Uh oh!
There was an error while loading. Please reload this page.
update dpnp.dot implementation - #1669
Conversation
View rendered docs @ https://intelpython.github.io/dpnp/index.html |
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
vtavana
commented
Feb 1, 2024
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
antonwolfy
commented
Feb 6, 2024
The issue #1693 relate to previous changes in |
* dot_func * using mkl::dotu instead mkl::dotc for complex * fix a test * fix negative strides * add a temporary workaround * address comments * add a TODO comment * call dpt.vecdot for integer data types * update doc string * pass argument by reference * update doc to add boolean dtype --------- Co-authored-by: Anton <100830759+antonwolfy@users.noreply.github.com> ac1fca7
* dot_func * using mkl::dotu instead mkl::dotc for complex * fix a test * fix negative strides * add a temporary workaround * address comments * add a TODO comment * call dpt.vecdot for integer data types * update doc string * pass argument by reference * update doc to add boolean dtype --------- Co-authored-by: Anton <100830759+antonwolfy@users.noreply.github.com> ac1fca7

In this PR,
dpnp.dotis updated. 1) For integer and boolean data types,dpctl.tensor.vecdotform the Data Parallel Control library is used, 2) For real-valued floating point data types,dotroutines from BLAS library of OneMKL are used, and 3) For complex data types,doturoutines from BLAS library of OneMKL are used.Also,
dpnp.matmulis updated to return a 1D array instead of 0D array when one of the inputs is 1D and the other one is inherently 1D.