[HPGMG Forum] Solver options used for FV results?

Jeff Hammond jeff.science at gmail.com
Tue Aug 19 14:20:06 UTC 2014


On Mon, Aug 18, 2014 at 10:28 PM, Jed Brown <jed at jedbrown.org> wrote:
> Jeff Hammond <jeff.science at gmail.com> writes:
>
>> I wasn't aware that K supported GCC but in any case, I'd imagine it's
>> dramatically worse than the Fujitsu C compiler. Do you need C99?
>
> The finite-element code currently uses for loop initializers and basic
> VLA pointers.  Neither is essential, but I'm not sure if it's worth
> pandering to a vendor that hasn't updated their C compiler in 15 years.

It is not worth it - I am sure of this.  C99 is amazing and it blows
my mind that people are still wasting their time trying to avoid it...
<cough> <cough> Argonne <cough> :-D

> Note that K's compilers claim to support C99:
>
>   http://www.fujitsu.com/downloads/MAG/vol48-3/paper10.pdf

It's hard to imagine how this can be said if something as simple as
loop initializers is unsupported.

> The finite-volume code does not need C99 flags.

That's too bad.

Jeff

-- 
Jeff Hammond
jeff.science at gmail.com
http://jeffhammond.github.io/


More information about the HPGMG-Forum mailing list