[petsc-dev] strangness in Chebyshev estimate of eigenvalues

[Mark Adams]

My hash function is:

            PetscScalar v = ((PetscScalar)((PETSC_HASH_FACT*idx)%100) -
49.5)/50.0;
            ierr = VecSetValues(bb, 1, &idx, &v,
INSERT_VALUES);CHKERRQ(ierr);


include/petscctable.h:#define PETSC_HASH_FACT 79943


On Thu, Aug 27, 2015 at 7:53 AM, Mark Adams <mfadams at lbl.gov> wrote:

>
>
> On Wed, Aug 26, 2015 at 3:30 PM, Jed Brown <jed at jedbrown.org> wrote:
>
>> Mark Adams <mfadams at lbl.gov> writes:
>> > I really think that it does not matter.  Just as long as it is noisy.
>> IMO,
>> > but we have no data, and I'm not sure how we could study it.  What is
>> the
>> > chance that your generator will generate a low frequency vector on some
>> god
>> > knows what grid?  I can not imaging this happening ... and I'm not sure
>> how
>> > I could even generate data to 'prove' it.
>>
>> Your hokey hash function generates stripes of various sorts on grids
>> with leading dimension 51, 100, and other combinations.
>
>
> You know there are more prime number, 100% of them are larger.
>
> What do you want to do?
>
> If dran48 is one line then I can just copy that and move on.
>
>
>
>> For an
>> anisotropic operator aligned to the grid, that would be devastatingly
>> low energy.
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20150827/243feb4d/attachment.html>