Commit Graph

485 Commits

Author SHA1 Message Date
Mateusz Jablonski
d25237c104 fix: pass ReadOnly flag only for page-misaligned input ptr
Related-To: NEO-12986
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2025-02-11 10:42:42 +01:00
Chandio, Bibrak Qamar
7149743162 fix: Set vmbind user fence when makeMemoryResident
Related-To: NEO-11977, GSD-10293

Signed-off-by: Chandio, Bibrak Qamar <bibrak.qamar.chandio@intel.com>
2025-02-10 14:20:09 +01:00
Szymon Morek
844a934c15 fix: Add debug flag to enable dummy pages
Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2025-02-07 20:25:49 +01:00
Szymon Morek
b11322332c performance: enable staging write for cl buffers
Related-To: NEO-13529

Also, add size threshold on iGPU on Linux,
and disable staging if imported host ptr could
be reused

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2025-02-05 09:41:17 +01:00
Mateusz Hoppe
6e35d055f2 feature: make contextGroupSize dependent on number of processes
Related-To: NEO-12952

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2025-02-03 18:13:33 +01:00
Mateusz Jablonski
a924761ed1 fix: correct flags for creating allocation with existing system memory
on Windows setup ReadOnly flag, it makes there is no ProbeForWrite call within
CreateAllocation2 function call

it doesn't force read-only access from GPU perspective, that access it
defined in MapGpuVa function

Related-To: NEO-12986, HSD-15016952776
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2025-02-03 11:24:55 +01:00
Compute-Runtime-Validation
3556d9ded6 Revert "fix: set ReadOnly flag when creating allocation with existing system ...
This reverts commit bddb24124d.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2025-01-31 12:15:12 +01:00
Compute-Runtime-Validation
d23249b061 Revert "fix: Set vmbind user fence when makeMemoryResident"
This reverts commit 80dc4fb43a.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2025-01-31 11:36:29 +01:00
Mateusz Jablonski
bddb24124d fix: set ReadOnly flag when creating allocation with existing system memory
setting ReadOnly flag makes there is no ProbeForWrite call within
CreateAllocation2 function call

it doesn't make read-only access from GPU perspective, that access it
defined in MapGpuVa function

Related-To: NEO-12986, HSD-15016952776
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2025-01-30 19:43:44 +01:00
Andrzej Koska
1abb48c3e0 Revert "feature: Enable Tile64 Optimization Flag"
This reverts commit 105a586615.

Signed-off-by: Andrzej Koska <andrzej.koska@intel.com>
2025-01-28 22:32:58 +01:00
Chandio, Bibrak Qamar
80dc4fb43a fix: Set vmbind user fence when makeMemoryResident
Related-To: NEO-11977, GSD-10293

Signed-off-by: Chandio, Bibrak Qamar <bibrak.qamar.chandio@intel.com>
2025-01-28 22:04:37 +01:00
Szymon Morek
254e7c5c6a fix: set notify enable flag when flushing monitor fence
Related-To: NEO-13848

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2025-01-24 22:19:24 +01:00
Chodor, Jaroslaw
dd3d5c6460 feature: Add logic for reading loaded library's version
getLoadedLibVersion is intended for reading version of a library that was
already loaded into the process prior to this call (i.e. new instance of
the library willnot be loaded).
Note: Currently, only windows version of this functionality is implemented.
Related-To: GSD-10248

Signed-off-by: Chodor, Jaroslaw <jaroslaw.chodor@intel.com>
2024-12-20 17:29:19 +01:00
Andrzej Koska
105a586615 feature: Enable Tile64 Optimization Flag
Related-To: NEO-12134

Signed-off-by: Andrzej Koska <andrzej.koska@intel.com>
2024-12-16 11:09:26 +01:00
Marcel Skierkowski
e27a6dc280 refactor: use virtualFileSystem in ULTs
reducing the number of tests that have interactions with filesystem.
writeDataToFile() saves filename and content in std::map.
fileExistsHasSize() checks if file was previously written to virtualFileSystem
loadDataFromVirtualFile() fetches data from std::map based on filename

Related-To: NEO-7006
Signed-off-by: Marcel Skierkowski <marcel.skierkowski@intel.com>
2024-12-13 17:57:07 +01:00
Chandio, Bibrak Qamar
8cf4804fcd fix: Overhead in zeDeviceGetGlobalTimestamps
Related-To: NEO-11908

There is overhead when submission method is used for
zeDeviceGetGlobalTimestamps. This fixes it.

Signed-off-by: Chandio, Bibrak Qamar <bibrak.qamar.chandio@intel.com>
2024-12-12 08:54:19 +01:00
Lukasz Jobczyk
c2093990d4 fix: Flush monitor fence only to context where needed
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-12-11 10:06:43 +01:00
Fabian Zwoliński
d2ce3badfc fix: bindlessHeapsHelper handle unavailable external heap
This PR handles the situation in which a component
has reserved a front window space for itself in the external heap,
so that the Compute Runtime cannot access this area.

In such a situation, we perform the following steps:
1. reserve 4GB chunk in heapStandard
2. split our chunk into 2 parts: heapFrontWindow, heapRegular
3. from this point on, map all linearStream allocations in reserved 4GB
chunk

Patch applies to Windows and WSL.
Patch only applies when the bindless global allocator is enabled.

Related-To: HSD-16025889919
Signed-off-by: Fabian Zwoliński <fabian.zwolinski@intel.com>
2024-12-05 14:18:01 +01:00
Chandio, Bibrak Qamar
ab2e831a4a fix: zeDeviceGetGlobalTimestamp to use submisison
Related-To: GSD-10253, GSD-9467, GSD-9381, NEO-11908

When EnableGlobalTimestampViaSubmission is set then
zeDeviceGetGlobalTimestamp uses immediate cmd submission
method to get GPU time.

Signed-off-by: Chandio, Bibrak Qamar <bibrak.qamar.chandio@intel.com>
2024-12-04 19:10:07 +01:00
Szymon Morek
d5e5c8e59e fix: call flushMonitorFence on BCS CSR
Related-To: NEO-12477

Also, make sure that global fence is always resident

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2024-12-03 12:48:54 +01:00
Compute-Runtime-Validation
f5cda5ba53 Revert "fix: call flushMonitorFence on BCS CSR"
This reverts commit fc8aa8246b.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-11-30 22:31:30 +01:00
Szymon Morek
fc8aa8246b fix: call flushMonitorFence on BCS CSR
Related-To: NEO-12477

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2024-11-29 16:31:40 +01:00
Mateusz Jablonski
d4e201db86 test: remove not used usings/typedefs/variables in shared tests
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-11-29 10:46:10 +01:00
Mateusz Jablonski
2039b1c41b refactor: remove not needed code
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-11-28 16:16:30 +01:00
Mateusz Jablonski
974e8ae63f performance: don't call evict with empty handle list
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-11-27 09:25:36 +01:00
Mateusz Jablonski
db6fe7892c fix: remove destroyed allocations from eviction lists
mark explicitly made resident allocations

Related-To: NEO-13246, GSD-10319
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-11-25 22:19:54 +01:00
Wenbin Lu
2ba80ce114 feature: support physical host memory
Related-To: NEO-11981

Signed-off-by: Wenbin Lu <wenbin.lu@intel.com>
2024-11-20 08:19:52 +01:00
Lukasz Jobczyk
7f3896d05f performance: Ensure hostptrs removed before creating new one
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-11-12 13:55:15 +01:00
Marcel Skierkowski
49d999abe6 fix: remove defaultProfilingTimerResolution from RuntimeCapabilityTable
Related-To: NEO-12275
Signed-off-by: Marcel Skierkowski <marcel.skierkowski@intel.com>
2024-11-04 12:02:11 +01:00
Lukasz Jobczyk
e687e11ab1 performance: Add CCS Optimization
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-10-23 11:35:57 +02:00
Maciej Bielski
45e78fea76 fix: use productHelper in getPatIndexInfoString() on Windows
Fix the PAT-index reporting in logger as currently on Windows reported
values are simply wrong.

The changed logic dependends on `RootDeviceEnvironment` and in order to
avoid introducing such dependencies into logger.[ch] the
`logAllocation()` is no longer a member of `FileLogger` but
a free-function instead (and a separate .cpp file). This is important
because the source files `logger.[ch]` are also used by ocloc library
and there is no point to contaminate ocloc code structure with
unnecessary dependencies.

Related-To: NEO-9421
Signed-off-by: Maciej Bielski <maciej.bielski@intel.com>
2024-10-22 19:27:13 +02:00
Wenbin Lu
a8a40d2afd feature: support SVM heap in reserveVirtualMem
Related-To: NEO-11981

Signed-off-by: Wenbin Lu <wenbin.lu@intel.com>
2024-10-22 16:47:14 +02:00
Compute-Runtime-Validation
e10998db45 Revert "performance: Add CCS Optimization"
This reverts commit e7b3a40aa7.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-10-22 05:52:14 +02:00
Michal Mrozek
18d828421d performance: add debug flag to control huge chunk size on wddm.
Signed-off-by: Michal Mrozek <michal.mrozek@intel.com>
2024-10-21 16:51:03 +02:00
Lukasz Jobczyk
8a647f6a39 Revert "performance: Ensure hostptrs removed before creating new one"
This reverts commit 5b2f2f3d83.

Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-10-21 10:36:06 +02:00
Lukasz Jobczyk
e7b3a40aa7 performance: Add CCS Optimization
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-10-21 09:59:14 +02:00
Mateusz Jablonski
4154e6666b fix: correct loading L0 loader functions
- don't load ze_loader.dll from file system
- to perform self-open on Windows use getModuleHandleA with proper module name
- don't free library loaded with getModuleHandleA
- as loader may be not available during runtime teardown:
- load translate handle function during global setup
- load setDriverTeardown function during global teardown
- when loader is not available during teardown, unset translate handle
function

Related-To: GSD-10147

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-10-17 15:39:20 +02:00
Compute-Runtime-Validation
f117b5fb68 Revert "fix: correct loading L0 loader functions"
This reverts commit 9c7b3c5e19.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-10-17 08:20:19 +02:00
Mateusz Jablonski
9c7b3c5e19 fix: correct loading L0 loader functions
on Windows use getModuleHandleA with proper module name
don't load ze_loader.dll from file system

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-10-16 19:53:05 +02:00
Mateusz Jablonski
2ce3022030 refactor: remove not needed code
Related-To: NEO-12949
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-10-15 15:33:12 +02:00
Compute-Runtime-Validation
a55a846284 Revert "fix: update slm size based on gt system info in wddm path"
This reverts commit d433fd585a.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-10-11 03:43:37 +02:00
Mateusz Jablonski
d433fd585a fix: update slm size based on gt system info in wddm path
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-10-10 11:29:19 +02:00
Wojciech Konior
6b40f9bc5a refactor: engineInstancedType removed
Related-To: NEO-12594

Signed-off-by: Wojciech Konior <wojciech.konior@intel.com>
2024-10-09 16:30:48 +02:00
Lukasz Jobczyk
5b2f2f3d83 performance: Ensure hostptrs removed before creating new one
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-10-08 13:04:56 +02:00
Compute-Runtime-Validation
60afb83b3b Revert "performance: Ensure hostptrs removed before creating new one"
This reverts commit a890ed5648.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-10-08 07:45:55 +02:00
Szymon Morek
a915ef4b7b fix: Don't program redundant paging fence semaphores
Related-To: NEO-12197

Don't program semaphore to wait for paging fence if it was
already programmed with the same value

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2024-10-07 11:42:06 +02:00
Mateusz Hoppe
e5658e5327 fix: fix freeing allocations in aub mode
- free in operations handler with aub dump must be called to evict
allocations from residentAllocations list
- prior to this change - stale allocations where in the container
leading to crash in aub modes

Related-To: NEO-12869

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-10-03 17:14:07 +02:00
Compute-Runtime-Validation
41df1a6f47 Revert "feature: support SVM heap in reserveVirtualMem"
This reverts commit bfaeeb01d6.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-10-03 14:53:50 +02:00
Mateusz Jablonski
9819b1e2e4 refactor: remove not needed flag ftrGpGpuMidThreadLevelPreempt
the flag was used for mid thread preemption support on pre-gen12 platforms

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-10-01 12:52:12 +02:00
Szymon Morek
468e309e11 fix: call evict under lock
Related-To: NEO-12657

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2024-09-25 10:26:01 +02:00