[Beowulf] Frontier Announcement

Tony Brian Albers tba at kb.dk
Thu May 9 07:51:26 PDT 2019 

Please stop, both of you.

/tony


On Thu, 2019-05-09 at 09:31 -0500, Richard Walsh wrote:
> 
> Jeffrey Layton wrote:
> 
> >  I've been watching for OpenMP-Nazis and it looks like I found one.
> 
> This is a joke, just like your apology.  You know nothing about me
> and to 
> suggest (without any evidence) that I threatened your family is a
> measure
> of limited depth of your standard of proof in general I would say ...
> if I do not
> like it ... it must not be true.
> 
> Peter K. essentially repeated my thoughts ... I guess he must be a
> Nazi also.
> 
> rbw
> 
> 
> On Thu, May 9, 2019 at 7:49 AM Jeffrey Layton <laytonjb at gmail.com>
> wrote:
> > Very serious. Someone does need help - and culture needs changing
> > (i.e. OpenMP). But apologies for posting this here. I've been
> > watching for OpenMP-Nazis and it looks like I found one.
> > 
> > Again - apologies for posting about this to the list.
> > 
> > An now back to your regularly scheduled discussion where Carol
> > Merrill is standing.
> > 
> > Jeff
> > 
> > 
> > On Thu, May 9, 2019 at 2:16 AM John Hearns via Beowulf <beowulf at beo
> > wulf.org> wrote:
> > > Seriously? Wha.. what? Someone needs to get help.
> > > And it wasn't me. I am a member of the People's Front of Julia.
> > > 
> > > (contrived Python reference intentional)
> > > 
> > > On Wed, 8 May 2019 at 22:57, Jeffrey Layton <laytonjb at gmail.com>
> > > wrote:
> > > > I wrote some OpenACC articles for HPC Admin Magazine. A number
> > > > of pro-OpenMP people attacked me on twitter (you know, OpenACC
> > > > sucks, OpenMP is great). I received a private email threatening
> > > > to kill me and my family if I didn't stop writing about
> > > > OpenACC. Given your pro-OpenMP, anti-OpenACC stance, using the
> > > > same tone as the threatening email, I wondered if that email
> > > > came from you.
> > > > 
> > > > 
> > > > 
> > > > On Wed, May 8, 2019, 17:48 Richard Walsh <rbwcnslt at gmail.com>
> > > > wrote:
> > > > > Huh ... ??  Weird, scary ... 
> > > > > 
> > > > > Just MHO. Dropping off this thread now ... 
> > > > > 
> > > > > rbw
> > > > > 
> > > > > Sent from my iPhone
> > > > > 
> > > > > On May 8, 2019, at 4:29 PM, Jeffrey Layton <laytonjb at gmail.co
> > > > > m> wrote:
> > > > > 
> > > > > > I was just pointing out that gcc has Open ACC capability on
> > > > > > AMD GPUs.
> > > > > > 
> > > > > > I didn't realize you part of the OpenMP Nazis. Were you the
> > > > > > one that threatened me and my family because I wrote about
> > > > > > OpenACC?
> > > > > > 
> > > > > > 
> > > > > > 
> > > > > > On Wed, May 8, 2019, 15:48 Richard Walsh <rbwcnslt at gmail.co
> > > > > > m> wrote:
> > > > > > > Jeffry/All,
> > > > > > > 
> > > > > > > Yes ... but given the choice of using OpenACC or OpenMP
> > > > > > > (if you are not going to write CUDA-HIP code for that
> > > > > > > extra 10% of performance) which captures most (all?) of
> > > > > > > the features of OpenACC, is a standard likely to outlive
> > > > > > > OpenACC, and should run on any vendor’s  accelerators,
> > > > > > > including whatever Intel comes up with ... why would you
> > > > > > > write in OpenACC ... ??  
> > > > > > > 
> > > > > > > GNU supports OpenMP too ... in my view, PVM is to MPI as
> > > > > > > OpenACC is to OpenMP 4.5-5.0 ... 
> > > > > > > 
> > > > > > > Cheers!
> > > > > > > 
> > > > > > > rbw 
> > > > > > > 
> > > > > > > Sent from my iPhone
> > > > > > > 
> > > > > > > On May 8, 2019, at 2:36 PM, Jeffrey Layton <laytonjb at gmai
> > > > > > > l.com> wrote:
> > > > > > > 
> > > > > > > > Don't forget that gcc supports both NV and AMD GPUs
> > > > > > > > with OpenACC. That's one of the lead compilers listed
> > > > > > > > on the Frontier specs.
> > > > > > > > 
> > > > > > > > Jeff
> > > > > > > > 
> > > > > > > > 
> > > > > > > > On Wed, May 8, 2019 at 3:29 PM Richard Walsh <rbwcnslt@
> > > > > > > > gmail.com> wrote:
> > > > > > > > > All,
> > > > > > > > > 
> > > > > > > > > Cray has deprecated support for in OpenACC in light
> > > > > > > > > of the OpenMP 4.5 and 5.0 standards, and their target
> > > > > > > > > and data directives. NVIDIA’s PGI Compiler group will
> > > > > > > > > keep OpenACC going for a while, but on AMD devices
> > > > > > > > > ... maybe not.  That Cray will support only OpenMP on
> > > > > > > > > Frontier seems to be a logical certainty.
> > > > > > > > > 
> > > > > > > > > So if you or yours want to run at speed on Frontier
> > > > > > > > > you should bone up on ROCm, HIP and OpenMP 4.5-5.0
> > > > > > > > > ... 
> > > > > > > > > 
> > > > > > > > > :-)
> > > > > > > > > 
> > > > > > > > > Cheers!
> > > > > > > > > 
> > > > > > > > > rbw
> > > > > > > > > 
> > > > > > > > > Sent from my iPhone
> > > > > > > > > 
> > > > > > > > > > On May 8, 2019, at 12:47 PM, Jörg Saßmannshausen <s
> > > > > > > > > assy-work at sassy.formativ.net> wrote:
> > > > > > > > > > 
> > > > > > > > > > Dear all,
> > > > > > > > > > 
> > > > > > > > > > I think the answer to the question lies here:
> > > > > > > > > > 
> > > > > > > > > > https://en.wikipedia.org/wiki/OpenACC
> > > > > > > > > > 
> > > > > > > > > > As I follow these things rather loosely, my
> > > > > > > > > understanding was that OpenACC 
> > > > > > > > > > should run on both nVidia and other GPUs. So maybe
> > > > > > > > > that is the reason why it 
> > > > > > > > > > is a 'pure' AMD cluster where both GPUs and CPUs
> > > > > > > > > are from the same supplier?
> > > > > > > > > > IF all of that is working out and if it is really
> > > > > > > > > true that you can compile 
> > > > > > > > > > and run OpenACC code on both types of GPUs, it
> > > > > > > > > would a be big win for AMD.
> > > > > > > > > > 
> > > > > > > > > > Time will tell!
> > > > > > > > > > 
> > > > > > > > > > All the best from my TARDIS!
> > > > > > > > > > 
> > > > > > > > > > Jörg
> > > > > > > > > > 
> > > > > > > > > > Am Dienstag, 7. Mai 2019, 16:59:48 BST schrieben
> > > > > > > > > Sie:
> > > > > > > > > >>>  I think it is interesting that they are using
> > > > > > > > > AMD for
> > > > > > > > > >>> 
> > > > > > > > > >>> both the CPUs and GPUs
> > > > > > > > > >> 
> > > > > > > > > >> I agree. That means a LOT of codes will have to be
> > > > > > > > > ported from CUDA to
> > > > > > > > > >> whatever AMD uses. I know AMD announced their HIP
> > > > > > > > > interface to convert
> > > > > > > > > >> CUDA code into something that will run on AMD
> > > > > > > > > processors, but I don't
> > > > > > > > > >> know how well that works in theory. Frankly, I
> > > > > > > > > haven't heard anything
> > > > > > > > > >> about it since it was announced at SC a few years
> > > > > > > > > ago.
> > > > > > > > > >> 
> > > > > > > > > >> I would not be surprised if AMD pursued this bid
> > > > > > > > > quite agressively,
> > > > > > > > > >> possibly at a significant loss, for the
> > > > > > > > > opportunity to prove their GPUs
> > > > > > > > > >> can compete with NVIDIA and demonstrate that codes
> > > > > > > > > can be successfully
> > > > > > > > > >> converted from CUDA to something AMD GPUs can use
> > > > > > > > > to demonstrate GPU
> > > > > > > > > >> users don't need to be locked in to a single
> > > > > > > > > vendor. If so, this could
> > > > > > > > > >> be a costly gamble for the DOE and AMD, but if it
> > > > > > > > > pays off, I imagine it
> > > > > > > > > >> could change AMD's fortunes in HPC.
> > > > > > > > > >> 
> > > > > > > > > >>  "Win on Sunday, sell on Monday" doesn't apply
> > > > > > > > > just to cars.
> > > > > > > > > >> 
> > > > > > > > > >> Prentice
> > > > > > > > > >> 
> > > > > > > > > >>> On 5/7/19 4:43 PM, Jörg Saßmannshausen wrote:
> > > > > > > > > >>> Hi Prentice,
> > > > > > > > > >>> 
> > > > > > > > > >>> that looks interesting and I hope it means I will
> > > > > > > > > finally get the neutron
> > > > > > > > > >>> structure which was measured last year there! :-)
> > > > > > > > > >>> 
> > > > > > > > > >>> On a more serious note: I think it is interesting
> > > > > > > > > that they are using AMD
> > > > > > > > > >>> for both the CPUs and GPUs. It sounds at least
> > > > > > > > > very fast of what they
> > > > > > > > > >>> want to build, lets hope their design will work
> > > > > > > > > as planned as well.
> > > > > > > > > >>> 
> > > > > > > > > >>> All the best from London
> > > > > > > > > >>> 
> > > > > > > > > >>> Jörg
> > > > > > > > > >>> 
> > > > > > > > > >>> Am Dienstag, 7. Mai 2019, 10:32:42 BST schrieb
> > > > > > > > > Prentice Bisbal via 
> > > > > > > > > > Beowulf:
> > > > > > > > > >>>> ORNL's Frontier System has been announced:
> > > > > > > > > >>>> 
> > > > > > > > > >>>> https://www.hpcwire.com/2019/05/07/cray-amd-exas
> > > > > > > > > cale-frontier-at-oak-ridg
> > > > > > > > > >>>> e/
> > > > > > > > > > 
> > > > > > > > > > _______________________________________________
> > > > > > > > > > Beowulf mailing list, Beowulf at beowulf.org sponsored
> > > > > > > > > by Penguin Computing
> > > > > > > > > > To change your subscription (digest mode or
> > > > > > > > > unsubscribe) visit https://beowulf.org/cgi-bin/mailma
> > > > > > > > > n/listinfo/beowulf
> > > > > > > > > _______________________________________________
> > > > > > > > > Beowulf mailing list, Beowulf at beowulf.org sponsored
> > > > > > > > > by Penguin Computing
> > > > > > > > > To change your subscription (digest mode or
> > > > > > > > > unsubscribe) visit https://beowulf.org/cgi-bin/mailma
> > > > > > > > > n/listinfo/beowulf
> > > > > > > > > 
> > > > 
> > > > _______________________________________________
> > > > Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin
> > > > Computing
> > > > To change your subscription (digest mode or unsubscribe) visit 
> > > > https://beowulf.org/cgi-bin/mailman/listinfo/beowulf
> > > > 
> > > 
> > > _______________________________________________
> > > Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin
> > > Computing
> > > To change your subscription (digest mode or unsubscribe) visit ht
> > > tps://beowulf.org/cgi-bin/mailman/listinfo/beowulf
> > > 
> > 
> > _______________________________________________
> > Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin
> > Computing
> > To change your subscription (digest mode or unsubscribe) visit http
> > s://beowulf.org/cgi-bin/mailman/listinfo/beowulf
> > 
> 
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin
> Computing
> To change your subscription (digest mode or unsubscribe) visit https:
> //beowulf.org/cgi-bin/mailman/listinfo/beowulf
-- 
Tony Albers - Systems Architect - IT Development
Royal Danish Library, Victor Albecks Vej 1, 8000 Aarhus C, Denmark
Tel: +45 2566 2383 - CVR/SE: 2898 8842 - EAN: 5798000792142

More information about the Beowulf mailing list