Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Statistical kernels are not supported on integrated graphics #1325

Closed
AlexanderKalistratov opened this issue Mar 3, 2023 · 1 comment
Closed
Assignees

Comments

@AlexanderKalistratov
Copy link
Collaborator

Integrated graphics doesn't support double data type.

cov supports only double input and output:

fmap[DPNPFuncName::DPNP_FN_COV][eft_INT][eft_INT] = {eft_DBL, (void*)dpnp_cov_default_c<double>};

max, min, mean, median, var and std supports only double output for integer input:

fmap[DPNPFuncName::DPNP_FN_MEAN][eft_INT][eft_INT] = {eft_DBL, (void*)dpnp_mean_default_c<int32_t, double>};

@vlad-perevezentsev
Copy link
Collaborator

Most statistical functions have been updated and no longer use the kernels from dpnp_krnl_statistics.cpp (#1396 , #1602, #1632, #1635)
Those functions that still use kernels from this file (like median) have been updated to support float data type to run on Iris Xe #1494.

All statistical functions run on integrated graphics without support for double precision types

Example of running test_statistics on Iris XE

$ pytest -v tests/test_statistics.py

============================================= 1030 passed in 19.66s ==============================================

@AlexanderKalistratov If you don't have any additional questions, could you please close this issue?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants