Commit Graph

352 Commits

Author SHA1 Message Date
Dominik Dabek b2fc7345cf performance: redesign usm alloc reuse mechanism
Dedicated pools for different allocations size ranges.
Additional reused allocations will create their own pools.
Do not reuse allocations >256MB.

Related-To: NEO-6893, NEO-12299, NEO-12349

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2024-09-10 18:12:14 +02:00
Damian Tomczak 41e24635e3 fix: correct maxMemAllocSize
Resolves: NEO-11857

Signed-off-by: Damian Tomczak <damian.tomczak@intel.com>
2024-09-10 09:15:18 +02:00
Mateusz Hoppe 9d5f8eb587 refactor: save hpCopyEngine in Device
- remove loop iteration to get hpCopyEngine

Related-To: NEO-11983

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-09-06 17:56:53 +02:00
Mateusz Hoppe ae3edba369 feature: limit contexts in group based on number of CCS engines
Related-To: NEO-12258, HSD-14023076222

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-08-30 19:55:31 +02:00
Mateusz Hoppe 482e077252 feature: add regular and hp contexts in group without dedicated hp engine
- if no hp copy engine available, create group with regular and hp
contexts

Related-To: NEO-11983

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-08-28 14:01:10 +02:00
Mateusz Hoppe d9864eca7a feature: add context group support for root device engine
Related-To: NEO-12257

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-08-27 10:07:28 +02:00
Lukasz Jobczyk d1c354d5bf fix: Init device with first submission when metrics enabled
Resolves: NEO-7642

Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-08-20 14:34:16 +02:00
Bartosz Dunajski ce0ecf5939 fix: download L0 Event TBX allocation on related subdevice if needed
Related-To: HSD-18038498579

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-08-19 18:34:58 +02:00
Bartosz Dunajski ec34656e0e fix: debug flag to defer first device submission
Related-To: HSD-18039343751

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-08-02 09:31:25 +02:00
Mateusz Hoppe b6299b8a21 feature: add support for HP copy engine context
- add support for contect group with HP copy engine
- choose HP copy engine when available

Related-To: NEO-11983

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-07-19 12:23:03 +02:00
Compute-Runtime-Validation 7ad15639fc Revert "feature: add support for HP copy engine context"
This reverts commit 3fbcbcaef2.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-07-18 21:02:14 +02:00
Mateusz Hoppe 3fbcbcaef2 feature: add support for HP copy engine context
- add support for contect group with HP copy engine
- choose HP copy engine when available

Related-To: NEO-11983

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-07-18 16:07:07 +02:00
Maciej Plewka 85e708819a fix: Add per product cache line size property
Signed-off-by: Maciej Plewka <maciej.plewka@intel.com>
2024-07-18 12:47:47 +02:00
Mateusz Hoppe 14a1bbfe12 feature: allow dynamic count of HP contexts in context group
- HP contexts may use up to half contexts in group

Related-To: NEO-7824

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-07-09 20:41:38 +02:00
Mateusz Jablonski a14d5c52cb fix: Extend scope of OverrideSlmSize debug key - apply key during initialization
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-07-05 18:03:45 +02:00
Filip Hazubski 922286633b fix: Correct debugger and SIP init logic
Initialize debugger and SIP kernel explicitly once during root-device init.

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2024-07-05 12:04:13 +02:00
Jitendra Sharma 51fc9059c2 feature: program debug surface address
Related-To: NEO-7986
Signed-off-by: Jitendra Sharma <jitendra.sharma@intel.com>
2024-07-01 16:30:14 +02:00
Bartosz Dunajski ca77ace883 fix: download TBX Event allocation on Secondary CSRs
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-06-26 16:33:52 +02:00
Bartosz Dunajski 6d70304e7b refactor: debug flag to override num HP contexts
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-06-25 16:46:16 +02:00
Filip Hazubski 230d3e903b fix: Move first submission programming to a separate function
Call the new function after debugger and sip kernel are initialized.

Related-To: HSD-18038645398, HSD-18038819112

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2024-06-25 10:10:02 +02:00
Filip Hazubski 2f6eaf149a fix: Update SIP kernel initialization logic
Initialize SIP kernel when shared device is being initialized
instead of api-specific device.

Initialize debugger when shared device is being initialized
instead of during platform or driver initialization.

Add missing makeResident calls for SIP kernel in heapless paths.

Related-To: HSD-18038645398, HSD-18038819112

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2024-06-20 18:30:46 +02:00
Bartosz Dunajski eff450c5ef fix: ensure CSR is initialized during first submission
Related-To: NEO-7824

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-06-07 13:23:44 +02:00
Bartosz Dunajski 5e8590d513 feature: skip first secondary context if interrupt hint is passed
Related-To: NEO-8179

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-06-07 13:14:23 +02:00
Bartosz Dunajski da9c009b88 feature: assign unique interrupt to queue
Related-To: NEO-8179

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-06-07 10:06:31 +02:00
Bartosz Dunajski e88127dcc3 refactor: remove not used multi context per engine code
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-06-06 11:41:15 +02:00
Bartosz Dunajski bc37e14e87 refactor: fallback paths for secondary context debug flags
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-06-05 12:02:16 +02:00
Bartosz Dunajski 88ed909e57 feature: assign secondary BCS engines at queue creation
Related-To: NEO-7824

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-06-03 09:41:43 +02:00
Bartosz Dunajski fe6809ac04 feature: prework to initialize BCS state in heapless mode
Related-To: NEO-7824

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-05-31 10:06:58 +02:00
Mrozek, Michal 38d7e02c28 refactor: remove not needed code
Signed-off-by: Mrozek, Michal <michal.mrozek@intel.com>
2024-05-29 13:24:02 +02:00
Bartosz Dunajski 913d5dc3b1 feature: create secondary contexts for different engine types
Related-To: NEO-7824

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-05-24 15:14:24 +02:00
Bartosz Dunajski 0a74d718f2 refactor: allow using different engine types for secondary contexts
Related-To: NEO-7824

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-05-24 09:39:46 +02:00
Mateusz Hoppe bfc3e8fcf0 fix: assign unique ContextId for secondary engines
- add secondary engines to allRegisteredEngines

Related-To: NEO-11219

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-05-20 13:19:20 +02:00
Szymon Morek 10ed479b16 performance: share inter-module ISA allocations
Related-To: NEO-10258

Currently each module has it's own GA
for kernel ISA's. This change allows new modules to
reuse existing allocation.

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2024-05-09 08:43:55 +02:00
Mateusz Hoppe d35d8727e5 fix: use primaryCsr allocations when csr has primaryCsr set
- global fence allocation
- global stateless heap allocation
- preemption allocation
- debug surface allocation

all above are shared from primary csr

Related-To: NEO-7824


Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-05-07 02:41:24 +02:00
Szymon Morek 83e8ae4a20 performance: Reuse GPU timestamp instead of KMD escape
This can be enabled only if related
debug flag will be set.

Related-To: NEO-10615

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2024-05-06 14:46:30 +02:00
Katarzyna Cencelewska ce3bb1327e feature: add method to switch enable/disable mid thread preemption
Resolves: NEO-8089

Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2024-04-15 16:26:52 +02:00
Kamil Kopryk cfc3b6a8c2 refactor: add heaplessModeEnabled argument to isHeaplessStateInitEnabled
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
2024-04-11 10:08:10 +02:00
Mateusz Jablonski cb2b572e94 feature: add support for null aub mode
In this mode AUB csr will be created, however, no aub file will be created

Related-To: NEO-11097
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-04-09 16:59:42 +02:00
Mateusz Hoppe d6675b14b3 fix: release secondaryContexts prior to primary context
Related-To: NEO-10920

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-04-05 16:34:39 +02:00
Mateusz Hoppe 0cec5ccc32 fix: check if any of devices has BindlessHeapHelper enabled
- report Bindless_images extension in driver only when at least one
device has global bindless heaps enabled

Related-To: NEO-10352

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-03-20 20:50:58 +01:00
Wenju He 03078541d7 feature: make global bindless heaps resident when created
Make bindless heaps resident right after heap allocation.
Motivation is that SYCL bindless image can be passed as a value argument
or through memory. Therefore, we're not able to make its bindless heap
resident during kernel initialization or setting kernel arguments.

This fixes SYCL bindless image read_write_*D.cpp tests on DG2.

Related-To: NEO-7063
Signed-off-by: Wenju He <wenju.he@intel.com>
2024-03-20 14:40:11 +01:00
Raiyan Latif 40471da591 fix: properly retrieve sub-device count in allocateRTDispatchGlobals
Previous method used for retrieiving sub-device count in this path
did not take into account device affinity mask nor device hiearchy
mode, resulting in segmentation faults when attempting to allocate
rtDispatchGlobals structure using improper deviceBitFields.

Related-To: NEO-8422

Signed-off-by: Raiyan Latif <raiyan.latif@intel.com>
2024-03-20 09:23:30 +01:00
Ben Ashbaugh e00230a6c2 fix: correctly report support for SPIR-V 1.0 through 1.3
Related-To: NEO-10336

Signed-off-by: Ben Ashbaugh <ben.ashbaugh@intel.com>
2024-03-12 17:54:09 +01:00
Kamil Kopryk 168445784e feature: introduce states programming at driver init heapless ocl
Related-To: NEO-7824
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
2024-03-08 12:29:44 +01:00
Mateusz Jablonski 8e7e5ca58f fix: return same UUID format independently on pci bus info presence
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-03-07 16:56:30 +01:00
Dunajski, Bartosz 00964275fb fix: create preemption allocation for Root Csr
Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
2024-02-29 16:46:24 +01:00
Compute-Runtime-Validation 9570d02134 Revert "fix: correctly report support for SPIR-V 1.0 through 1.3"
This reverts commit 5120ec2f93.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-02-15 04:00:31 +01:00
Ben Ashbaugh 5120ec2f93 fix: correctly report support for SPIR-V 1.0 through 1.3
Related-To: NEO-10336

Signed-off-by: Ben Ashbaugh <ben.ashbaugh@intel.com>
2024-02-14 09:21:06 +01:00
Kacper Nowak ab1ffea592 refactor: Create function to check if OCL2.1 caps are enabled or enforced
Signed-off-by: Kacper Nowak <kacper.nowak@intel.com>
2024-02-06 08:27:09 +01:00
Kacper Nowak 6cf71cce81 refactor: Allow for grouping BCS1+ copy engines in more engine group types
Do not group copy engines from BCS1+ in linked copy group by default.

Signed-off-by: Kacper Nowak <kacper.nowak@intel.com>
2024-01-23 12:13:09 +01:00