@gera-dev-w7 on 2014/07/08 15:08:51
ECR #304775 - Device enqueuing
	- Add printing of the waiting events
	- Add early exit in the scheduler if nothing to launch

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuschedcl.cpp#19 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuvirtual.cpp#321 edit


[ROCm/clr commit: 977f314094]
2014-07-08 15:30:04 -04:00
foreman af11d3e66c P4 to Git Change 1052832 by gandryey@gera-dev-w7 on 2014/07/07 18:44:29
ECR #304775 - Device enqueuing
	- Update the scheduler to handle event mask

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuschedcl.cpp#18 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuvirtual.cpp#320 edit


[ROCm/clr commit: cd3fefb00d]
2014-07-07 18:58:52 -04:00
foreman e1ddad4e50 P4 to Git Change 1052441 by gandryey@gera-dev-w7 on 2014/07/04 18:58:11
ECR #304775 - Device enqueuing
	- Match the printed value width with the argument size

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuvirtual.cpp#319 edit


[ROCm/clr commit: 0d42b943fe]
2014-07-04 19:03:58 -04:00
foreman ad3093e80c P4 to Git Change 1052436 by gandryey@gera-dev-w7 on 2014/07/04 17:46:53
ECR #304775 - Device enqueuing
	- Added debug print for the generated child kernels. GPU_PRINT_CHILD_KERNEL=N, where N is the number of child kernels for dump.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuvirtual.cpp#318 edit
... //depot/stg/opencl/drivers/opencl/runtime/utils/flags.hpp#205 edit


[ROCm/clr commit: e4b73f0896]
2014-07-04 17:54:13 -04:00
foreman f80f2f233c initial commit
[ROCm/clr commit: 3694ab2ce8]
2014-07-04 16:17:05 -04:00