Re: [petsc-users] hypre-ILU vs hypre Euclid

2023-06-23 Thread Alexander Lindsay 
Based on https://github.com/hypre-space/hypre/issues/937 it sounds like
hypre-ILU is under active development and should be the one we focus on
bindings for. It does support 64 bit indices and GPU

On Fri, Jun 23, 2023 at 8:36 AM Alexander Lindsay 
wrote:

> Thanks all for your replies. Mark, I’m a little unclear on what you said.
> My understanding is that PETSc ILU is serial only (or can be used as the
> sub PC in DD PCs).
>
> On Jun 23, 2023, at 3:35 AM, Mark Adams  wrote:
>
> 
> Alexander, let me add that Ilu is pretty commodity, and is available with
> our vendor, back ends, and that is probably the more reliable route.
> Hyper’s AMG Solver is state of the art, but their ilu is not their focus.
>
> Mark.
>
> On Fri, Jun 23, 2023 at 12:11 AM Jed Brown  wrote:
>
>> It looks like Victor is working on hypre-ILU so it is active. PETSc used
>> to have PILUT support, but it was so buggy/leaky that we removed the
>> interface.
>>
>> Alexander Lindsay  writes:
>>
>> > Haha no I am not sure. There are a few other preconditioning options I
>> will explore before knocking on this door some more.
>> >
>> >  On Jun 22, 2023, at 6:49 PM, Matthew Knepley 
>> wrote:
>> >
>> >  On Thu, Jun 22, 2023 at 8:37 PM Alexander Lindsay <
>> alexlindsay...@gmail.com> wrote:
>> >
>> >  I know that PETSc has hooks for Euclid but I discovered today that it
>> does not support 64 bit indices, which many MOOSE
>> >  applications need. This would probably be more appropriate for a hypre
>> support forum (does anyone know if such a forum
>> >  exists other than opening GitHub issues?), but does anyone here know
>> what the difference between hypre-ILU and
>> >  hypre-Euclid are? From the docs it seems they are both supposed to be
>> parallel ILU solvers.
>> >
>> >  If hypre-ILU worked with 64 bit indices (I can probably check this
>> sifting through the sources), then I would probably add
>> >  hooks for it in PETSc (AFAICT those don't exist at present).
>> >
>> >  My understanding was that two different people were working on them. I
>> do not know if either is still actively supported. We
>> >  would of course like a binding to whatever is supported.
>> >
>> >  Are you sure you want to run ILU?
>> >
>> >THanks,
>> >
>> >   Matt
>> >  --
>> >  What most experimenters take for granted before they begin their
>> experiments is infinitely more interesting than any results to
>> >  which their experiments lead.
>> >  -- Norbert Wiener
>> >
>> >  https://www.cse.buffalo.edu/~knepley/
>>
>

Re: [petsc-users] hypre-ILU vs hypre Euclid

2023-06-23 Thread Alexander Lindsay 
Thanks all for your replies. Mark, I’m a little unclear on what you said. My understanding is that PETSc ILU is serial only (or can be used as the sub PC in DD PCs).On Jun 23, 2023, at 3:35 AM, Mark Adams  wrote:Alexander, let me add that Ilu is pretty commodity, and is available with our vendor, back ends, and that is probably the more reliable route. Hyper’s AMG Solver is state of the art, but their ilu is not their focus.Mark.On Fri, Jun 23, 2023 at 12:11 AM Jed Brown  wrote:It looks like Victor is working on hypre-ILU so it is active. PETSc used to have PILUT support, but it was so buggy/leaky that we removed the interface.

Alexander Lindsay  writes:

> Haha no I am not sure. There are a few other preconditioning options I will explore before knocking on this door some more. 
>
>  On Jun 22, 2023, at 6:49 PM, Matthew Knepley  wrote:
>
>  On Thu, Jun 22, 2023 at 8:37 PM Alexander Lindsay  wrote:
>
>  I know that PETSc has hooks for Euclid but I discovered today that it does not support 64 bit indices, which many MOOSE
>  applications need. This would probably be more appropriate for a hypre support forum (does anyone know if such a forum
>  exists other than opening GitHub issues?), but does anyone here know what the difference between hypre-ILU and
>  hypre-Euclid are? From the docs it seems they are both supposed to be parallel ILU solvers.
>
>  If hypre-ILU worked with 64 bit indices (I can probably check this sifting through the sources), then I would probably add
>  hooks for it in PETSc (AFAICT those don't exist at present).
>
>  My understanding was that two different people were working on them. I do not know if either is still actively supported. We
>  would of course like a binding to whatever is supported.
>
>  Are you sure you want to run ILU?
>
>    THanks,
>
>       Matt
>  -- 
>  What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to
>  which their experiments lead.
>  -- Norbert Wiener
>
>  https://www.cse.buffalo.edu/~knepley/

Re: [petsc-users] hypre-ILU vs hypre Euclid

2023-06-23 Thread Mark Adams 
Alexander, let me add that Ilu is pretty commodity, and is available with
our vendor, back ends, and that is probably the more reliable route.
Hyper’s AMG Solver is state of the art, but their ilu is not their focus.

Mark.

On Fri, Jun 23, 2023 at 12:11 AM Jed Brown  wrote:

> It looks like Victor is working on hypre-ILU so it is active. PETSc used
> to have PILUT support, but it was so buggy/leaky that we removed the
> interface.
>
> Alexander Lindsay  writes:
>
> > Haha no I am not sure. There are a few other preconditioning options I
> will explore before knocking on this door some more.
> >
> >  On Jun 22, 2023, at 6:49 PM, Matthew Knepley  wrote:
> >
> >  On Thu, Jun 22, 2023 at 8:37 PM Alexander Lindsay <
> alexlindsay...@gmail.com> wrote:
> >
> >  I know that PETSc has hooks for Euclid but I discovered today that it
> does not support 64 bit indices, which many MOOSE
> >  applications need. This would probably be more appropriate for a hypre
> support forum (does anyone know if such a forum
> >  exists other than opening GitHub issues?), but does anyone here know
> what the difference between hypre-ILU and
> >  hypre-Euclid are? From the docs it seems they are both supposed to be
> parallel ILU solvers.
> >
> >  If hypre-ILU worked with 64 bit indices (I can probably check this
> sifting through the sources), then I would probably add
> >  hooks for it in PETSc (AFAICT those don't exist at present).
> >
> >  My understanding was that two different people were working on them. I
> do not know if either is still actively supported. We
> >  would of course like a binding to whatever is supported.
> >
> >  Are you sure you want to run ILU?
> >
> >THanks,
> >
> >   Matt
> >  --
> >  What most experimenters take for granted before they begin their
> experiments is infinitely more interesting than any results to
> >  which their experiments lead.
> >  -- Norbert Wiener
> >
> >  https://www.cse.buffalo.edu/~knepley/
>

Re: [petsc-users] hypre-ILU vs hypre Euclid

2023-06-22 Thread Jed Brown 
It looks like Victor is working on hypre-ILU so it is active. PETSc used to 
have PILUT support, but it was so buggy/leaky that we removed the interface.

Alexander Lindsay  writes:

> Haha no I am not sure. There are a few other preconditioning options I will 
> explore before knocking on this door some more. 
>
>  On Jun 22, 2023, at 6:49 PM, Matthew Knepley  wrote:
>
>  On Thu, Jun 22, 2023 at 8:37 PM Alexander Lindsay 
>  wrote:
>
>  I know that PETSc has hooks for Euclid but I discovered today that it does 
> not support 64 bit indices, which many MOOSE
>  applications need. This would probably be more appropriate for a hypre 
> support forum (does anyone know if such a forum
>  exists other than opening GitHub issues?), but does anyone here know what 
> the difference between hypre-ILU and
>  hypre-Euclid are? From the docs it seems they are both supposed to be 
> parallel ILU solvers.
>
>  If hypre-ILU worked with 64 bit indices (I can probably check this sifting 
> through the sources), then I would probably add
>  hooks for it in PETSc (AFAICT those don't exist at present).
>
>  My understanding was that two different people were working on them. I do 
> not know if either is still actively supported. We
>  would of course like a binding to whatever is supported.
>
>  Are you sure you want to run ILU?
>
>THanks,
>
>   Matt
>  -- 
>  What most experimenters take for granted before they begin their experiments 
> is infinitely more interesting than any results to
>  which their experiments lead.
>  -- Norbert Wiener
>
>  https://www.cse.buffalo.edu/~knepley/

Re: [petsc-users] hypre-ILU vs hypre Euclid

2023-06-22 Thread Alexander Lindsay 
Haha no I am not sure. There are a few other preconditioning options I will explore before knocking on this door some more. On Jun 22, 2023, at 6:49 PM, Matthew Knepley  wrote:On Thu, Jun 22, 2023 at 8:37 PM Alexander Lindsay  wrote:I know that PETSc has hooks for Euclid but I discovered today that it does not support 64 bit indices, which many MOOSE applications need. This would probably be more appropriate for a hypre support forum (does anyone know if such a forum exists other than opening GitHub issues?), but does anyone here know what the difference between hypre-ILU and hypre-Euclid are? From the docs it seems they are both supposed to be parallel ILU solvers.If hypre-ILU worked with 64 bit indices (I can probably check this sifting through the sources), then I would probably add hooks for it in PETSc (AFAICT those don't exist at present).
My understanding was that two different people were working on them. I do not know if either is still actively supported. We would of course like a binding to whatever is supported.Are you sure you want to run ILU?  THanks,     Matt-- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.-- Norbert Wienerhttps://www.cse.buffalo.edu/~knepley/

Re: [petsc-users] hypre-ILU vs hypre Euclid

2023-06-22 Thread Matthew Knepley 
On Thu, Jun 22, 2023 at 8:37 PM Alexander Lindsay 
wrote:

> I know that PETSc has hooks for Euclid but I discovered today that it does
> not support 64 bit indices, which many MOOSE applications need. This would
> probably be more appropriate for a hypre support forum (does anyone know if
> such a forum exists other than opening GitHub issues?), but does anyone
> here know what the difference between hypre-ILU and hypre-Euclid are? From
> the docs it seems they are both supposed to be parallel ILU solvers.
>
> If hypre-ILU worked with 64 bit indices (I can probably check this sifting
> through the sources), then I would probably add hooks for it in PETSc
> (AFAICT those don't exist at present).
>

My understanding was that two different people were working on them. I do
not know if either is still actively supported. We would of course like a
binding to whatever is supported.

Are you sure you want to run ILU?

  THanks,

 Matt
-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener

https://www.cse.buffalo.edu/~knepley/

[petsc-users] hypre-ILU vs hypre Euclid

2023-06-22 Thread Alexander Lindsay 
I know that PETSc has hooks for Euclid but I discovered today that it does
not support 64 bit indices, which many MOOSE applications need. This would
probably be more appropriate for a hypre support forum (does anyone know if
such a forum exists other than opening GitHub issues?), but does anyone
here know what the difference between hypre-ILU and hypre-Euclid are? From
the docs it seems they are both supposed to be parallel ILU solvers.

If hypre-ILU worked with 64 bit indices (I can probably check this sifting
through the sources), then I would probably add hooks for it in PETSc
(AFAICT those don't exist at present).