Commit Graph

12 Commits

Author SHA1 Message Date
German Andryeyev 8904848abc Set CPU access flag for SVM
Make sure all GPUs have CPU access flag for the fine grain buffer.

Change-Id: Ifc843c2807e70a271b269192ae7859205ff458f3
2020-05-13 16:05:46 -04:00
Jason Tang b4f1239f34 device/rocm: split gfxVersion to major/minor/stepping
Change-Id: I1e437eaee30794147713d9516229211670f01d90
2020-05-12 12:17:13 -04:00
Michael LIAO 503ef06555 Clear executable permission.
Change-Id: Ia0d363b1ba89d7947e5b5a55cb67edba86f0515e
2020-05-07 10:38:58 -04:00
kjayapra-amd 7458bf9964 SWDEV-229840 - Improve error messages on ROCCLR Layer.
Change-Id: Iab7d9156cdc206db86385aa05023a0095ed40f92
2020-04-19 20:01:49 -04:00
Vladislav Sytchenko c781f4d419 Don't call updateFreeMemory() if the allocation failed
Change-Id: I978cb2e463914f6a48b3d4a9057c0f67e7bdb646
2020-04-09 18:41:11 -04:00
Alex Xie 43b9863e17 SWDEV-229731 - [Lnx][Rocm][Navi]Support images in full Opencl Conformance tests
1. Enable pitch workaround
2. When we use copy image, we don't need to create the custom pitch image
3. wrtBackImageBuffer_ stores device memory object, not amd image object.

Tests:
conformance kernel read / write test pass with this code change.

Change-Id: I7dca3127adde6ac83e78dd270a2256ebed55c60d
2020-04-04 09:43:03 -04:00
Jason Tang 9cfb973c3f SWDEV-207312 - Fix images_write_cube failure for gfx10
Change-Id: I7f6a8c77eb712103e4c0a5f36bba95093ffbff98
2020-03-12 17:25:52 -04:00
Vladislav Sytchenko 696d00e71b Correct pitch alignment calculation
When we're aligning rowPitch to imagePitchAlignment, rowPitch is in pixels,
but imagePitchAlignment_ is bytes, so we end up overaligning the pitch.

Convert imagePitchAlignment_ to pixels before doing any logic.

Change-Id: Ia5ab9d54bed150fe974e86b060dbadc196165b29
2020-02-20 18:00:27 -05:00
Jason Tang 290b42625a SWDEV-207312 - Comment out incorrect cl-gl inter-op code to facilitate Linux-Pro testing
Change-Id: I273476465332ffcda50a6095f839e1bc07863c8e
2020-02-19 17:56:45 -05:00
German Andryeyev 374f612b7c SWDEV-193956
[hipclang-vdi-rocm][perf]~45% to 50% of Performance drop on
rocBLAS_int8 test

- Enable AMD_OPT_FLUSH optimization by default to match HCC
- Disable CPU writes to GPU memory on boards with large bar,
because it requires HDP flush tracking.
- Enable L2 cache on kernel arguments, because L2 will be
invalidated on memory reuse .

Change-Id: I124cf250bdd4d19c523ce542c163813828f8fbdc
2020-02-18 14:26:00 -05:00
Laurent Morichetti b4c6143a2f Update copyright info
Change-Id: Ia4f9ff0f5f873b4223a8cca154188bb0d2f1abba
2020-02-04 09:26:14 -08:00
Laurent Morichetti 20c7173849 Merge branch 'origin/pghafari/vdi-prototype' into lmoriche/amd-master
Change-Id: Id3b833d405596735becb3346f3b08c6da57033fe
2020-01-30 20:12:13 -08:00