81–90 of 198 posts
Earlier quoted context omitted.
But this time it's (almost) full windows, and it can run x86 application.
Clarification: Only 32 but apps. Not 64 bit apps, which the Windows world has almost fully moved to. And I’m sure they would run at reduced performance as well.
Barely insightful. The rumor of Intel buying Broadcomm, ... is most likely being perpetuated by Broadcomm itself
This article makes it seem that being a major player in mobile is the only way Intel could survive. Mobile is a decelerating market, while cloud and servers, where Intel has a huge lead is an expanding market and profitable. In 2017, Intel's revenue hit an all-time high, as it did the year before, and its stock is an all time high. This is not a company "fighting for its future."
The headline is definitely sensationalist but there are threats and missed opportunities on Intel's horizon and past. The lure of mobile is the upgrade cycle as those are much shorter. The same, short upgrade cycles were a boon for intel in the hey day of desktops/laptops. The bigger threat the article fails to focus on is ARM. Singularity for apps has always hinged on write-once-run-everywhere. Intel hasn't been abl…
From app developer perspective, processor architecture is low on the list of the problems you have to solve to run anywhere. User-facing apps aren't written in C or anything that close to metal anymore; I suspect that most of the Java, Kotlin and Swift code is completely architecture agnostic: add new capability to the compiler, change the build command and your ARM app becomes x86 app.
Earlier quoted context omitted.
But as cloud service provider grow stronger, they will begin to make their own chips, like TPU. That is a much bigger threat to Intel. Essentially, as Moore's law is effectively put into an end, people will seek to make their own customized chips, CPU will be marginalized as time goes by.
Except they wouldn't. It took Apple a decade to up their silicon to a level where it is now. For most of that time, their activity was a money sink for them. And even today, even with some of the best talent in IC design one can get in USA (PA Semi and its veteran designers from HP, Sun, Freescale, and whatever else was remaining of American semi industry), their chips are still half outsourced, 70% external IP if yo…
And what Chinese fabless was that?
Earlier quoted context omitted.
Yes. I must say I chuckled a bit when I saw the headline. Intel is still churning out 63% gross margins. On semiconductors. That is exceptional. Intel is not fighting for its life by any stretch of the imagination.
Intel is fine as long as ARM doesn’t compete in their core markets. Then it’s a sudden and violent change. I guess you don’t see ARM attaining parity with x86?
Earlier quoted context omitted.
But as cloud service provider grow stronger, they will begin to make their own chips, like TPU. That is a much bigger threat to Intel. Essentially, as Moore's law is effectively put into an end, people will seek to make their own customized chips, CPU will be marginalized as time goes by.
Except they wouldn't. It took Apple a decade to up their silicon to a level where it is now. For most of that time, their activity was a money sink for them. And even today, even with some of the best talent in IC design one can get in USA (PA Semi and its veteran designers from HP, Sun, Freescale, and whatever else was remaining of American semi industry), their chips are still half outsourced, 70% external IP if yo…
Earlier quoted context omitted.
The headline is definitely sensationalist but there are threats and missed opportunities on Intel's horizon and past. The lure of mobile is the upgrade cycle as those are much shorter. The same, short upgrade cycles were a boon for intel in the hey day of desktops/laptops. The bigger threat the article fails to focus on is ARM. Singularity for apps has always hinged on write-once-run-everywhere. Intel hasn't been abl…
Yeah but the upgrade cycle won't stay short for long, just like it hasn't for desktops. Give it 10 years and people will be keeping their phones until they break. Also anyone can make an ARM CPU so the margins are thin, whereas with x86 CPUs Intel will basically only ever have one competitor - AMD. They want to find another monopoly.
Earlier quoted context omitted.
> Yet, what took them at least 5 years, and extremely expensive to make reticule limit chips, was bested by a no-name Chinese fabless that blew TPUs out of the water on power/cost/performance ratio. I'm not familiar with what company or product you're referring to. What is it?
Novumind, they showed their engineering samples few months ago. And yes, they got quite close to their initial promise of 3 teraflops per watt. Cambricon - possibly a scam, though also claims performance on an equal level with moderately sound technical approach. They never released much info other than fancy CG videos.
NovuMind only does 3x3 convolutions. Literally a hard burnt ASIC. Also, the TPU's efficiency is close. (Btw both their "FLOPs/watt " numbers are useless, they don't account for utilization)
And Cambricon... well, I'll just leave it at that...
Earlier quoted context omitted.
Novumind, they showed their engineering samples few months ago. And yes, they got quite close to their initial promise of 3 teraflops per watt. Cambricon - possibly a scam, though also claims performance on an equal level with moderately sound technical approach. They never released much info other than fancy CG videos.
I can't find much information about Novumind (NovuTensor is their product) nor Cambricon. NovuTensor seems to be only doing inference , and only a specific type of inference (judging from their hardware, probably only images), which is a lot simpler to speed up than the training that TPUs do. I'd take the statement "Unknown Chinese company beating Google with fewer resources and less time" with a huge grain of salt.
I think this article misses the point: Intel has nothing to gain by entering the commodity smartphone processor market. The margins are slim, and therefore doesn't make good use of Intel's advantage in semiconductor fabrication. Serial performance is important in processors, and as long as Intel is king in this area, they will have a home in server racks and workstations. I keep hearing about how cheap arrays of ARM…
If margins in mobile are razor thin, then Intel can make bigger margins than anyone else by owning the fabs. And mobile's volume can justify owning fabs.