5 ms·
> And they buy them in servers from OEMs who volume buy even bigger number of chip than any of them individually. No, they build their own servers. They defini
by coder543 6y ago
> And they buy them in servers from OEMs who volume buy even bigger number of chip than any of them individually.
No, they build their own servers. They definitely aren't racking Dell or HP servers.
Even Amazon's x86 based servers are extremely custom. That whole "Nitro" system isn't exactly a USB stick, and AWS is often using customized variants of Intel processors in their servers.
Let me repeat that: Intel literally makes custom processors specifically for AWS.
> If they were buying directly, Intel would have even more leverage to milk them, than if they stood behind the back of some beefy OEM.
That's not how bulk buying works... at all.
The scale of AWS is so much larger than you seem to think it is.
- monocasa 6y ago> Intel literally makes custom processors specifically for AWS Sort of. It's the same mask for everyone with the per customer special features either fused off, or just requiring a special MSR knock sequence.
- coder543 6y agoThat’s an implementation detail — an optimization on Intel’s side of things. It’s the same effect either way, and I can’t just ask Intel to customize their processors for me. AWS operates at a large enough scale that they do exactly that, and AWS already purchases directly from Intel, both of which are my main points. Baybal2 doesn’t seem to be aware of either of these facts based on their comment.
- baybal2 6y agoWell, the thing is, it doesn't seem to be really customer specific feature. An option to fuse an individual performance profile, or few more ME features doesn't amount to be something really custom.
- monocasa 6y agoIt's real custom silicon on the masks; it's not a binning thing. What I've mainly heard of is custom system integration glue for their boards.
- baybal2 6y agoEven if that it is, it must be something really small. It's hard to believe for me that they will put even a square millimetre of effectively dead silicon onto xeon for <10% of their sales.
- jeffbee 6y agoHere's an example of the kinds of speedups you can get from something really small: 3% fleet-wide (estimated) speedup from 0.006% marginal silicon area. https://www.samxi.org/papers/kanev_asplos2017.pdf https://www.samxi.org/papers/kanev_asplos2017.pdf
- monocasa 6y agoWhy? Particularly when the practicalities of layout, particularly around the uncore area, mean there's some free space anyway. Also, a square millimeter is around 30-40 million transistors on TSMC 7nm (about Pentium III level gate count), which is almost certainly way more than any of these features would need by orders of magnitude.
- coder543 6y agoYou're basically admitting that you don't know much about these parts of the datacenter industry, but you keep posting comments authoritatively like you do. Why can't you just ask questions instead of misleading people?
- fomine3 6y agoI'm just curious so ask you a question, what's the special feature on Intel processor for AWS? I had imagined something like special turbo-boost parameter to avoid performance hit.
- 6y ago
- baybal2 6y ago> No, they build their own servers. They definitely aren't racking Dell or HP servers. Yes, they buy from Quanta, Supermicro, Wistron etc, the ones who make servers for Dell, and HP. > Let me repeat that: Intel literally makes custom processors specifically for AWS. Some unlockable features, and perf profiles. Those weird SKUs popup on liquidation sale websites from time to time.
- jeffbee 6y agoQuanta is an ODM. You said OEM. Unless this is some kind of performance art, you really should stop commenting and switch to reading.
- coldtea 6y ago>Quanta is an ODM. You said OEM. Same difference, compared to building their own from parts. I don't think parent's point lies in the OEM vs ODM semantics detail.