topic Gianluca, yes, we are working in Intel® oneAPI Math Kernel Library

LAPACKE_zgesvxx with small size systems

Gianluca_G_1 — Mon, 03 Apr 2017 10:09:35 GMT

We use LAPACKE_zgesvxx function in order to resolve a huge system of linear equations.

This function is very fast with this kind of matrices, but if we have small matrix Size like < 10x10, it is very slow.
Did you provide any optimized method for this kind of size?

We have to resolve plenty of small small systems.

Best regards

Gianluca

Gianluca, yes, we are working

Gennady_F_Intel — Mon, 03 Apr 2017 15:27:43 GMT

Gianluca, yes, we are working on further improvement such tiny cases. What mkl version do you use?

I downloaded just today the

Gianluca_G_1 — Mon, 03 Apr 2017 16:13:35 GMT

I downloaded just today the MKL version 2017.0.2.1 .

have you tried the

Gennady_F_Intel — Tue, 04 Apr 2017 05:21:39 GMT

have you tried the performance with this update?

Yes, I just verified it, and

Gianluca_G_1 — Wed, 05 Apr 2017 13:17:38 GMT

Yes, I just verified it, and it is much faster than before, but not enough.
If we use a simple LU function written in C# this is still much better.

Let me know if you improve more.

thank you

Gianluca