German Andryeyev
d17108e8d0
SWDEV-303560 - Remove coarse grain setup by default
...
The original logic was left after initial testing when HMM
couldn't handle xnack properly
Change-Id: I0abf01805704171e931dfba8b6d95bfe87d5fab1
2021-10-05 17:20:59 -04:00
kjayapra-amd
3081f7ca53
SWDEV-295277 - Report max waves per cu from ROCr backend.
...
Change-Id: Ie170b26b53f1cc2da851034c96b21de38ce7b563
2021-10-05 12:38:44 -04:00
anusha GodavarthySurya
34e86bf0c3
SWDEV-24806 - Added support to update memory command params
...
Change-Id: Ib518eaedeeb820023a05278a017a9716e5601dca
2021-10-05 10:51:14 -04:00
Tao Sang
10abe8ab37
SWDEV-305884 - Clear up codes
...
Fix a log typo error
Change-Id: I887ecbdcfe414c2119247228bdd1255b8308da1d
2021-10-04 18:11:32 -04:00
pghafari
1cb0641831
SWDEV-245531 - GLInterop: removing GL dependency
...
Change-Id: I58133e330f5f5e90965fc05003c2d901d4ae6382
2021-10-04 16:35:40 -04:00
German Andryeyev
9a9d10a10b
SWDEV-296301 - Avoid deadlocks in the hostcall path
...
Change the scope of hostcall buffer access lock during destruction.
Make sure wait() returns the signal value after timeout. That
matches ROCr behaviour for HSA signal wait.
Change-Id: I3df34207e0c2e21972ec8052777e5742bda1dca0
2021-10-04 15:00:44 -04:00
Julia Jiang
3098324ef7
SWDEV-302493 - Comgr API backward compatible
...
Change-Id: I3d796d10fa1dbb7edb0510f8ed7d44e3b4993bd7
2021-09-29 11:43:43 -04:00
Chauncey Hui
5788e2b476
SWDEV-2 - Change OpenCL version number from 3372 to 3373
2021-09-28 03:00:06 -04:00
Sarbojit Sarkar
22a847f3ce
SWDEV-301823 - Optimize hipMemset2D/3D
...
Change-Id: Ibe560149a263c2ac6b08e4eb1a1d331d2aeac78c
2021-09-27 14:10:06 -04:00
Sourabh
cbb8d82bdb
SWDEV-292525 - [vdi] Path to streamOps shaders
...
Implementation to use a blit kernel to perform
a hipStreamWait/write instead of an AQL packet.
Change-Id: I462671ed5cec37144dfe97ff66439249196117c1
2021-09-27 13:59:35 -04:00
Chauncey Hui
60295a9da9
SWDEV-2 - Change OpenCL version number from 3371 to 3372
2021-09-25 03:00:07 -04:00
Jason Tang
e0bd4aad63
SWDEV-1 - Only check agent-binary ISA compatibility for online device
...
Change-Id: I5c8395c646280b8316e2f1af9f1e624359166cba
2021-09-23 16:03:01 -04:00
Chauncey Hui
6aca5576b1
SWDEV-2 - Change OpenCL version number from 3370 to 3371
2021-09-23 03:00:06 -04:00
German Andryeyev
6b47e50350
SWDEV-292408 - Keep tracking of subbuffers for multiple devices
...
Change-Id: I568b5ea79614ef507cf1f0efb3dca38ab6cc2db7
2021-09-22 19:24:39 -04:00
Chauncey Hui
f27d17acd3
SWDEV-2 - Change OpenCL version number from 3369 to 3370
2021-09-22 03:00:07 -04:00
Jason Tang
f212fc91ca
SWDEV-1 - More 'delete' clean up
...
info_.extensions_ and settings_ are deleted at amd::Device()::~Device().
Change-Id: I06f240a42e5c131dbd4e61a759f905bcdf84b45a
2021-09-21 11:17:24 -04:00
Chauncey Hui
a0c9bf4414
SWDEV-2 - Change OpenCL version number from 3368 to 3369
2021-09-21 03:00:05 -04:00
German Andryeyev
51556711dc
SWDEV-292408 - Don't force high clock for HIP
...
Change-Id: I501f4e5272124025068b1d3cb637ee8061b06467
2021-09-20 18:44:11 -04:00
German Andryeyev
6da9d18140
SWDEV-292408 - Disable cache coherency tracking for HIP
...
Cache coherency layer is OCL feature to support multiple devices in
single OCL context.
Change-Id: Ic66df9551fad5b0c4df95ab3e1db1da259919f25
2021-09-20 16:13:36 -04:00
Chauncey Hui
41fdd996cb
SWDEV-2 - Change OpenCL version number from 3367 to 3368
2021-09-17 03:00:05 -04:00
agunashe
82e73c096b
SWDEV-301069 - current device id saved while allocating memory in VDI
...
Change-Id: Ic5426895fb1d152e4e6e1baf8b938f35c85fa3d8
2021-09-16 14:00:07 -04:00
Chauncey Hui
c3be1d41f1
SWDEV-2 - Change OpenCL version number from 3366 to 3367
2021-09-16 03:00:09 -04:00
German Andryeyev
f116959b54
SWDEV-302383 - Get active state from device
...
The queue can be destroyed at the time the app will request
the event status. Hence just get the active state from the device.
Change-Id: I887ecb0cfe414c2119247228b0d1255b8308da1e
2021-09-14 19:01:44 -04:00
Chauncey Hui
e6c0e0fe8b
SWDEV-2 - Change OpenCL version number from 3365 to 3366
2021-09-14 03:00:07 -04:00
German Andryeyev
d8a86e4870
SWDEV-282419 - Use HSA_AMD_SVM_ATTRIB_AGENT_ACCESSIBLE for unset
...
When unsetting runtime should use HSA_AMD_SVM_ATTRIB_AGENT_ACCESSIBLE
for the agent and not HSA_AMD_SVM_ATTRIB_AGENT_ACCESSIBLE_IN_PLACE
Change-Id: I3814802d1fb3b72c54e7566defafafed6b0d5cee
2021-09-13 15:05:20 -04:00
German Andryeyev
65ddfcc6a8
SWDEV-294669 - Keep one more slot for HW processing
...
The original logic left only one slot for HW processing in the queue.
For some reason there is a race condition on CPU overwrite of the slot
before the current active. The workaround is to avoid the previous to
the current active slot for possible unfinished HW processing.
Change-Id: I565495a8feeaedffc9fc8a505edbee5ff5816975
2021-09-13 13:56:05 -04:00
Chauncey Hui
7102264d6a
SWDEV-2 - Change OpenCL version number from 3364 to 3365
2021-09-10 03:00:05 -04:00
Jason Tang
73967c3b17
SWDEV-1 - Some 'delete' clean up
...
Change-Id: I02564f0f0e349375bde1471e9f82df268703367b
2021-09-09 12:12:40 -04:00
Jason Tang
5549007856
SWDEV-1 - switch to C++17
...
Change-Id: I7e6a28a20ca163fe86dd6d53b031e9bb6f9b22c9
2021-09-09 10:16:14 -04:00
Chauncey Hui
df862e7ef6
SWDEV-2 - Change OpenCL version number from 3363 to 3364
2021-09-09 03:00:06 -04:00
Vladislav Sytchenko
d934612948
SWDEV-1 - Prepare for c++17 switch
...
std::mem_fun() and std::bind2nd() are removed in c++17. Switch to
simpler logic that does not require those functions.
Change-Id: I19a31f076e1813e367615bd377b424046ce144c7
2021-09-08 16:18:33 -04:00
Vladislav Sytchenko
dd23379ac8
SWDEV-301636 - Manually find NUMA
...
CMake does not provide a way to query the NUMA library, hence we need
to find it manually.
Change-Id: I370b286acdee75cbebc21340da3c432c79f8ffa7
2021-09-08 15:04:15 -04:00
Vladislav Sytchenko
c68f024b35
SWDEV-1 - Fix Windows build
...
std: :mem_fun() is removed in c++17. Simplify logic to not require it.
Change-Id: Ic9a4753b48dd13fcb20cd5b90ff73c3df3211b9f
2021-09-08 12:59:48 -04:00
Saleel Kudchadker
21ba34d0fe
SWDEV-297448 - Add 64bit and 16bit write support
...
For the fillBuffer shader, if there are two 32bit writes to a MMIO
register, it can get dropped. It has to be a single 64bit write.
Add optimization to fillBuffer to write 64bit and 16bit writes.
Change-Id: I3aa78e027898f8ae01e9c8f09004615673720c2b
2021-09-08 12:30:04 -04:00
Chauncey Hui
1786fd8851
SWDEV-2 - Change OpenCL version number from 3362 to 3363
2021-09-08 03:00:04 -04:00
Jatin Chaudhary
b5a57327bb
SWDEV-299162 - Add new comgr API to demangle names
...
Change-Id: I84174d7e40e9afff9a8ea135abadb69d580478e0
2021-09-08 00:35:34 -04:00
Sarbojit Sarkar
42d33029dc
SWDEV-300655 - Added thread ID to hip trace
...
Change-Id: I9234d4ec93e7687cd0a5d1bd930bd4f80936311b
2021-09-06 00:22:42 -04:00
Chauncey Hui
1b218d6861
SWDEV-2 - Change OpenCL version number from 3361 to 3362
2021-09-04 03:00:05 -04:00
Saleel Kudchadker
e29b9c00ee
SWDEV-301667 - Kern arg placement
...
Add a env var ROC_USE_FGS_KERNARG to toggle kernel arg placement
By default its in Fine Grain Kernel arg segment for supported asics.
Change-Id: I3d57ed69a1a4db2b392b0438ead499f3ddca4716
2021-09-02 12:36:49 -04:00
Chauncey Hui
6ebbc0a8af
SWDEV-2 - Change OpenCL version number from 3360 to 3361
2021-09-02 03:00:06 -04:00
jujiang
f63115cec6
SWDEV-286322 - clean up trailing white space
...
Change-Id: I01f3a559cbd1835aa2fdad7abe2bd685d90fc6a8
2021-09-01 11:45:47 -04:00
Jason Tang
1d0364e590
SWDEV-294768 - Fix PCMark10 performance drop
...
PCMark10 counts the time spent in clCreateKernel as part of execution
time, so as workaround for the PAL path, move code object loading
back to clBuildProgram.
Change-Id: I3b9cf1879ece08ab59f447ec165b0525bc8593a4
2021-09-01 09:25:40 -04:00
Jason Tang
7f83bcdb45
SWDEV-1 - Disable OpenCL support for gfx8 in ROCm path
...
Change-Id: Ie1e0c0d6273edf6b734909447c2a08252cba305b
2021-08-31 12:48:47 -04:00
Chauncey Hui
037fc843ba
SWDEV-2 - Change OpenCL version number from 3359 to 3360
2021-08-30 03:00:04 -04:00
Vladislav Sytchenko
215853fd54
SWDEV-298985 - Calm down build warnings
...
This resolves -Wreorder warning.
Change-Id: I28851d66e19a70c4851ac056819d2daadbdc7113
2021-08-29 13:58:48 -04:00
Chauncey Hui
d5f51db5b1
SWDEV-2 - Change OpenCL version number from 3358 to 3359
2021-08-28 03:00:07 -04:00
Satyanvesh Dittakavi
169cc857fd
SWDEV-298985 - hipMemPrefetchAsync should prefetch the data to the specified destination device
...
Pass the device agent specified by the user to the ROCr api instead of passing the device agent attached to the specified stream
Change-Id: I86c98935b9dc404eaa6d47ccdd082a8c3678fb36
2021-08-27 05:12:07 -04:00
Chauncey Hui
e43e773283
SWDEV-2 - Change OpenCL version number from 3357 to 3358
2021-08-27 03:00:06 -04:00
Satyanvesh Dittakavi
48c1b895c0
SWDEV-299491 - Update data index in GetSvmAttributes appropriately
...
Fixes Seg fault caused when the attribute hipMemRangeAttributeAccessedBy
is queried using hipMemRangeGetAttribute
Change-Id: I2ceb2267d89bfc31a55d9eae2685610c7ad89b1f
2021-08-26 13:36:35 -04:00
Chauncey Hui
1e8da33d49
SWDEV-2 - Change OpenCL version number from 3356 to 3357
2021-08-26 03:00:08 -04:00