[GPU] Improve reference kernel coverage. #2366
Conversation
github-actions
bot
added
platform:gpu-intel
|
make test |
There was a problem hiding this comment.
My 2 cents on targeting specific implementations in our ci coverage: it would increase ci testing scope with no clear value since ref impls should be tested by HW configuration that do not dispatch optimized ones.
For that reason, I would keep --global-impl flag out of the input files, and have it used for development purpose only.
| @@ -33,3 +33,19 @@ | |||
| --attr-post-ops=,linear:2:1 | |||
| --runtime_dims_masks=1:0,3:3 | |||
| --batch=shapes_2d_ci | |||
|
|
|||
| # Decompression | |||
There was a problem hiding this comment.
these do not seem to cover decompression as fp-mathmode is not set here. Am I missing something?
| @@ -689,7 +689,7 @@ void def_eltwise_alg_kinds(compute::kernel_ctx_t &kernel_ctx); | |||
|
|
|||
| bool post_ops_with_binary_ok(const primitive_attr_t *attr, | |||
| const data_type_t dst_dt, const int max_ndims_supported = 2, | |||
| const int prelu_mask_supported = 3); | |||
| const int prelu_mask_supported = 0xFFFF); | |||
There was a problem hiding this comment.
should the default value just be dropped if it doesnt restrict any value?
Addresses https://jira.devtools.intel.com/browse/MFDNN-12444.
Broaden ocl:ref coverage and add some smoke tests for key compute bound primitives.