I think this is a matter of testing. For HPC applications performance is the
main goal. If RHEL6 (or SL6) is performing better than RHEL 5.6, go RHEL6.


Paulo Estrela
http://tabugado.com



On Sat, Jul 16, 2011 at 3:56 PM, Stephen John Smoogen <[email protected]>wrote:

> On Sat, Jul 16, 2011 at 11:36, Yasha Karant <[email protected]> wrote:
> > We are obtaining a new GPU cluster research compute engine, Nvidia CUDA 4
> > conforming.  Because of the way the funding agency evaluates proposals,
> we
> > effectively had to use an integrator who is well respected in the
> community
> > associated with the particular funding source.
> >
> > Speaking with the professional staff at the integrator, it appears that
> the
> > consensus is that RHEL 6 is not really ready for production, and that
> > production engines are being kept on RHEL 5.6 (CentOS 5.6, SL 5.6).  When
> I
> > enquired about SUSE Enterprise current, I received similar comments from
> the
> > same source.  It was noted that RHEL 6 was withdrawn for a while after
> > production release.  Does anyone reading this list have any observations
> on
> > the production stability of RHEL 6?
>
> I am sorry but the question is rather nebulous to answer. What is your
> value of production ready? What is your integrators? What criteria do
> you use for evaluation of that production readiness. It is an N
> parameter non-linear equation with different answers on what those
> criteria are.
>
> I know integrators who only are now trusting RHEL-4 for their projects
> now that it is reaching end of life. There are large sites running
> RHEL-6 to render movies and various scientific clustering items but
> they each have different reasons for their choices and each considers
> it "production stability."
>
> So unless you define what one means by production stability, the
> question is unanswerable.
>
> > This question is irrespective of new hardware support in RHEL 6 that may
> not
> > be operational (e.g., USB 3, nominally in RHEL 6.1).
> >
> > Yasha Karant
> >
>
>
>
> --
> Stephen J Smoogen.
> "The core skill of innovators is error recovery, not failure avoidance."
> Randy Nelson, President of Pixar University.
> "Let us be kind, one to another, for most of us are fighting a hard
> battle." -- Ian MacLaren
>

Reply via email to