Commit Graph

2793 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
Bartosz Dunajski
4f1262645b refactor: pass extra walker params
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-09-10 15:30:03 +02:00
Filip Hazubski
beaddb55e0 fix: Minor code changes
Remove needless checks for module and cmdList.
Add explicit check for surfaceState in setArgImageWithMipLevel.
Explicitly initialize WddmResidencyController::csr.

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2024-09-10 13:31:15 +02:00
Wenbin Lu
bfaeeb01d6 feature: support SVM heap in reserveVirtualMem
Related-To: NEO-11981

Signed-off-by: Wenbin Lu <wenbin.lu@intel.com>
2024-09-09 23:22:04 +02:00
Zbigniew Zdanowicz
0628d97ee1 feature: update processing kernel residency
- save position of kernel internal container when allocation can change
- reuse the same position when new allocation arrives
- add index container for additional allocation of image argument
- save position of additional allocation of image argument
- reuse position when for new image argument

Related-To: NEO-11719

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2024-09-09 14:10:02 +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
Fabian Zwoliński
38e1614f4a fix: create and use new allocation type for syncBuffer
Related-To: NEO-11533
Related-To: HSD-18039788811
Signed-off-by: Fabian Zwoliński <fabian.zwolinski@intel.com>
2024-09-06 14:11:33 +02:00
Lukasz Jobczyk
a54a3bf624 performance: Optimize heap handling when mitigate dc flush
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-09-06 04:33:41 +02:00
Kamil Kopryk
95b035a071 fix: global stateless heap creation in heapless path
Related-To: NEO-10681
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
2024-09-05 16:56:00 +02:00
Andrzej Koska
b0e7a11e9a refactor: Improving information transfer about the copy engine
Related-To: NEO-11934

Signed-off-by: Andrzej Koska <andrzej.koska@intel.com>
2024-09-05 16:11:52 +02:00
Compute-Runtime-Validation
d842f65cf1 Revert "fix: submit dummy exec to pin memory during zeContextMakeMemoryReside...
This reverts commit f9b87d53e6.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-09-05 03:28:03 +02:00
Mateusz Jablonski
14fd9f3f24 fix: correct using L0 loader functions
use zelLoaderTranslateHandle for translating handle to internal handle
get pointer to zelSetDriverTeardown during global ctor
don't load loader library by name
get loader function pointers directly from current process

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-09-04 17:30:25 +02:00
Maciej Plewka
f9b87d53e6 fix: submit dummy exec to pin memory during zeContextMakeMemoryResident call
Signed-off-by: Maciej Plewka <maciej.plewka@intel.com>

Related-To: NEO-11879
2024-09-04 14:07:29 +02:00
Tomasz Biernacik
c5e174da00 refactor: remove queue extended methods and files
Related-To: NEO-12103

Signed-off-by: Tomasz Biernacik <tomasz.biernacik@intel.com>
2024-09-04 11:11:33 +02:00
Zbigniew Zdanowicz
11405af267 performance: use simple if operation instead of std max
Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2024-09-04 08:31:13 +02:00
Mateusz Jablonski
383d786f31 fix: correct including compiler product helper definition
inl file should be included only in enabler cpp files

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-09-03 20:20:39 +02:00
Compute-Runtime-Validation
99f62ac866 Revert "feature: support SVM heap in reserveVirtualMem"
This reverts commit 93cde3ee12.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-09-03 20:20:25 +02:00
Mateusz Hoppe
37b7caa137 fix: correct program header generation for shared isa allocation
- when kernels share single allocation, LOAD address in program headers
should point to correct virtual address including kernel offset

Related-To: NEO-7788, GSD-9836

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-09-03 17:50:29 +02:00
Mateusz Jablonski
43e3957e66 fix: add nullptr check in cache reservation API
in case of AUB/TBX mode there is no os interface

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-09-03 15:43:35 +02:00
Mateusz Jablonski
7adfa344c0 performance: get reference from getBindlessInfo
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-09-03 12:50:52 +02:00
Wenbin Lu
93cde3ee12 feature: support SVM heap in reserveVirtualMem
Related-To: NEO-11981

Signed-off-by: Wenbin Lu <wenbin.lu@intel.com>
2024-09-03 11:38:51 +02:00
Mateusz Jablonski
c0cec096a5 fix: don't return gtpin init error from L0 API call
gtpin should be transparent for application

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-09-02 17:41:08 +02:00
Mateusz Hoppe
d5e275cb49 performance: optimize surface state programming
- eliminate read-modify-write on gfx memory when global bindless mode
enabled

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-09-02 15:31:30 +02:00
Aravind Gopalakrishnan
20aa853369 feature: Append recorded commandlist into immediate (5/N)
- add support for heapless mode

Related-To: NEO-10356

Signed-off-by: Aravind Gopalakrishnan <aravind.gopalakrishnan@intel.com>
2024-08-30 09:23:59 +02:00
Mateusz Jablonski
dec502e11b fix: add magic identifier to L0 command queue object
add tests for toInternalType function

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-08-29 16:39:29 +02:00
Aravind Gopalakrishnan
a9cb7876a5 feature: Append recorded commandlist into immediate (4/N)
- add changes for copy queue support for immediate cmdlist submit

Related-To: NEO-10356

Signed-off-by: Aravind Gopalakrishnan <aravind.gopalakrishnan@intel.com>
2024-08-29 09:35:31 +02:00
Fabian Zwoliński
6a50320f3c fix: use gpuAddress when patching sync buffer
Related-To: NEO-11533
Related-To: HSD-18039788811
Signed-off-by: Fabian Zwoliński <fabian.zwolinski@intel.com>
2024-08-28 17:47:35 +02:00
Mateusz Jablonski
d45c16dfc2 fix: add fallback for invalid handles in extension functions
handle context, commandlist, driver, device, event, image and kernel handles

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-08-28 17:22:35 +02:00
Bartosz Dunajski
db611962f7 fix: improve task count handling in tbx download path
Related-To: HSD-18039789178

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-08-28 15:32:15 +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
ce07b86d87 fix: return valid LP csr from root device
Related-To: NEO-12257

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-08-27 23:33:57 +02:00
Mateusz Jablonski
b7cedb99a4 fix: move gtpin initialization to first zeDriverGet call
Currently gtpin is intialized during zeInit call. However, it is not
correct as gtpin is calling other L0 API functions, like zeDriverGet
which should not be called until zeInit function finishes.

With this commit gtpin initialization is performed during first
zeDriverGet function call. In case of initialization failure
ZE_RESULT_ERROR_DEPENDENCY_UNAVAILABLE will be returned every
zeDriverGet call.

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-08-27 15:09:04 +02:00
Bartosz Dunajski
85359331ce fix: ensure couterbased timestamps writes before access
Related-To: HSD-18039452525

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-08-27 15:04:32 +02:00
Bartosz Dunajski
1795642da9 fix: disallow enabling synchronized dispatch on non-ioq queues
Related-To: HSD-18039265269

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-08-23 19:30:24 +02:00
Mateusz Hoppe
c9e8931a18 fix: global bindless offsets for implicit args
- do not patch bindless offsets for implicit args in with global bindless
allocator, the offsets are patched at kernel initialization time and
do not change

Related-To: NEO-7063

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2024-08-23 13:21:27 +02:00
Bartosz Dunajski
696b02bfd3 fix: improve TBX downloading after L0 Event sync
Related-To: HSD-18038498579

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-08-23 10:42:17 +02:00
Mateusz Jablonski
5b51146673 fix: fail L0 device creation when invalid state save area header was delivered
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-08-22 16:24:40 +02:00
Bartosz Dunajski
716c557cdb feature: update region barrier size allocation
Related-To: NEO-11031

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-08-22 12:21:11 +02:00
Winston Zhang
0590b34cfa feature: refactor and rewrite setErrorDescription
Related-To: NEO-8379

Signed-off-by: Winston Zhang <winston.zhang@intel.com>
2024-08-21 17:26:25 +02:00
Szymon Morek
f2fd669011 fix: call wait after post-processing
Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2024-08-21 12:46:18 +02:00
Mateusz Jablonski
579af57161 refactor: don't call OsLibrary::load directly, use function pointer
this allows mocking this call in ULT

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-08-20 08:59:26 +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
Aravind Gopalakrishnan
cb8063f71d feature: Append recorded command list into immediate (3/N)
- Use correct stream for dispatch
- Add capability to append signal event
- Check available space globally in immediate append call

Related-To: NEO-10356

Signed-off-by: Aravind Gopalakrishnan <aravind.gopalakrishnan@intel.com>
2024-08-16 17:40:28 +02:00
Compute-Runtime-Validation
9b652f4a34 Revert "feature: Improving information transfer about the copy engine"
This reverts commit 17ffdff4f1.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-08-15 22:06:31 +02:00
Szymon Morek
26769205a9 fix: unlock mutex before wait
Signed-off-by: Szymon Morek <szymon.morek@intel.com>
2024-08-14 14:27:43 +02:00
Andrzej Koska
17ffdff4f1 feature: Improving information transfer about the copy engine
Related-To: NEO-11934

Signed-off-by: Andrzej Koska <andrzej.koska@intel.com>
2024-08-14 11:28:29 +02:00
Compute-Runtime-Validation
1cca5e24c1 Revert "performance: disable dispatch task count update when direct submissio...
This reverts commit 6d486459fe.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-08-13 16:37:31 +02:00
Kamil Kopryk
c4850dbd4b refactor: set containsAnyKernel also in heapless
Related-To: NEO-10641
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
2024-08-13 14:44:36 +02:00
Zbigniew Zdanowicz
6d486459fe performance: disable dispatch task count update when direct submission active
Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2024-08-13 13:19:26 +02:00
Bartosz Dunajski
d76ac1d1de fix: scratch controller residency
Related-To: HSD-18039519400

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2024-08-09 14:47:56 +02:00