4 ms·
This is a great example of capitalism working properly: Intel has been 'phoning it in' for nearly a decade now, because their products had a significant lead o
by johnvanommen 8y ago
This is a great example of capitalism working properly:
Intel has been 'phoning it in' for nearly a decade now, because their products had a significant lead over AMD, particularly in performance-per-watt.
Now AMD has taken the lead, and it's going to take years for Intel to catch up. (Because CPU innovation moves at a glacial pace.)
The winners are the consumers.
- smaddox 8y agoAMD has taken the lead in performance per watt??? Do you have a source? Everything I've seen shows Intel is either matched or better. If you also take capitol cost into account, maybe AMD is in the lead; I don't really know. Either way, Intel is soon to release the first 10nm process product, which should provide a significant boost, as long as the architecture can take advantage of it.
- eip 8y ago"soon"
- zaarn 8y agoYou mean the 10nm that has been delayed for 6 years now? That has so bad yields Intel has internally just renamed 12nm to 10nm and use that? The 10nm that has 50% yields while being 12nm THAT 10nm? For Performance per Watt you can check this article: http://www.pcgameshardware.de/CPU-CPU-154106/Tests/CPU-Test-Benchmark-AMD-Intel-675663/ http://www.pcgameshardware.de/CPU-CPU-154106/Tests/CPU-Test-... The current Intel top dog against AMD in terms of perf per watt is the Core i3 8300.
- smaddox 8y agoNot sure where you heard that Intel renamed 12nm to 10nm, but you're probably thinking of another fab. Intel's 10nm cell density is comparable to Global Foundry's 7nm [1]. And yes, Intel's 10nm process is very late; that doesn't mean it's not coming. As far as yields go, they're constantly improving, and they depend on the processor, so your claim that the yield is 50% is either ignorant or intentionally misleading. I checked out your link, but all I saw was price per performance numbers, not performance per watt. Price per performance is an important metric for gamers, but not so much for servers, which is Intel's bread and butter. In the server space, performance per watt is what matters. If you have some evidence that AMD has taken the lead in performance per watt, I would love to see it. [1] https://www.semiwiki.com/forum/content/7191-iedm-2017-intel-versus-globalfoundries-leading-edge.html https://www.semiwiki.com/forum/content/7191-iedm-2017-intel-...
- zaarn 8y ago>Intel's 10nm cell density is comparable to Global Foundry's 7nm [1]. And yes, Intel's 10nm process is very late; that doesn't mean it's not coming. That was Q4/17. Latest information: https://www.semiaccurate.com/2018/08/02/intel-guts-10nm-to-get-it-out-the-door/ https://www.semiaccurate.com/2018/08/02/intel-guts-10nm-to-g... The 10nm parts due for Q4/19 (!) are 12nm more than 10nm. The report also mentions the real 10nm yields, 8-10%, targets at 50-60%. I admit that I'm not sure what the current yield on their 12nm is but I would guess it would have to be the target.
- smaddox 8y agoThat article is completely devoid of any technical information supporting the claim that Intel nerfed the 10nm scaling.
- monocasa 8y agoEh, I don't think it's capitalism per se, but more a direct consequence of the state of the art of the technology. Moore's law is coming to an end, Intel institutionally has always relied on their process advantages to stay ahead. But now it's only expected that others are catching up as we see that moore's law is really an s curve. Intel will get another bump with 10nm. They're taking on EUV 'early' where no one else is (which is what's taking so long) and it'll probably mean that they have a head start there once they figure it out, but that'll be one of the last bumps for them. BEUV is sort of intrinsically flawed.