Commit Graph

609 Commits

Author SHA1 Message Date
Mateusz Jablonski da957d1a37 refactor: correct naming of enum class constants 1/n
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2023-12-05 14:26:42 +01:00
Dominik Dabek 6cf6a8def8 performance: add power throttling debug key
Set windows process power throttling hint to HIGH on wddm init

Related-To: NEO-8215

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2023-12-04 15:03:33 +01:00
John Falkowski 911acd81a2 feature: add SetBOChunkingSize debug variable
Resolves: NEO-9562

Signed-off-by: John Falkowski <john.falkowski@intel.com>
2023-12-04 06:21:01 +01:00
Compute-Runtime-Validation 9add9f12dc Revert "feature: add SetBOChunkingSize debug variable"
This reverts commit e1df8f9112.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2023-12-03 03:28:16 +01:00
Dunajski, Bartosz 2c921ec940 feature: support to use mi_atomic for signalling in-order counter
Related-To: NEO-7966

Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
2023-12-01 15:35:12 +01:00
John Falkowski e1df8f9112 feature: add SetBOChunkingSize debug variable
Resolves: NEO-9562

Signed-off-by: John Falkowski <john.falkowski@intel.com>
2023-12-01 09:39:47 +01:00
Mateusz Jablonski c9664e6bad refactor: rename global debug manager to debugManager
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2023-11-30 13:00:59 +01:00
Mateusz Jablonski 36194c4e7d refactor: correct variable namings
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2023-11-29 23:49:03 +01:00
Dunajski, Bartosz 5772b17924 refactor: Add debug flag to check Device State on failed Wddm submission
Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
2023-11-29 18:44:25 +01:00
Dunajski, Bartosz aba1cd8f9c feature: improve waiting and signaling Events via KMD calls
Related-To: NEO-8179

Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
2023-11-23 14:53:27 +01:00
Baj, Tomasz c49a9b9787 refactor: remove ReturnSubDevicesAsApiDevices from shared code
Related-To: NEO-9437

Signed-off-by: Baj, Tomasz <tomasz.baj@intel.com>
2023-11-22 15:13:29 +01:00
Kacper Nowak 1b932bf119 fix: allow legacy device binary validation logic for Blender on DG2 and MTL
Temporarily opt-out from additional compatibility checks
on DG2 and MTL for Blender and its derivatives AOT-compiled kernels.
This prevents a long kernel recompilation.

Additionally, same behavior can be enforced for other applications
manually via NEO debug key named DoNotUseProductConfigForValidationWa.

Signed-off-by: Kacper Nowak <kacper.nowak@intel.com>
Related-To: NEO-9240
2023-11-21 16:05:17 +01:00
Dominik Dabek 6562828095 performance: prealloc internal heap on mtl
Preallocate 1 internal heap allocation per csr on mtl

Related-To: NEO-8152

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2023-11-17 13:36:21 +01:00
Dunajski, Bartosz a0beb96db8 feature: initial support for implicit convertion to CounterBased Events
Related-To: NEO-8145

Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
2023-11-08 14:59:12 +01:00
John Falkowski f0175b3916 feature: set device allocation chunking as default
Device allocation chunking only applies for multi-tile mode for implicit scaling

Related-To: NEO-9051

Signed-off-by: John Falkowski <john.falkowski@intel.com>
2023-11-07 10:58:17 +01:00
Zbigniew Zdanowicz 20c3f45998 refactor: add ulls diagnostic flag to select monitor fence input
Related-To: NEO-8395

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2023-11-06 15:09:36 +01:00
Zbigniew Zdanowicz e0ce08bb77 fix: detect gpu hang or page fault at direct submission flush to gpu
Related-To: NEO-8395

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2023-11-06 14:22:02 +01:00
Michal Mrozek ed897c302d performance: Implement V2 version of tg dispatch size algorithm.
Signed-off-by: Michal Mrozek <michal.mrozek@intel.com>
Related-To: NEO-6989

-Prevent imbalance in multi dimensional dispatches
-Make sure to utilize as much Eus as possible
-Prefer highest possible tg dspatch count possible
-Make sure that xe_core doesn't have uneven workgroups
2023-11-03 15:54:04 +01:00
Zbigniew Zdanowicz 19586277ca refactor: add debug flag to control delay after waiting for paging fence on cpu
Related-To: NEO-8395

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2023-11-03 12:49:39 +01:00
Dominik Dabek 39cf653959 performance(ocl): cmd buffer prealloc per cmdqueue
Add mechanism to preallocate cmd buffer allocations in command stream
receiver reusable allocations list per command queue initialized.

This should limit additional allocations during hot loop.

Needs to be enabled in subsequent commits by setting product helper
method.

Related-To: NEO-8152

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2023-10-27 16:56:29 +02:00
Mateusz Hoppe 5d572b9c8f feature: allow freeing memory in aubstream
Related-To: NEO-2707

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2023-10-26 17:16:23 +02:00
Compute-Runtime-Validation 69f614a8c2 Revert "fix: allow legacy device binary validation logic for Blender on DG2 p...
This reverts commit d3d15542fb.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2023-10-24 21:00:19 +02:00
Kacper Nowak d3d15542fb fix: allow legacy device binary validation logic for Blender on DG2 platforms
Temporarily opt-out from additional compatibility checks
on DG2 for Blender AOT-compiled kernels.
This prevents a long kernel recompilation.

Additionally, same behavior can be enforced for other applications
manually via NEO debug key named DoNotUseProductConfigForValidationWa.

Signed-off-by: Kacper Nowak <kacper.nowak@intel.com>
Related-To: NEO-9240
2023-10-23 18:20:37 +02:00
Mateusz Jablonski 8da4a9cbc7 fix: add debug flag to control non walker signalling in in-order cmdlist
set to false by default

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2023-10-23 14:43:40 +02:00
John Falkowski f156a74f54 fix: split chunking prefetch flags
Related-To: NEO-9120

Signed-off-by: John Falkowski <john.falkowski@intel.com>
2023-10-18 19:20:42 +02:00
Dominik Dabek 75c4844987 feature(internal): logging kernel dispatch params
Use debug flag PrintKernelDispatchParameters to print params used in
thread group dispatch size heuristic when encoding kernel dispatch.

Related-To: NEO-6989

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2023-10-17 17:31:54 +02:00
Compute-Runtime-Validation 30b066c40e Revert "fix: synchronize host and device timers to avoid device timer overflow"
This reverts commit dae8c34f81.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2023-10-16 11:16:23 +02:00
Dunajski, Bartosz 0592390e2b refactor: print gmm compression settings
Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
2023-10-16 09:14:52 +02:00
Mateusz Jablonski dae8c34f81 fix: synchronize host and device timers to avoid device timer overflow
Related-To: NEO-8394
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2023-10-13 17:40:45 +02:00
Dunajski, Bartosz 06a02552ce refactor: debug flag to override PAT index for given memory type
Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
2023-10-12 15:47:28 +02:00
Filip Hazubski 08e92d154f fix: Add getDefaultDeviceHierarchy call to GfxCoreHelper
Added getDefaultDeviceHierarchy call that describes default device
hierarchy for a gfx core. Refactored L0 and OCL paths to use this
value by default and override this value when user sets
ZE_FLAT_DEVICE_HIERARCHY environment variable or
ReturnSubDevicesAsApiDevices debug key.

Updated ReturnSubDevicesAsApiDevices to force COMPOSITE device hierarchy
when set to 0.

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2023-10-06 12:32:41 +02:00
Mateusz Jablonski 110164a52a fix: remove invalid std::forward
Related-To: NEO-9038
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2023-10-03 16:41:41 +02:00
Mateusz Jablonski a033df33ff fix: remove preferSmallWorkgroupSizeForKernel method
Related-To: HSD-18033866078
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2023-09-29 11:55:09 +02:00
Dunajski, Bartosz 4e8600d8d0 feature: initial support for RelaxedOrdering of in-order Events chaining
Disabled by default.

Related-To: NEO-7966

Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
2023-09-27 16:45:20 +02:00
Dunajski, Bartosz 42496ac96d feature: initial support for patching regular in-order CmdList
Related-To: NEO-7966

Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
2023-09-21 14:20:50 +02:00
Dunajski, Bartosz b94f58abaa feature: debug flag to enable in-order events
Related-To: NEO-7966

Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
2023-09-21 11:22:48 +02:00
Dominik Dabek 1b7e178b25 performance(ocl): program barrier pc in taskStream
Program barrier to task stream, before next enqueue kernel.
This will reduce the number of batch buffer starts for sequences of
enqueue, barrier, enqueue, ... .

Related-To: NEO-8147

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2023-09-19 11:48:02 +02:00
Mrozek, Michal 451c48fc2f refactor: remove not needed code.
Signed-off-by: Mrozek, Michal <michal.mrozek@intel.com>
2023-09-12 10:51:35 +02:00
Mateusz Jablonski 46288b8efd fix: setup correct non-release key name in getStringWithFlags
unify function for getting env

Related-To: NEO-8347
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2023-09-08 15:39:36 +02:00
Joshua Santosh Ranjan 91784a87cc fix: Return success for system address in setArg
This patch avoids returning error for system addresses in setArg

Related-To: GSD-3597

Signed-off-by: Joshua Santosh Ranjan <joshua.santosh.ranjan@intel.com>
2023-09-08 05:27:55 +02:00
Zbigniew Zdanowicz cb641226b5 fix: add debug key to provide alternative directory for wddm residency logs
Related-To: NEO-8211

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2023-09-01 10:15:09 +02:00
Mateusz Hoppe 9e89704624 feature: debug flag to disable DriverStore path enforcement
Resolves: NEO-8320

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2023-08-31 08:20:53 +02:00
John Falkowski d49190f4ae feature: Add debug/release variables prefixes
Add debug/release variables with prefixes for Level Zero, OpenCL and NEO

Resolves: NEO-6357

Signed-off-by: John Falkowski <john.falkowski@intel.com>
2023-08-22 15:15:45 +02:00
Dunajski, Bartosz 7e6e0da978 feature: flush task count on cmd list hostSynchronize if needed
Related-To: NEO-7966

Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
2023-08-22 14:29:14 +02:00
Dunajski, Bartosz f3b2458a9c fix: Use immediate command queue instead of CSR to obtain TaskCount.
Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
2023-08-21 15:04:46 +02:00
Artur Harasimiuk f6e0c0cf89 Revert "feature: Add debug/release variable prefixes"
This reverts commit ec95d9314a.

Signed-off-by: Artur Harasimiuk <artur.harasimiuk@intel.com>
2023-08-18 12:42:39 +02:00
John Falkowski 2403212dcd fix: chunking prefetch add USER_FENCE
Add USER_FENCE before PREFETCH call and after the BIND

Related-To: NEO-8098

Signed-off by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
Signed-off-by: John Falkowski <john.falkowski@intel.com>
2023-08-17 21:32:47 +02:00
John Falkowski ec95d9314a feature: Add debug/release variable prefixes
Resolves:  NEO-6357
Signed-off-by: John Falkowski <john.falkowski@intel.com>
2023-08-10 14:01:09 +02:00
Fabian Zwolinski 6fca8ee195 refactor: Remove SourceLevelDebugger
Removed:
- SourceLevelDebugger (with tests)
- DebuggerLibrary
- DebuggerLibraryRestore
- debuggerSupported field from hwInfo.capabilityTable
- HasSourceLevelDebuggerSupport matcher
- ExperimentalEnableSourceLevelDebugger debug var
- EnableMockSourceLevelDebugger debug var
- DebuggerOptDisable debug var
- lib_names.h.in file
- third_party/source_level_debugger/igfx_debug_interchange_types.h

Related-To: NEO-7213
Signed-off-by: Fabian Zwolinski <fabian.zwolinski@intel.com>
2023-08-10 11:14:02 +02:00
Dominik Dabek 12ab74fe96 performance: flag to program barrier in task cs
Add debug flag ProgramBarrierInCommandStreamTask to program barrier
pipe control in task command stream instead of csr command stream.
This will reduce the number of batch buffer starts.

Related-To: NEO-8147

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2023-08-02 10:26:34 +02:00