3 ms·

Shrinking feature size alone has had diminishing returns on efficiency for the last couple generations, and I haven't heard any indications that sub 22 nm would

by brigade 13y ago

Shrinking feature size alone has had diminishing returns on efficiency for the last couple generations, and I haven't heard any indications that sub 22 nm would reverse that. Obviously it still increases your transistor budget, which is admittedly important for server CPUs since caches are generally the single largest use of transistors. But efficiency-wise it's FinFET, not feature size, that has given Intel the biggest advantage as of late. And indeed the ARM foundries might not have volume shipping FinFET SoCs until 2015. I have no idea why you're comparing TI/Qualcomm to Intel when discussing foundry R&D - it's TSMC, GloFo, and Samsung that are relevant there. Also, Tom's hardware has its years off - Intel isn't shipping 14nm CPUs until next year, with 20nm ARM SoCs expected around the same timeframe. Then 10nm isn't expected until 2016 at the earliest, again with 14-16nm ARM SoCs in about the same timeframe. So yes, Intel is about a generation ahead of its competition, and even more factoring in FinFET. But that's not the end of the world - IBM, AMD, and Oracle still make server CPUs despite this. Not as successfully as Intel obviously, but enough that there could be room for one or two makers of ARM microservers to enter.

akiselev 13y ago
I forgot about FinFET and just extrapolated from the i-series jump but the only important differences I've seen in processors in the last 3-4 years has been power efficiency, cache, and bus speed. Clock speeds (especially with TurboBoost thrown in) became more and more erratic as a metric for my use cases so at this point it's all about Intel's microarchitecture and process (in my perspective). They've also been one of the most advanced firms materials wise so as we start to get smaller and smaller, I think Intel will begin to use new materials that will increase the gains on FinFET and process. IIRC in 2012 they were two "generations" ahead on high dielectric materials and their silicon straining process for mass production. I just picked two companies off the cuff that are relatively well known and that I think could make an impact in the server market with ARM. I think comparing ARM mobile SoCs to Intel's x86 or AMDs AMD64 is disingenuous and since all TSMC and Global Foundries can do is play catchup (hence me touting Intel's process advantage), the brunt is left on the microarchitecture designers and the integrators to really make a server that can beat Intel's Xeon. I'm sure TSMC/Global Foundries are deeply involved in the design aspect but I think other companies will make or break ARM in the server market. As for the dates, that's a shame to hear :(. However, I just looked it up and TSMC finished their 20nm design this year so I really doubt 10nm chips will be in a server-ready state in 2016. I think that the next few years are on ARM's turf but if Intel can hit the cost sweet spot that ARM is at (or even just in the ballpark) with Intel's process and performance, it might be the end of ARM as a non-mobile/embedded contender.

by brigade 13y ago

akiselev 13y ago

I forgot about FinFET and just extrapolated from the i-series jump but the only important differences I've seen in processors in the last 3-4 years has been power efficiency, cache, and bus speed. Clock speeds (especially with TurboBoost thrown in) became more and more erratic as a metric for my use cases so at this point it's all about Intel's microarchitecture and process (in my perspective). They've also been one of the most advanced firms materials wise so as we start to get smaller and smaller, I think Intel will begin to use new materials that will increase the gains on FinFET and process. IIRC in 2012 they were two "generations" ahead on high dielectric materials and their silicon straining process for mass production. I just picked two companies off the cuff that are relatively well known and that I think could make an impact in the server market with ARM. I think comparing ARM mobile SoCs to Intel's x86 or AMDs AMD64 is disingenuous and since all TSMC and Global Foundries can do is play catchup (hence me touting Intel's process advantage), the brunt is left on the microarchitecture designers and the integrators to really make a server that can beat Intel's Xeon. I'm sure TSMC/Global Foundries are deeply involved in the design aspect but I think other companies will make or break ARM in the server market. As for the dates, that's a shame to hear :(. However, I just looked it up and TSMC finished their 20nm design this year so I really doubt 10nm chips will be in a server-ready state in 2016. I think that the next few years are on ARM's turf but if Intel can hit the cost sweet spot that ARM is at (or even just in the ballpark) with Intel's process and performance, it might be the end of ARM as a non-mobile/embedded contender.