Commit Graph

4725 Commits

Author SHA1 Message Date
Compute-Runtime-Validation
f5c433c8f8 Revert "performance: Use lock pointer copy with sfence for dc flush mitigation"
This reverts commit 8c3c703ec0.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-10-26 14:28:26 +02:00
Young Jin Yoon
ebdded1bb9 fix: change error message for GPU page fault
Change the error message for GPU page fault to match
with the message from gdb output

Related-To: NEO-13093
Signed-off-by: Young Jin Yoon <young.jin.yoon@intel.com>
2024-10-26 01:29:50 +02:00
Dominik Dabek
741101551e fix: add infrastructure to limit device usm reuse max memory used
Related-To: NEO-12924

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2024-10-25 21:54:41 +02:00
Kamil Kopryk
3891e887c1 feature: use heapless builtins for images
Related-To: NEO-12744
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
2024-10-25 16:42:27 +02:00
Zbigniew Zdanowicz
bbdf1ac7b6 refactor: change encoder for thread group over dispatch 1/n
- change method name to more meaningful
- add all inputs of the algorithm as explicit function arguments
- position all implementations accordingly
- rename unit test names to fit new method name
- fix unit test to have correct initial command values
- fix unit test to have consistent input values with command values
- fix unit test to change input values together with command values

Related-To: NEO-12639

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2024-10-25 16:31:39 +02:00
Lukasz Jobczyk
8c3c703ec0 performance: Use lock pointer copy with sfence for dc flush mitigation
Resolves: NEO-12898

Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-10-25 16:24:40 +02:00
Wenbin Lu
1c46ed9b40 feature: enable reservation from SVM range by default
Related-To: NEO-11981

Signed-off-by: Wenbin Lu <wenbin.lu@intel.com>
2024-10-24 20:01:18 +02:00
Dominik Dabek
22a3ec948b fix: ulls controller sleep, windows
Request higher resolution for windows periodic timers for ulls
controller sleep.

Allows for controller thread to sleep with granularity of 1ms.

Related-To: NEO-10800

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2024-10-24 18:07:04 +02:00
Joshua Santosh Ranjan
c9e48d0d2b refactor: support oa programmable metric group
Related-To: NEO-12184


Signed-off-by: Joshua Santosh Ranjan <joshua.santosh.ranjan@intel.com>
2024-10-24 08:35:12 +02:00
Zbigniew Zdanowicz
a4f34e6fcc refactor: unify programming of preferred slm size 8/n
- remove not needed release helper method

Related-To: NEO-12639

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2024-10-24 08:32:17 +02:00
Filip Hazubski
715313a02c test: Fix typo
Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2024-10-24 03:49:47 +02:00
Naklicki, Mateusz
a429f520fb fix: use proper cpu caching for USM allocations
When using USM allocation, placement of BO can be simultaneously set to
local and system memory. This indicatates that migration between them
is allowed, but then we must not allow CPU caching to prevent corruptions.

Related-To: HSD-18040382956
Signed-off-by: Naklicki, Mateusz <mateusz.naklicki@intel.com>
2024-10-24 03:24:29 +02:00
Chodor, Jaroslaw
5f908ce092 feature: adding support for custom compiler backends
This adds abbility to load different versions of the backend
compiler based on underlying device.

Related-To: NEO-12747

Signed-off-by: Chodor, Jaroslaw <jaroslaw.chodor@intel.com>
2024-10-23 19:55:36 +02:00
Bartosz Dunajski
9d76158c1f feature: debug flag to change ULLS BCS timeout
Related-To: HSD-18040119232

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-10-23 18:18:42 +02:00
Szymon Morek
fca544b178 performance: enable idle csr detection in ULLS controller
Related-To: NEO-12991

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2024-10-23 12:10:45 +02:00
Lukasz Jobczyk
e687e11ab1 performance: Add CCS Optimization
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-10-23 11:35:57 +02:00
Jitendra Sharma
171f1e27a3 fix: Add debug variables for configurable timeouts in debugger
Related-To: NEO-13046
Signed-off-by: Jitendra Sharma <jitendra.sharma@intel.com>
2024-10-23 10:30:03 +02:00
Jaroslaw Warchulski
2ba5ee2f6b Revert "fix: use full size for HEAP_EXTENDED initialization"
This reverts commit 5afc63df93.

Related-To: GSD-8948
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
2024-10-22 19:27:31 +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
Bartosz Dunajski
eed69f45ed fix: Enable 64k pages for TSB allocation
Related-To: HSD-18040274716

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-10-22 15:12:42 +02:00
Szymon Morek
01a0b8e7f7 performance: improve ULLS controller timeout detection
Related-To: NEO-12991

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2024-10-22 13:53:25 +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
Mateusz Jablonski
afe5f105a8 test: don't use pre-gen12 AOT names
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-10-21 18:33:27 +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
Naklicki, Mateusz
a5a11f4a0e feature: add debug key to set MaxSubSlicesSupported
Related-To: HSD-16025421624
Signed-off-by: Naklicki, Mateusz <mateusz.naklicki@intel.com>
2024-10-21 13:34:43 +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
Bartosz Dunajski
ff80a02fcb refactor: parse extra zebin params
Related-To: NEO-12591

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-10-21 09:59:33 +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
Jitendra Sharma
26709ba124 fix: Implement polling of SW FIFO
Related-To: NEO-12955
Signed-off-by: Jitendra Sharma <jitendra.sharma@intel.com>
2024-10-21 07:19:42 +02:00
Jaroslaw Warchulski
5afc63df93 fix: use full size for HEAP_EXTENDED initialization
Related-To: GSD-8948
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
2024-10-18 18:14:34 +02:00
Tomasz Biernacik
46f3d27339 feature: initialize regionCount
Related-To: NEO-12138

Signed-off-by: Tomasz Biernacik <tomasz.biernacik@intel.com>
2024-10-18 13:33:14 +02:00
Oskar Hubert Weber
6730011c3b test: don't load igsc in unit tests
Fixes test timeouts caused by previous changes to dlopen mock

Signed-off-by: Oskar Hubert Weber <oskar.hubert.weber@intel.com>
2024-10-18 12:20:49 +02:00
Mateusz Hoppe
c72ebb263d feature: enable global bindless allocator on xe2hpg
Related-To: NEO-7063, NEO-11533

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-10-18 09:37:37 +02:00
Compute-Runtime-Validation
d7a4dab4ca Revert "performance: Enable timestamp wait for queues on Xe2"
This reverts commit 9185760a0e.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-10-18 08:46:03 +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
Maciej Plewka
9d6d6e85f1 fix: align thread group to dss size if kernel uses slm
Related-To: NEO-12133
Signed-off-by: Maciej Plewka <maciej.plewka@intel.com>
2024-10-17 15:30:19 +02:00
Tomasz Biernacik
c982981dde feature: add number of rt stacks to capability table
Related-To: NEO-12138

Signed-off-by: Tomasz Biernacik <tomasz.biernacik@intel.com>
2024-10-17 14:46:19 +02:00
Chodor, Jaroslaw
5463ddea06 feature: New forward-compatibility model for zeinfo
Up till now, NEO ignored uknown attributes in zeinfo
which could lead to undefined behavior. With this change
NEO will emit an error whenever an unknown attribute is
encountered.

Note : old behavior can be restored using new
IgnoreZebinUnknownAttributes debug environment variable

Resolves: NEO-11762

Signed-off-by: Chodor, Jaroslaw <jaroslaw.chodor@intel.com>
2024-10-17 14:03:01 +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
Chodor, Jaroslaw
e11ceb4a20 feature: Adding support for parsing sip_surface_offset
Adding support for parsing sip_surface_offset in accordance
with zeinfo specification. This field is not used because
surface offset calculations are dynamic and zebin binary can
rely only on sip_surface_bti entry. However, failing to parse
this unused entry blocks NEO from moving to stricter
zebin validation.

Related-To: NEO-11762

Signed-off-by: Chodor, Jaroslaw <jaroslaw.chodor@intel.com>
2024-10-17 00:10:43 +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
Maciej Plewka
deb27d0363 fix: align thread group count to fit within dss
Related-To: NEO-12133
Signed-off-by: Maciej Plewka <maciej.plewka@intel.com>
2024-10-16 18:34:39 +02:00
Szymon Morek
9fe53bdc60 Revert "performance: Enable direct submission on DG2 Windows"
This reverts commit 8c5be5c55a.

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2024-10-16 14:50:09 +02:00
Bartosz Dunajski
52e9a6e07f feature: Initial CB events IPC support
Related-To: NEO-11925

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-10-16 13:33:59 +02:00
Tomasz Biernacik
bf6d9fc416 feature: read new params from blob
Related-To: NEO-12138

Signed-off-by: Tomasz Biernacik <tomasz.biernacik@intel.com>
2024-10-16 10:29:22 +02:00
Jack Myers
10f67bcd1a feature: finalizes 2d block load/store query
Finalized the implementation of the 2d block
load/store extension query. Namely, this adds
the extension info to the `DriverHandleImp`.

Also fixed support matrix in the current
implementation that incorrectly includes
MTL and ARL in the supported products. ULTs
and the implementation have both been changed
to match the true support matrix.

Related-To: NEO-11592
Signed-off-by: Jack Myers <jack.myers@intel.com>
2024-10-16 02:19:22 +02:00
Compute-Runtime-Validation
2098e64dc1 Revert "feature: adding support for custom compiler backends"
This reverts commit 8098bcc48d.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-10-16 02:07:25 +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
Jaroslaw Warchulski
5626295e61 fix: don't split HEAP_EXTENDED among root devices
Related-To: GSD-8948
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
2024-10-15 15:01:29 +02:00