intel/compute-runtime
1
0
Fork 0
mirror of https://github.com/intel/compute-runtime.git synced 2025-12-20 08:53:55 +08:00
Code Issues Packages Projects Releases Wiki Activity
12,252 Commits 42 Branches 655 Tags
01017a5df39bf5403bddefca6c5c3d4deb4ff7e6
Commit Graph

223 Commits

Author SHA1 Message Date
Compute-Runtime-Validation
ddb7eb07b0 Revert "Align external host ptr allocation gpu va to 2MB" 
This reverts commit d24f403cc4.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
 2023-01-12 09:33:16 +01:00
Lukasz Jobczyk
d24f403cc4 Align external host ptr allocation gpu va to 2MB 
Related-To: NEO-7116

Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
 2023-01-11 13:45:06 +01:00
Lukasz Jobczyk
39df92a16f Align only CPU VA for SVM_CPU 
Resolves: NEO-7608

Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
 2023-01-10 16:52:29 +01:00
Compute-Runtime-Validation
af81c79076 Revert "Deferred deletion of allocations in main thread" 
This reverts commit c268e30189.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
 2023-01-04 18:33:39 +01:00
HeFan2017
c268e30189 Deferred deletion of allocations in main thread 
Add a clearQueueTillFirstFailure interface to DeferredDeleter, which
iterates the queue from the front and delete the allocations in the
queue till a failure. It is called by defer deletion of allocations
occupied by mutliple contexts to unlock the execution in main thread

Related-To: NEO-7532

Signed-off-by: HeFan2017 <fan.f.he@intel.com>
 2023-01-03 13:51:15 +01:00
Warchulski, Jaroslaw
0eac749fcc Cleanup includes 19 
Cleaned up files:
opencl/source/api/cl_types.h
shared/source/compiler_interface/external_functions.h
shared/source/compiler_interface/linker.h
shared/source/device_binary_format/elf/elf.h
shared/source/helpers/preamble.h
shared/source/memory_manager/definitions/storage_info.h
shared/source/memory_manager/memory_manager.h
shared/source/memory_manager/os_agnostic_memory_manager.h
shared/source/program/program_info.h

Related-To: NEO-5548
Signed-off-by: Warchulski, Jaroslaw <jaroslaw.warchulski@intel.com>
 2023-01-03 11:02:05 +01:00
Warchulski, Jaroslaw
c3a86232d9 Cleanup includes 15 
Cleaned up files:
shared/source/command_stream/command_stream_receiver.h
shared/source/command_stream/submissions_aggregator.h
shared/source/helpers/flat_batch_buffer_helper.h

Related-To: NEO-5548
Signed-off-by: Warchulski, Jaroslaw <jaroslaw.warchulski@intel.com>
 2022-12-23 12:36:51 +01:00
Neil R Spruit
a829ce8941 Fix ReserveGPUAddress to reserve on specific RootDevice indices 
Related-To: LOCI-3790

Signed-off-by: Neil R Spruit <neil.r.spruit@intel.com>
 2022-12-20 23:08:02 +01:00
Warchulski, Jaroslaw
803bbb89d2 Cleanup includes 13 
Cleaned up files:
shared/source/helpers/blit_commands_helper.h
shared/source/helpers/heap_assigner.h
shared/source/memory_manager/alignment_selector.h
shared/source/memory_manager/gfx_partition.h
shared/source/memory_manager/memory_manager.h
shared/source/os_interface/os_memory.h
shared/source/utilities/heap_allocator.h

Related-To: NEO-5548
Signed-off-by: Warchulski, Jaroslaw <jaroslaw.warchulski@intel.com>
 2022-12-16 12:36:27 +01:00
Mateusz Jablonski
ecea487cf0 Reduce usage of global gfx core helper getter [2/n] 
Related-To: NEO-6853
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2022-12-12 13:43:21 +01:00
Kamil Kopryk
03b687881f Rename HwHelper -> GfxCoreHelper 
Related-To: NEO-6853
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
 2022-12-09 10:29:06 +01:00
Jaime Arteaga
675272a32c Return a unique pointer for multiple calls to openIpcHandles 
This to make zexMemOpenIpcHandles and zeMemOpenIpcHandle
behave similarly.

Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
 2022-12-08 21:56:21 +01:00
Slawomir Milczarek
5c1b50bccf Add memory prefetch modes for single and multiple subdevices 
Single-subdevice prefetch for cmd list copy-only (with bcs) and acc mode.
Multi-subdevice prefetch (default) for shared allocation with multiple BOs.

Related-To: NEO-6740

Signed-off-by: Slawomir Milczarek <slawomir.milczarek@intel.com>
 2022-12-07 15:25:32 +01:00
Warchulski, Jaroslaw
c10aa90815 Cleanup includes 11 
Related-To: NEO-5548
Signed-off-by: Warchulski, Jaroslaw <jaroslaw.warchulski@intel.com>
 2022-12-06 12:25:30 +01:00
Milczarek, Slawomir
f08e552f4d Create multiple-BOs shared allocation in single lmem regions in non-acc mode 
Related-To: NEO-6839

Signed-off-by: Milczarek, Slawomir <slawomir.milczarek@intel.com>
 2022-12-05 11:25:34 +01:00
Jaime Arteaga
99655d34f9 Return a unique pointer for multiple calls to openIpcHandle 
This to follow specification, which says:

zeMemOpenIpcHandle:

- Multiple calls to this function with the same IPC handle will return
unique pointers.

Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
 2022-12-01 18:42:54 +01:00
Spruit, Neil R
d81b0b14a1 L0 Virtual Memory Reservation support 
- Enable support for L0 Virtual Memory reservation on Linux and Windows.
- Excludes support for Linux to allow pStart option

Related-To: LOCI-3397, LOCI-1543

Signed-off-by: Spruit, Neil R <neil.r.spruit@intel.com>
 2022-12-01 00:55:19 +01:00
Jaime Arteaga
4dfdbd612d Fail when handle cannot be obtain for an allocation 
If a handle cannot be obtained, like PRIME_HANDLE_TO_FD, then
properly check for the error and propagate it upwards.

Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
 2022-11-30 20:59:02 +01:00
Mateusz Jablonski
f7c11e8a94 Fix typos 
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2022-11-29 17:12:10 +01:00
Michal Mrozek
a104636b31 Improve address alignment logic. 
Resolves: NEO-7543

Signed-off-by: Michal Mrozek <michal.mrozek@intel.com>
 2022-11-24 14:51:50 +01:00
Dunajski, Bartosz
f267c34327 Check completion fence for AllocationType::DEFERRED_TASKS_LIST 
Related-To: NEO-7458

Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
 2022-11-23 14:18:35 +01:00
Compute-Runtime-Validation
565d44e59c Revert "Fail when handle cannot be obtain for an allocation" 
This reverts commit 4391ad21bb.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
 2022-11-23 05:45:01 +01:00
Jaime Arteaga
4391ad21bb Fail when handle cannot be obtain for an allocation 
If a handle cannot be obtained, like PRIME_HANDLE_TO_FD, then
properly check for the error and propagate it upwards.

Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
 2022-11-22 19:29:19 +01:00
Slawomir Milczarek
ba003dd7dc Prefetch kmd-migrated shared allocation to multiple subdevices 
Related-To: NEO-6740

Signed-off-by: Slawomir Milczarek <slawomir.milczarek@intel.com>
 2022-11-17 15:24:54 +01:00
Mateusz Jablonski
6648adef15 Handle error in register memory allocation methods 
Related-To: NEO-7412
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2022-11-15 17:23:55 +01:00
Szymon Morek
3281dea1fe Allow debug allocs to use copyMemoryToAllocation 
Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2022-11-14 17:50:32 +01:00
Michal Mrozek
508ad66313 Align GPU VA to previous power of 2. 
Aligning to next power of 2 was excessive.
We really need previous power of 2.

Signed-off-by: Michal Mrozek <michal.mrozek@intel.com>
 2022-11-10 16:57:57 +01:00
Lukasz Jobczyk
ff500e0de6 Add support for external hostptr in shared allocations 
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
 2022-11-09 14:48:01 +01:00
Szymon Morek
c9fad505d7 Replicate CONSTANT_SURFACE across tiles 
Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2022-11-08 20:35:43 +01:00
Michal Mrozek
d9762c0337 Add debug flag to control alignment. 
Signed-off-by: Michal Mrozek <michal.mrozek@intel.com>
 2022-11-08 19:42:19 +01:00
Wrobel, Patryk
4c05a54c2b Use correct heap when opening device memory IPC handle 
This change extends the logic of opening IPC
memory handles and fills the gap between this
logic and allocation of USM device buffers.

When HEAP_EXTENDED is available, then it is
preferred.

Signed-off-by: Wrobel, Patryk <patryk.wrobel@intel.com>
 2022-11-08 12:08:04 +01:00
Patryk Wrobel
9b26e96b11 Do not copy std::vector of OS handles when it is not needed 
This change replaces unneeded copying of std::vectors
with usage of const references. Furthermore, it adds
reserve() call before filling the container via push_back().

Signed-off-by: Patryk Wrobel <patryk.wrobel@intel.com>
 2022-10-21 10:28:58 +02:00
Mateusz Hoppe
0fdb308207 Use correct context for validateHostPtr 
- with per context VMs, context with subdeviceBitfield must
be selected

Resolves: NEO-7426

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
 2022-10-18 13:10:28 +02:00
Milczarek, Slawomir
9bc7c48206 Enable kmd-migrated shared allocation with multiple BOs by default 
Related-To: NEO-7395

Signed-off-by: Milczarek, Slawomir <slawomir.milczarek@intel.com>
 2022-10-14 09:51:14 +02:00
Milczarek, Slawomir
6edbfa04a0 Add regkey to create kmd-migrated shared allocation with multiple BOs 
Related-To: NEO-7395

Signed-off-by: Milczarek, Slawomir <slawomir.milczarek@intel.com>
 2022-10-11 16:16:43 +02:00
Jaime Arteaga
a3b3f3e86e Add support for PRELIM_I915_PARAM_SET_PAIR 
This extension allows pairing two buffer objects so they can be
exported using a single dma-buf handle. When imported, a single
buffer object is created with a total size of the two buffer
objects.

Related-To: LOCI-3355

Sync to
https://github.com/intel-gpu/drm-uapi-helper/releases/tag/v2.0-rc15

Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
 2022-09-29 23:41:39 +02:00
Michal Mrozek
2fbc1f652b Choose alignment as next power of 2 for HEAP_EXTENDED allocations. 
This way we will get as big pages as possible without leftovers.

Signed-off-by: Michal Mrozek <michal.mrozek@intel.com>
 2022-09-26 16:42:59 +02:00
Luzynski, Sebastian Jozef
bac85ddb25 Move kernel globals from SVM to USM device 
With this change, module's data sections will be allocated in USM device
pool instead of SVM or USM shared.

Signed-off-by: Luzynski, Sebastian Jozef <sebastian.jozef.luzynski@intel.com>
 2022-09-23 23:06:15 +02:00
Michal Mrozek
3d5e34f727 Reduce the size of masks to 4. 
32 is not required.

Signed-off-by: Michal Mrozek <michal.mrozek@intel.com>
 2022-09-19 21:53:40 +02:00
Compute-Runtime-Validation
45c8124d8f Revert "Move kernel globals from SVM to USM device" 
This reverts commit 706a5a7a8c.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
 2022-09-18 12:49:44 +02:00
Sebastian Luzynski
706a5a7a8c Move kernel globals from SVM to USM device 
With this change, module's data sections will be allocated in USM device pool
instead of SVM or USM shared.

Signed-off-by: Sebastian Luzynski <sebastian.jozef.luzynski@intel.com>
 2022-09-15 16:50:12 +02:00
Mateusz Hoppe
75b117b82f L0Debug - fixes for ZE AFFINITY MASK 
- pass correct deviceIndex when notifying cmdQs
- fix allocating debug surface - single tile allocation when single
bit set in storageInfo

Related-To: NEO-6980

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
 2022-08-31 11:43:19 +02:00
Dunajski, Bartosz
4809a2a9b3 Allow debug flag to apply allocation alignment for heaps 
Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
 2022-08-29 15:17:56 +02:00
Neil R Spruit
ada9b5d4a9 Handle if L0 Event Memory is Shareable 
- Properly check for IPC event handle flag to determine if the event
pool memory is sharable between processes.
- Given Host Visible Event Pool, a check is done to determine if the
Host memory can be shared between the processes.
- Enabled handling if Event Host Memory is shareable for DRM
- If Event Pool Memory is Not shareable, then retrieving the IPC Event
Pool Handle returns unsupported.

Signed-off-by: Neil R Spruit <neil.r.spruit@intel.com>
 2022-08-05 00:11:05 +02:00
Milczarek, Slawomir
9f36b20423 Bind buffer object before calling vm prefetch 
Related-To: NEO-6740

Signed-off-by: Milczarek, Slawomir <slawomir.milczarek@intel.com>
 2022-08-01 15:03:21 +02:00
Mateusz Jablonski
db2d1a53b9 Cleanup includes in drm_memory_manager.h 
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2022-07-26 20:53:21 +02:00
Mateusz Jablonski
c94b6581c2 Make drm memory manager independent on drm headers 
Related-To: NEO-6999
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2022-07-26 20:46:22 +02:00
Mateusz Jablonski
b2489a4c28 Add enum value for i915 mmap offset wb and wc 
Related-To: NEO-6999
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2022-07-26 18:28:08 +02:00
Mateusz Jablonski
3d1c990e37 Move set domain cpu logic to ioctl helper 
Related-To: NEO-6999
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2022-07-25 16:16:34 +02:00
Mateusz Jablonski
b49e2237c5 Pass drm to MemoryInfo ctor 
Related-To: NEO-6999
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2022-07-25 15:44:10 +02:00