Enable RISC-V platform support #2
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
There was a missing check here before enforcing a local size that would cause uneven WGs.
cl_khr_sub_group_ballot seems to implicitly assume max 128 sized SGs due to the return value type uint4. However, the max SG size is not stated anywhere in the specs and thus should not affect the other sub group functionality to my understanding. The 128 assumption is used for WI masking for the basic subgroup test, causing a crash when exceeding it. This just ups the limit to 1280 to push the limit up to work around the issue. A proper fix would be to use dynamic bit vector size here or define a max SG size in the specs (which doesn't make sense).
…s: fix Skip logic
mutable_dispatch_global_size was failing because the update_global_size = 3 was not a multiple of the Local WS (original command's LWS = GWS / 64 and PoCL doesn't support non-uniform WGs). Fixed by increasing the GWS to 256K and LWS to 16K. The test should really detect the max Local WS of the device and use multiples of that value - that ensures changing global WS without changing local WS will still result in an acceptable GWS/LWS combination. Fixes another test which had an out-of-bounds access because it hardcoded old value of GWS.
If the queried type was "signed" or "unsigned" (with an implied int), the returned size was zero.
Reverts part of "cl_khr_command_buffer_mutable_dispatch & negative_command_buffer tests: fix Skip logic" commit. This was found by review to be invalid.
the negative_command_buffer_command_copy_image_sync_points_null_or_num_zero test was using incorrect buffer size (1 byte per pixel instead of 4), this resulted in PoCL returning the error: CL_INVALID_VALUE dst_origin+region is outside the dst_buffer which the test assumed is invalid result because it expected CL_INVALID_SYNC_POINT_WAIT_LIST_KHR.
Member
|
Thanks @soccercheng! Can you send this PR directly to the upstream CTS in the KhronosGroup project repo? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This request contains patches for RISC-V target platform support.