Intel® oneAPI Base Toolkit
Support for the core tools and libraries within the base toolkit that are used to build and deploy high-performance data-centric applications.
419 Discussions

LAPACKE_cheevx equivalent function in oneMKL DPC++ for GPU

TestDR
Beginner
740 Views

Hi,

Could you please advise when LAPACKE_cheevx equivalent function can be included in oneMKL DPC++ for GPU?

Best regards,

Rajitha

0 Kudos
6 Replies
PraneethA_Intel
Moderator
715 Views

Hi Rajitha,

 

Thanks for posting in Intel communities.

 

>>LAPACKE_cheevx equivalent function in oneMKL DPC++

Could you please let us know if any of the following routines would satisfy your requirements.

  1. heevd
  2. heevs(USM Version)

 

Thanks and Regards,

Praneeth Achanta

 

0 Kudos
TestDR
Beginner
691 Views

Hi,

Thanks for your reply. Cheevx calculate selected number of eigen values whereas heevd calculate them all. We are developing a time sensitive application and we want something like cheevx. When can we expect something like cheevx with GPU support?

Thanks,

Rajitha

0 Kudos
PraneethA_Intel
Moderator
656 Views

Hi Rajitha, 


This is to kindly inform you that, there is no exact date when heevx would be available with GPU support. We will inform you as soon as we have an update. As of now, heevd is the only alternative available. Kindly let us know if we could close this case on our end.


Thanks and Regards,

Praneeth Achanta


0 Kudos
TestDR
Beginner
616 Views

We may evaluate “heevd” for calculating all the eigen values. However, I could not find an example for “heevd” within /opt/intel/oneapi/mkl/2023.0.0/examples/examples_dpcpp.tgz. Can you please point me to an example?  My aim is to see whether it is profitable to offload heevd computation into a GPU using Intel Advisor.

0 Kudos
PraneethA_Intel
Moderator
588 Views

Hi Rajitha,


As per your request, we will be communicating the required information to you privately via email.


Thanks and Regards,

Praneeth Achanta


0 Kudos
PraneethA_Intel
Moderator
563 Views

Hi Rajitha,


As informed earlier, the requested information will be communicated to you privately via email and this thread will no longer be monitored by Intel. If you need any further assistance, please post a new question.


Thanks and Regards,

Praneeth Achanta


0 Kudos
Reply