Thank you Marek!
I am hoping that this is a brief turbulent period confronting a dilemma
that poses a real threat to economical science in the US. 8 years ago we
could not have known NVIDIA would increase in value by a factor of 40, nor
could we have expected the amazing resurgence of neural networks. And both
of these are the result of an amazing amount of hard work and forethought,
and NVIDIA's success is richly deserved.
But I've worked with CUDA since the beginning. And without a low-cost
desktop option, CUDA is no longer developer-friendly. While I love what
Volta offers, I have not enjoyed working over network connections with
datacenter-bound GPUs in the least so far. The development experience is
absolutely craptastic and parts of NVIDIA's tool chain do not even work
with CUDA 9 over a network. If this doesn't change, I sincerely believe it
will cost them dearly in the long run.
But NVIDIA just deprecated Fermi class GPUs from 2011 in the past year so
this is not a moment to panic. When I say pmemd.cuda will bitrot without
something changing here, I do not mean that tomorrow it will just stop
running. I mean that it will become impractical to support future NVIDIA
technology because the cost will be prohibitive going forward, especially
in the age of Trump and assuming he gets 2 terms. In the meantime, GTX
1080 Ti and Titan XP are fantastic hardware as well, and they will carry
pmemd.cuda for at least the next few years. What NVIDIA will miss out on
is beating Anton with one or two DGX-1Vs. That's something I think could
be achieved in a matter of months, but it requires rewriting the inner
loops from the ground up to parallelize everything across multiple GPUs
(especially the FFTs), switching to one process per server with thread per
GPU (Volta is the first GPU where how one invokes things matters), and
atomic synchronization between kernels instead of returning to the CPU.
I'm busy these days with my deep learning framework (I'm done fighting the
tide so to speak), but I am happy to advise anyone willing to pursue this
at this level of sophistication.
And besides, the future isn't Java or Python, it's manycore and one can
either choose to sail off into the sunset or embrace change.
Scott
On Mon, Oct 30, 2017 at 6:15 PM, Marek Maly <marek.maly.ujep.cz> wrote:
> Hi all,
>
> first of all thanks to Ross, Scott and the others for the information even
> if they are a bit incomprehensible to me.
>
> If NVIDIA has started to realize that a significant portion of its
> high-end GTX sales is not going to gamers but to scientists (as Gerald
> Monard noticed) they should be just happy and support the expansion of GTX
> to the world of scientific/technical applications, because even if GTX
> GPUs inhibit somewhat or maybe significantly sales of TESLA models, the
> NVIDIAs total financial gain, based on GTX expansion to this new
> application segment, probably more than compensates this.
>
> Their idea that if they start to complicate or even block the use of GTX
> for scientific/technical computations they will gain because Tesla market
> will be revived, is as naive as the idea that if from some store is
> removed pate people will start significantly more buy caviar instead and
> so the total finantial gain of that store will be significantly increased,
> which is naturally not true, because we all are significantly limited by
> our budgets.
>
> I really do not understand such a behavior especially if the another giant
> AMD is here ready to provide comparable alternative. I think that AMD will
> be really pleased with such unreasonable behavior of NVIDIA.
>
> Anyway, I would like to express my great thanks to Ross, Scott, Professor
> Case and all the other developers of this miracle called Amber that I have
> been using for almost 10 years. I strongly hope that this hangover of
> NVIDIA's managers will soon be gone and even if it did not, Amber will
> successfully continue to develop, including that excellent GPU support.
>
> Best wishes,
>
> Marek
>
>
>
>
>
> Dne Mon, 30 Oct 2017 19:57:47 +0100 Ross Walker <ross.rosswalker.co.uk>
> napsal/-a:
>
> > Dear All,
> >
> > In the spirit of open discussion I want to bring the AMBER community's
> > attention to a concern raised in two recent news articles:
> >
> > 1) "Nvidia halts distribution partners from selling GeForce graphics
> > cards to server, HPC sectors" -
> > http://www.digitimes.com/news/a20171027PD200.html
> >
> > 2) "Nvidia is cracking down on servers powered by Geforce graphics
> > cards" - https://www.pcgamesn.com/nvidia-geforce-server
> >
> > I know many of you have benefitted greatly over the years from the GPU
> > revolution that has transformed the field of Molecular Dynamics. A lot
> > of the work in this field was provided by people volunteering their time
> > and grew out of the idea that many of us could not have access to or
> > could not afford supercomputers for MD. The underlying drive was to
> > bring supercomputing performance to the 99% and thus greatly extend the
> > amount and quality of science each of us could do. For AMBER this meant
> > supporting all three models of NVIDIA graphics card, Geforce, Quadro and
> > Tesla in whatever format or combination, you the scientist and customer,
> > wanted.
> >
> > In my opinion key to AMBER's success was the idea that, for running MD
> > simulations, very few people in the academic field, and indeed many R&D
> > groups within companies, small or large, could afford the high end tesla
> > systems, whose price has been steadily going up substantially above
> > inflation with each new generation (for example the $149K DGX-1). The
> > understanding, both mine and that of the field in general, has
> > essentially always been that assuming one was willing to accept the
> > risks on reliability etc, use of Geforce cards should be perfectly
> > reasonable. We are not after all running US air traffic control, or some
> > other equally critical system. It is prudent use of limited R&D funds,
> > or in many cases tax payer money and we are the customers after all so
> > should be free to choose the hardware we buy. NVIDIA has fought a number
> > of us for many years on this front but mostly in a passive aggressive
> > stance with the occasional personal insult or threat. As highlighted in
> > the above artic
> > les with the latest AI bubble they have cemented a worrying monopoly
> > and are now getting substantially more aggressive, using this monopoly
> > to pressure suppliers to try to effectively ban the use of Geforce cards
> > for scientific compute and restrict what we can buy to Tesla cards, that
> > for the vast majority of us are simply out of our price range.
> >
> > In my opinion this a very worrying trend that could hurt us all and have
> > serious repercussions on all of our scientific productivities and the
> > field in general. If this is a concern to you too I would encourage each
> > of you to speak up. Contact people you know at NVIDIA and make your
> > concerns heard. I am concerned that if we as a community do not speak up
> > now we could see our field be completely priced out of the ability to
> > make use of GPUs for MD over the next year.
> >
> > All the best
> > Ross
> >
> >
> >
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber
>
>
> --
> Vytvořeno poštovní aplikací Opery: http://www.opera.com/mail/
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Oct 31 2017 - 09:00:04 PDT
