By: AM (myname4rwt.delete@this.jee-male.com), October 12, 2018 3:48 am
Room: Moderated Discussions
anon (anon.delete@this.anon.com) on October 11, 2018 3:49 pm wrote:
> AM (myname4rwt.delete@this.jee-male.com) on October 11, 2018 11:02 am wrote:
> > Foo_ (foo.delete@this.nomail.com) on October 11, 2018 10:22 am wrote:
> > > AM (myname4rwt.delete@this.jee-male.com) on October 11, 2018 9:15 am wrote:
> > > > And considering Cloudflare's well-known stance wrt ARM vs x86 (see e.g. https://www.datacenterknowledge.com/design/cloudflare-bets-arm-servers-it-expands-its-data-center-network)
> > > >
> > > >
> > > >
> > > > I suspect -- assuming Cloudflare folks are negotiation-smart -- they got these rejects not just for
> > > > free, but for a hefty subzero price, part of their deal being this blog post mentioning they're upgrading
> > > > their fleet to an off-roadmap/custom Intel CPU "offering considerable value in performance per watt".
> > >
> > > Why would Intel pay to advertise a CPU model that they don't want to sell?
> >
> > Not to advertise a "CPU model", but for Cloudflare staying with
> > Intel -- their ARM intentions are well-known (link above).
> >
> > > > With Centriq, ThunderX2, and finally Ampere eMAG servers available (https://insidehpc.com/2018/10/ampere-augments-arm-servers-new-developer-platform/),
> > > >
> > > >
> > > >
> > > > I guess the value of such PR project targeted at various business execs is well worth
> > > > selling rejects at negative price to prevent someone size of Cloudflare switching to
> > > > ARM.
> > >
> > > There are not many companies "size of Cloudflare" (in # of servers) around the world, I think.
> > > And I doubt those companies make CPU architecture decisions by reading blog posts.
> >
> > Nothing more than sloppy wording on my part -- companies who consider, or are forced to consider,
> > power efficiency and computer density are surely not limited to size of Cloudflare.
>
> They're all already looking at ARM, AMD, some even POWER. They all have a strong interest to make non-Intel look
> good on their workloads even if they don't switch, to negotiate with Intel. PR has nothing to do with it.
The thing is, regardless of whether Intel had PR-related considerations in this deal or not, there's no denial they exist -- it simply wouldn't be smart for them not to take any PR advantage out of a deal like that. Just like it wouldn't be smart for Cloudflare not to do a bit of digging and tell Intel upfront they are being offered rejects of commercial bins.
Besides, the whole assumption that Intel went out of its way and made a custom part for Cloudflare tuned to suit their workloads is way too fishy to consider it seriously (they may have claimed whatever they wanted to Cloudflare in private talks when they approached them to keep them as a customer, but claims are one thing, facts are another).
And, quite frankly, after their notorious 28-core 5 GHz demo at Computex last summer the assumption of Intel doing a custom part for Cloudflare which some posters here seem to take for granted, is downright dumb. A nice video in this regard https://www.youtube.com/watch?v=ozcEel1rNKM
Interestingly enough, my comment to their blog post which says literally
"There's likely more to this story than meets the eye from the post
[link to this thread on RWT]"
clearly did not pass the moderation.
I posted another one mainly to check whether it would be blocked as well, but it's early to decide yet on this one:
"Considering that both ThunderX2 servers and Ampere eMAG server-type development systems are available for purchase (https://developer.amperecomputing.com), why don't you guys get hold of them and publish some performance and power consumption benchmarks as you did before? "
> >
> > > There has to be a more plausible explanation. Such as: Intel offered an usually low price
> > > for those "reject" CPUs and Cloudflare decided it was a better deal than to go with ARM.
> >
> > As I said, that's assuming they were smart talking to Intel in the first place. What
> > were the real terms of the deal is anyone's guess -- a downright dumb negotiator
> > could have paid Intel for these rejects Intel's list price for production bins.
>
>
> AM (myname4rwt.delete@this.jee-male.com) on October 11, 2018 11:02 am wrote:
> > Foo_ (foo.delete@this.nomail.com) on October 11, 2018 10:22 am wrote:
> > > AM (myname4rwt.delete@this.jee-male.com) on October 11, 2018 9:15 am wrote:
> > > > And considering Cloudflare's well-known stance wrt ARM vs x86 (see e.g. https://www.datacenterknowledge.com/design/cloudflare-bets-arm-servers-it-expands-its-data-center-network)
> > > >
> > > >
> > > >
> > > > I suspect -- assuming Cloudflare folks are negotiation-smart -- they got these rejects not just for
> > > > free, but for a hefty subzero price, part of their deal being this blog post mentioning they're upgrading
> > > > their fleet to an off-roadmap/custom Intel CPU "offering considerable value in performance per watt".
> > >
> > > Why would Intel pay to advertise a CPU model that they don't want to sell?
> >
> > Not to advertise a "CPU model", but for Cloudflare staying with
> > Intel -- their ARM intentions are well-known (link above).
> >
> > > > With Centriq, ThunderX2, and finally Ampere eMAG servers available (https://insidehpc.com/2018/10/ampere-augments-arm-servers-new-developer-platform/),
> > > >
> > > >
> > > >
> > > > I guess the value of such PR project targeted at various business execs is well worth
> > > > selling rejects at negative price to prevent someone size of Cloudflare switching to
> > > > ARM.
> > >
> > > There are not many companies "size of Cloudflare" (in # of servers) around the world, I think.
> > > And I doubt those companies make CPU architecture decisions by reading blog posts.
> >
> > Nothing more than sloppy wording on my part -- companies who consider, or are forced to consider,
> > power efficiency and computer density are surely not limited to size of Cloudflare.
>
> They're all already looking at ARM, AMD, some even POWER. They all have a strong interest to make non-Intel look
> good on their workloads even if they don't switch, to negotiate with Intel. PR has nothing to do with it.
The thing is, regardless of whether Intel had PR-related considerations in this deal or not, there's no denial they exist -- it simply wouldn't be smart for them not to take any PR advantage out of a deal like that. Just like it wouldn't be smart for Cloudflare not to do a bit of digging and tell Intel upfront they are being offered rejects of commercial bins.
Besides, the whole assumption that Intel went out of its way and made a custom part for Cloudflare tuned to suit their workloads is way too fishy to consider it seriously (they may have claimed whatever they wanted to Cloudflare in private talks when they approached them to keep them as a customer, but claims are one thing, facts are another).
And, quite frankly, after their notorious 28-core 5 GHz demo at Computex last summer the assumption of Intel doing a custom part for Cloudflare which some posters here seem to take for granted, is downright dumb. A nice video in this regard https://www.youtube.com/watch?v=ozcEel1rNKM
Interestingly enough, my comment to their blog post which says literally
"There's likely more to this story than meets the eye from the post
[link to this thread on RWT]"
clearly did not pass the moderation.
I posted another one mainly to check whether it would be blocked as well, but it's early to decide yet on this one:
"Considering that both ThunderX2 servers and Ampere eMAG server-type development systems are available for purchase (https://developer.amperecomputing.com), why don't you guys get hold of them and publish some performance and power consumption benchmarks as you did before? "
> >
> > > There has to be a more plausible explanation. Such as: Intel offered an usually low price
> > > for those "reject" CPUs and Cloudflare decided it was a better deal than to go with ARM.
> >
> > As I said, that's assuming they were smart talking to Intel in the first place. What
> > were the real terms of the deal is anyone's guess -- a downright dumb negotiator
> > could have paid Intel for these rejects Intel's list price for production bins.
>
>