Dear petsc-dev,
I'd like to try out doing SOR-like smoothing for a problem with MG
where on the pre-smooth I run forwards, and on the post-smooth backwards.
I can do this by setting up the PC and then spinning over the up and
down smoothers separately. I wonder if it would be possible to extend
Lawrence,
You analysis is correct; I don't know what I was thinking twenty+ years ago.
I think a fix could be to have an option that just creates a separate set
of post smoothers and gives them a a different suffix.
I'll try to do it on a branch and you can then follow the bran
On Tue, Feb 13, 2018 at 8:48 PM, Jed Brown wrote:
> Richard Tran Mills writes:
>
> > I haven't experimented very thoroughly with it (hmm... should probably do
> > such experiments), but I believe that, once matrix rows become
> sufficiently
> > long, then SELL doesn't provide an advantage over A
Richard Tran Mills writes:
> On Tue, Feb 13, 2018 at 8:48 PM, Jed Brown wrote:
>
>> Richard Tran Mills writes:
>>
>> > I haven't experimented very thoroughly with it (hmm... should probably do
>> > such experiments), but I believe that, once matrix rows become
>> sufficiently
>> > long, then SE
Hi Jonathan,
thanks for your message and the pointer.
The incomplete factorizations have been around for a while, and with
recent hardware they tend to be less competitive (note that they use a
Tesla 2050 in their benchmarks, which is ~7 years old).
The fine-grained parallel version here:
h
https://bitbucket.org/petsc/petsc/pull-requests/858/add-support-for-selecting-different-down/diff
> On Feb 14, 2018, at 5:50 AM, Lawrence Mitchell
> wrote:
>
> Dear petsc-dev,
>
> I'd like to try out doing SOR-like smoothing for a problem with MG
> where on the pre-smooth I run forwards, an