Commit Graph

146 Commits

Author SHA1 Message Date
Bartosz Dunajski 0fd8b850a6 Support for enabling perf counters with clCreateCommandQueueWithProperties
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2022-03-22 10:33:40 +01:00
Dominik Dabek d5fedf90c5 Fix for svm pointer arg caching
Previous version could cause segfaults.

Related-To: NEO-6737

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2022-03-08 12:13:15 +01:00
Dominik Dabek 6556d9a510 Improve caching in clSetKernelArgSVMPointer 2/n
Update allocIdMemoryManagerCounter on cache hit

Related-To: NEO-6737

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2022-03-02 18:10:52 +01:00
Dominik Dabek 7ab86d44d6 Improve caching in clSetKernelArgSVMPointer
Check allocId earlier and also reuse if allocationsCounter did not
change from last call.

Related-To: NEO-6737

Co-authored-by: Michal Mrozek <michal.mrozek@intel.com>

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2022-03-02 15:56:21 +01:00
Dominik Dabek 9bc364e7a7 Fix for clSetKernelArgSVMPointer optimization
Related-To: NEO-6737

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2022-03-01 12:56:04 +01:00
Dominik Dabek b9d8d8c0fd Optimize setKernelArgSVMPointer
If same pointer is already set, we don't need to set it again.

Related-To: NEO-6737

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2022-03-01 09:12:13 +01:00
Krzysztof Gibala e518a8f3f9 Add debug flag ForceExtendedUSMBufferSize
Forces extended buffer size by adding pageSize specify by number when
debug flag is >=1 in:
- clHostMemAllocINTEL
- clDeviceMemAllocINTEL
- clSharedMemAllocINTEL

Usage:
ForceExtendedUSMBufferSize=2
size += (2 * pageSize)

Signed-off-by: Krzysztof Gibala <krzysztof.gibala@intel.com>
2022-02-07 11:44:31 +01:00
Katarzyna Cencelewska 3d0c065183 Remove device enqueue part 16
-delete old unused flags

Related-To: NEO-6559
Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2022-02-03 19:38:41 +01:00
Katarzyna Cencelewska 58055aecdf Remove device enqueue part 12
remove:
- debug flag ForceDeviceEnqueueSupport
- functions isDeviceEnqueueSupported, supportsDeviceEnqueue

Related-To: NEO-6559
Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2022-01-28 13:38:34 +01:00
Katarzyna Cencelewska 3d9e1ea3a5 Remove device enqueue part 11
- remove templates from queue functions

Related-To: NEO-6559
Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2022-01-25 12:40:56 +01:00
Krzysztof Gibala b75f5d4c8b Add debug flag ForceExtendedBufferSize
Forces extended buffer size by adding pageSize specify by number when
debug flag is >=1 in:
- clCreateBuffer
- clCreateBufferWithProperties
- clCreateBufferWithPropertiesINTEL

Usage:
ForceExtendedBufferSize=2
size += (2 * pageSize)

Signed-off-by: Krzysztof Gibala <krzysztof.gibala@intel.com>
2022-01-19 18:52:10 +01:00
Katarzyna Cencelewska a06fbd2077 Remove device enqueue part 10
- remove DeviceQueue

Related-To: NEO-6559
Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2022-01-19 17:41:06 +01:00
Artur Harasimiuk 42115e4a1d fix identifier naming
Signed-off-by: Artur Harasimiuk <artur.harasimiuk@intel.com>
2021-10-14 21:10:58 +02:00
Mateusz Jablonski b1df167632 Move enable_product.inl to shared
extract api agnostic validators to shared
remove not needed opencl includes from neo shared


Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2021-10-12 22:02:16 +02:00
Fabian Zwolinski 332720920a Expose clGetKernelSuggestedLocalWorkSizeKHR func
Related-to: NEO-2922
Signed-off-by: Fabian Zwolinski <fabian.zwolinski@intel.com>
2021-10-12 20:06:57 +02:00
Mateusz Jablonski ec1e39bab8 Move OCL specific part of FileLogger to ClFileLogger
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2021-10-11 23:55:08 +02:00
Mateusz Jablonski 48f01f28f5 Move core part of MemoryPropertiesHelpers to shared
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2021-10-07 23:14:38 +02:00
Artur Harasimiuk a63b9b1273 move common files to shared/
Signed-off-by: Artur Harasimiuk <artur.harasimiuk@intel.com>
2021-09-27 11:10:20 +02:00
Kamil Kopryk 9ccf43e441 Correct branch_dir_suffix in cmake
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
Related-To: NEO-6245
2021-09-14 16:00:20 +02:00
Filip Hazubski 3d6d4acda2 Update isCooperativeDispatchSupported
Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2021-09-02 17:36:58 +02:00
Bartosz Dunajski 0345d9f707 Improve queries with number of SubDevices
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2021-09-01 14:01:57 +02:00
Bartosz Dunajski cb91474b1a Improve getDevice methods
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2021-08-26 14:45:00 +02:00
Filip Hazubski 63f8c9d98b Add new EngineUsage
Simplify verifying EngineUsage in tryGetEngine function.
Remove unused getGpgpuEnginesCount function.

Related-To: NEO-4940

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2021-08-23 17:41:13 +02:00
Szymon Morek 3c15023871 Add AllowUnrestrictedSize debug flag
This debug flag allows to allocate memory with size greater than
CL_DEVICE_MAX_MEM_ALLOC_SIZE.

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2021-08-05 15:35:26 +02:00
Michal Mrozek d18172c00e Add debug flag to control tryFlushEvent
Signed-off-by: Michal Mrozek <michal.mrozek@intel.com>
2021-07-15 11:40:13 +02:00
Szymon Morek 399dc60bf8 Remove processing extra properties
Related-To: NEO-6024

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2021-07-08 16:18:06 +02:00
Kamil Kopryk dbe3ef21c1 Prefer unique_ptr instead of raw pointer nv12 ults
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
Related-To: NEO-4692
2021-06-28 13:54:37 +02:00
Szymon Morek 8b48b33a0c Remove 4GB limit if shared allocations supported
Resolves: NEO-5984

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2021-06-24 13:28:59 +02:00
Maciej Dziuban 13b9f54c7d Check queue capabilities in SVM calls
Signed-off-by: Maciej Dziuban <maciej.dziuban@intel.com>
2021-06-23 16:14:31 +02:00
Mateusz Jablonski 0602a9c285 Use only one device when creating context from type
Related-To: NEO-3691
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2021-05-26 11:51:09 +02:00
Konstanty Misiak 074fc1d60f Fix clGetKernelSuggestedLocalWorkSizeINTEL
Related-To: NEO-5456

Signed-off-by: Konstanty Misiak <konstanty.misiak@intel.com>
2021-05-18 13:35:26 +02:00
Anton V. Gorshkov 187bfefe28 Invalid pointer passed to tracer
Fixed an issue with invalid return value passed to tracer for
clGetExtensionFunctionAddress

Signed-off-by: Anton V. Gorshkov <anton.v.gorshkov@intel.com>
2021-05-18 12:31:37 +02:00
lgotszal 3bd4bca911 Copyright header update
Dates corrected in copyright headers to reflect original publication date
(2018 for OpenCL, 2020 for Level Zero).

Signed-off-by: lgotszal <lukasz.gotszald@intel.com>
2021-05-17 20:38:19 +02:00
Sebastian Luzynski 95ba83c7e1 Fix ocl program creation on multi device contexts
Signed-off-by: Sebastian Luzynski <sebastian.jozef.luzynski@intel.com>

Related-To: NEO-5734
2021-05-10 12:30:09 +02:00
Krystian Chmielewski ef71915c71 Remove KernelArgInfo and use KernelDescriptor's args instead
Related-To: NEO-4729
Signed-off-by: Krystian Chmielewski <krystian.chmielewski@intel.com>
2021-04-23 15:12:09 +02:00
Mateusz Jablonski 1e97e4117e Respect platform in context properties when creating context
Resolves: NEO-5223
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2021-03-30 20:59:19 +02:00
Mateusz Jablonski d6bbe48175 Handle SVM allocations from multi root device contexts
Related-To: NEO-5001, NEO-3691
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2021-03-24 11:14:39 +01:00
Mateusz Jablonski 35ff284944 Cleanup Kernel class
move deviceVector to MultiDeviceKernel class
remove Device arg from Kernel's methods

Related-To: NEO-5001
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2021-03-24 09:17:41 +01:00
Mateusz Jablonski 7098e9c5f2 Store single KernelInfo in Kernel
remove root device index from Kernel's methods

Related-To: NEO-5001
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2021-03-22 21:30:03 +01:00
Filip Hazubski ecceddcab6 Correct clSetKernelExecInfo not to always require SVM support
Resolves: NEO-5269

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2021-03-22 21:08:03 +01:00
Filip Hazubski 8d55bfe21d Implement zeCommandListAppendLaunchCooperativeKernel
Resolves: NEO-4725


Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2021-03-22 19:26:41 +01:00
Mateusz Jablonski 487f10f999 Use MultiDeviceKernel instead of Kernel in clCloneKernel
Related-To: NEO-5001
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2021-03-18 20:19:54 +01:00
Mateusz Jablonski 04eca48ee0 Introduce MultiDeviceKernel class
Related-To: NEO-5001
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2021-03-15 11:46:52 +01:00
Katarzyna Cencelewska 7fdbf4f6ef Correct set pitch in clEnqueueCopyBufferRect
validate dst and src buffers

Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2021-03-10 12:41:51 +01:00
Krystian Chmielewski 41f3bd00ff Remove PatchTokens from KernelInfo
Use KernelDescriptor instead of patchTokens stored in KernelInfo's
patchInfo.
Removed: SPatchMediaInterfaceDescriptorLoad, SPatchAllocateLocalSurface,
SPatchMediaVFEState(slot 0), SPatchMediaVFEState(slot 1),
SPatchInterfaceDescriptorData, SPatchSamplerStateArray,
SPatchBindingTableState, SPatchDataParameterBuffer,
SPatchDataParameterStream, SPatchThreadPayload,
SPatchKernelAttributesInfo, SPatchAllocateStatelessPrivateSurface,
SPatchAllocateSyncBuffer,
SPatchAllocateStatelessConstantMemorySurfaceWithInitialization,
SPatchAllocateStatelessGlobalMemorySurfaceWithInitialization,
SPatchAllocateSystemThreadSurface.

Related-To: NEO-4729
Signed-off-by: Krystian Chmielewski <krystian.chmielewski@intel.com>
2021-03-09 17:09:13 +01:00
Mateusz Jablonski 056300f01f Pass internal kernel object to enqueueKernel method
Related-To: NEO-5001
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2021-03-08 19:14:08 +01:00
Compute-Runtime-Validation 68e01dd736 Revert "Correct set pitch in clEnqueueCopyBufferRect"
This reverts commit a0f3e868a1.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2021-02-26 13:48:41 +01:00
Jaime Arteaga 722a80079b Add experimental_properties for indirect access (2)
Add support to OpenCL to also make use of indirect access properties
provided by IGC and to make use of DisableIndirectAccess debug key.

Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
2021-02-25 11:41:24 +01:00
Katarzyna Cencelewska a0f3e868a1 Correct set pitch in clEnqueueCopyBufferRect
Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2021-02-25 10:39:57 +01:00
Maciej Dziuban 08506dc81e OpenCL Queue Families extension 14/n
Check capabilities in clEnqueueWaitForEvents

Related-To: NEO-5120
Signed-off-by: Maciej Dziuban <maciej.dziuban@intel.com>
2021-02-02 11:54:43 +01:00