Commit Graph

1470 Commits

Author SHA1 Message Date
Lukasz Jobczyk
26221678aa Enable engines round robin assign
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-04-28 11:49:15 +02:00
Lukasz Jobczyk
72f6ef7d2b Do not submit to gpgpu during blit in immediate dispatch mode
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-04-28 11:36:42 +02:00
Krzysztof Gibala
9b778863b4 Store GmmHelper in Gmm class
Store GmmHelper in Gmm class instead of GmmClientContext

Related-To: NEO-6523
Signed-off-by: Krzysztof Gibala <krzysztof.gibala@intel.com>
2022-04-27 15:45:49 +02:00
Lukasz Jobczyk
7eb70775ea Flush caches for cross CSR dependencies
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-04-27 15:03:27 +02:00
Filip Hazubski
3900c9d24a Report to StreamProperties whether large grf should be programmed with SCM
Add helper method to UnitTestHelper to query programmed grf values.

Related-To: NEO-6659

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2022-04-27 13:20:14 +02:00
Compute-Runtime-Validation
903cf766b3 Revert "Move shared helper files to shared directory"
This reverts commit 459524f129.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2022-04-27 13:12:16 +02:00
Krystian Chmielewski
e007ba499f feat(zebin): add support for build options section
This change:
* Adds support for build options section in zebinary - using
build options in binary when rebuilding.
* Appends "-cl-intel-allow-zebin" flag to build options when zebin is
used.

Resolves: NEO-6916

Signed-off-by: Krystian Chmielewski <krystian.chmielewski@intel.com>
2022-04-27 11:26:42 +02:00
Bartosz Dunajski
0d82216f43 Refactor handling unknown builtins
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2022-04-27 10:49:16 +02:00
Filip Hazubski
459524f129 Move shared helper files to shared directory
Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2022-04-27 10:34:42 +02:00
Compute-Runtime-Validation
e8e3355a93 Revert "Assign BCS at first blit enqueue"
This reverts commit 5f5a1a6f81.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2022-04-26 09:40:46 +02:00
Lukasz Jobczyk
5f5a1a6f81 Assign BCS at first blit enqueue
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-04-25 13:48:21 +02:00
Dominik Dabek
d0a32b1437 Fix: svm pointer caching
Given a following sequence of setting kernel arg svm pointers: A, nullptr, A
the kernel arg was not being set to A

Related-To: NEO-6895

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2022-04-25 07:59:47 +02:00
Filip Hazubski
944319b3d9 Correct media compression format for blitter operations on planar images
Set most significant bit for chroma planes.
Move common logic to helper function.

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2022-04-22 17:02:16 +02:00
Katarzyna Cencelewska
91840e6e07 Fix gap for forcing engine by debug flag NodeOrdinal
add scenario when command queue created with properties
and there is also applied NodeOrdinal flag
then value from flag will use to force engine

Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2022-04-22 10:48:04 +02:00
Compute-Runtime-Validation
f3bcbfb305 Revert "Enable engines round robin assign"
This reverts commit b81380a879.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2022-04-21 12:43:28 +02:00
Lukasz Jobczyk
b81380a879 Enable engines round robin assign
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-04-20 11:52:33 +02:00
Jobczyk, Lukasz
93fe8a4794 Optimize gpgpu csr's mutex lock in the enqueue blit
Signed-off-by: Jobczyk, Lukasz <lukasz.jobczyk@intel.com>
2022-04-20 11:40:59 +02:00
Kamil Kopryk
7f08a9d993 Remove adjustHwInfoForIgc helper
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
Related-To: NEO-6738
2022-04-20 10:04:36 +02:00
Mateusz Jablonski
b5b8762faf refactor direct submission: Pass command stream receiver to ctor
Related-To: NEO-6643

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2022-04-19 23:02:55 +02:00
Filip Hazubski
e47385dc11 Refactor ShaderChannelSelect logic for images using CL_LUMINANCE
Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2022-04-19 15:58:38 +02:00
Lukasz Jobczyk
6fdcb83020 Add missing tbx calls to timestamp wait logic
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-04-19 14:24:06 +02:00
Filip Hazubski
15fe7e06ce Add reporting of extension cl_intel_bfloat16_conversions
Resolves: NEO-6885

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2022-04-19 12:33:08 +02:00
Lukasz Jobczyk
7e4262839a Add timestamp wait support for BCS
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-04-15 13:26:23 +02:00
Dominik Dabek
8d1ad5a4f3 Refactor: use stack vector for root device indices
Stack vector will not cause dynamic allocations in most circumstances
ie. number of root device indices not more than 16

Related-To: NEO-6837

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2022-04-14 14:05:42 +02:00
Jaime Arteaga
7d4d22228f Check for null allocation in processResources
Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
2022-04-13 02:36:27 +02:00
Krystian Chmielewski
01a719cf33 Allow for zebin rebuild when IR is present
This commits adds rebuilding zebin binary.
If zebin is built for different device and has SPIRV, then new ze binary
will be built using SPIRV.

Signed-off-by: Krystian Chmielewski <krystian.chmielewski@intel.com>
2022-04-12 08:05:25 +02:00
Compute-Runtime-Validation
e9ebd10ab8 Revert "Correct ShaderChannelSelect value for images using CL_LUMINANCE"
This reverts commit 6f9ce94a17.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2022-04-12 06:47:24 +02:00
Filip Hazubski
6f9ce94a17 Correct ShaderChannelSelect value for images using CL_LUMINANCE
Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2022-04-11 17:13:57 +02:00
Krystian Chmielewski
2c1bfbb5b2 Encode number barriers
When programming number of barriers use BARRIER_SIZE enumeration.
Resolves: NEO-6785

Signed-off-by: Krystian Chmielewski <krystian.chmielewski@intel.com>
2022-04-08 10:32:23 +02:00
Compute-Runtime-Validation
64d3f95410 Revert "Allow for zebin rebuild when IR is present"
This reverts commit 2ae3f3e521.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2022-04-06 23:17:22 +02:00
Zbigniew Zdanowicz
f4407064a4 Refactor store register mem encoder to include partition parameter
Related-To: NEO-6811

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2022-04-06 14:00:56 +02:00
Dominik Dabek
76289be8c1 Avoid unnecessary allocation in enqueueHandler
Related-To: NEO-6837

kernelObjsForAuxTranslation will allocate only if needed

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2022-04-06 13:03:41 +02:00
Krystian Chmielewski
2ae3f3e521 Allow for zebin rebuild when IR is present
Signed-off-by: Krystian Chmielewski <krystian.chmielewski@intel.com>
2022-04-05 10:25:06 +02:00
Dominik Dabek
ff240c66db Update cmake for dynamic allocation tracking
If cmake variable ENABLE_DYNAMIC_MEMORY_TRACKING is true:
- add preprocessor definition to enable tacking
- change linker options to export all symbols

Related-To: NEO-6837

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2022-04-04 12:34:22 +02:00
Lukasz Jobczyk
9ace9356f8 Use timestamp wait only if cache flush not needed
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-04-04 09:36:03 +02:00
Bartosz Dunajski
e32f624bf4 Update extension name
Resolves: NEO-6831

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2022-03-31 17:17:33 +02:00
Dominik Dabek
308f9ccfff Do not allocate dynamic state heap if not needed
Dynamic state heap is only used for sampler data.

Related-To: NEO-6821

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2022-03-31 12:13:13 +02:00
Lukasz Jobczyk
aa46dd9cdf Assign data only to unused barrier timestamps
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-03-30 16:56:06 +02:00
Zbigniew Zdanowicz
9858438121 Limit multiple partition count to compute command lists
Related-To: NEO-6811

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2022-03-29 07:29:08 +02:00
Lukasz Jobczyk
1aeb097a63 Download timestamp allocation before checking completion
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-03-28 15:02:04 +02:00
Konstanty Misiak
174c27eb31 Fix CFEFusedEUDispatch debug flag
Signed-off-by: Konstanty Misiak <konstanty.misiak@intel.com>
2022-03-28 12:32:05 +02:00
Jobczyk, Lukasz
d77a6cbe4b Enable task count update from wait
Signed-off-by: Jobczyk, Lukasz <lukasz.jobczyk@intel.com>
2022-03-28 11:09:55 +02:00
Filip Hazubski
3123ab5bf9 Correct media compression format for planar images
Set most significant bit for chroma planes.

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2022-03-26 21:54:08 +01:00
Mateusz Jablonski
e11eb46bff Unify logic for programming mocs in post sync struct
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2022-03-25 17:01:51 +01:00
Lukasz Jobczyk
a230f267e1 Poll task count indefinitely on high throttle command queue
Resolves: NEO-6781

Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-03-25 10:06:16 +01:00
Compute-Runtime-Validation
0c064ccf4c Revert "Enable task count update from wait"
This reverts commit 5118a5d3a6.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2022-03-25 00:09:52 +01:00
Rafal Maziejuk
dea671a04c Correct IMAGE1D_BUFFER width size calculation in BCS
Buffer's default bytesPerPixel value always equals 1 and as
IMAGE1D_BUFFER is originally an image, X coordinate needs to be
multiplied by bytesPerPixel (implied by image format)
in both copySize and (src/dst)Size.

Signed-off-by: Rafal Maziejuk rafal.maziejuk@intel.com
Related-To: NEO-6134
2022-03-24 17:15:38 +01:00
Artur Harasimiuk
52da8f7260 allow external definition of NEO_KERNELS_BIN_DIR
Related-To: NEO-6783
Signed-off-by: Artur Harasimiuk <artur.harasimiuk@intel.com>
2022-03-24 12:00:46 +01:00
Lukasz Jobczyk
5118a5d3a6 Enable task count update from wait
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-03-23 11:38:50 +01:00
Patryk Wrobel
e4d56cde21 Detect GPU hangs in blocking enqueue handler calls
This change introduces detection of GPU hangs in blocking
calls to enqueueHandler() function. Moreover, usages of
this function template have been revised and adjusted to
check the exit code. Furthermore, enqueueBlit() and
dispatchBcsOrGpgpuEnqueue() functions returns value now.
ULTs have been added to cover new cases.

Signed-off-by: Patryk Wrobel <patryk.wrobel@intel.com>
Related-To: NEO-6681
2022-03-22 21:06:19 +01:00