[petsc-dev] refactoring petsccusp.h needed

Paul Mullowney paulm at txcorp.com
Sat Mar 16 09:36:03 CDT 2013


I agree. Just thought I would share something that works easily now for 
the small nv case.
-Paul
>
> thanks, this should work well for nv = 1, 2, maybe 3. However, it 
> won't help Jose a lot. Clearly, for nv >> 1, there are a bunch of 
> unnecessary loads of xarray. Thus, a two-kernel approach is necessary 
> to handle both extremes (nv \approx 1 and nv >> 1).
>
> Best regards,
> Karli
>
>
>
>
>>> Hi Jose,
>>>
>>>>> Since I just stumbled over VecMDot_SeqCUSP() when interfacing
>>>>> ViennaCL: Do you know what was the reason why the 'old' version was
>>>>> replaced by this expensive call to gemv() including the creation of
>>>>> temporaries, etc.? Just writing a custom kernel with one work group
>>>>> per dot-product should do the job perfectly, shouldn't it?
>>>>
>>>> My fault:
>>>> https://bitbucket.org/petsc/petsc-hg/commits/ec7a7de2acd477e5edd24cc5a3af441ce7a68a36 
>>>>
>>>>
>>>>
>>>> The motivation was that the previous version was even worse for me
>>>> (VecMDot is used a lot in SLEPc and GPU performance was really bad).
>>>> At that time I did not have the time to write a custom kernel. If you
>>>> write one, I could help in testing and measuring performance.
>>>
>>> Thanks for providing the context. It makes sense to me now, because
>>> for eigenvalue computations you typically have a lot more vectors
>>> taking part in mdot as compared to GMRES. This looks like an
>>> archetypal example for using two different kernels: The first is
>>> suitable for 'small' numbers of vectors (GMRES), while the second is
>>> more gemv-like and good for larger vector counts (SLEPc). I'll let you
>>> know as soon as it's ready for testing.
>>>
>>> Thanks and best regards,
>>> Karli
>>




More information about the petsc-dev mailing list