Nathan Brookwood
Nathan Brookwood
article clever computer computing faith history industry pc people performance says software somewhere thinking
There are 25 years of PC industry and 40 years of computer industry history that says there is never enough performance. My article in faith is that you will never have enough computing performance for things you want to do. Having said that, there are clever software people somewhere who are thinking of even cleverer things to do.
build capability close create devices graphics industry problem step
This is a big step forward. And if you're going to build in the capability for this I/O virtualization into devices like graphics cards, you're not going to want to do it twice. So if AMD can get the industry to close in around its standard, it could create a problem for Intel.
adequate companies consider deliver industry job motivation performance plans relevant seen standard sun themselves whether
But here the performance per watt is dramatically better than the industry standard and that to me is an adequate motivation to consider Sun. The cores themselves may not be very elegant, but what's relevant here is whether they'll do the job companies need them to do. From what I've seen Sun plans to deliver that.
deal easier problems space
Right now you have to scrunch things together. If you space out the components, it makes the thermal problems much easier to deal with.
deal easier problems space
Right now you have to scrunch things together, ... If you space out the components, it makes the thermal problems much easier to deal with.
address variety
PIC does a lot of different things. It does address a variety of needs.
intel last performance
I think they're getting back on track. The performance benchmarks that they demonstrated, these are all of the things Intel couldn't do last year.
chip decided dynamite forward larger performance reasons
Tualatin is a dynamite server chip. One of the reasons (Intel) decided not to go forward with the Foster-based Xeon chip was because Tualatin, with its larger cache, had better performance in server environments.
chip decided dynamite forward larger performance reasons
Tualatin is a dynamite server chip, ... One of the reasons (Intel) decided not to go forward with the Foster-based Xeon chip was because Tualatin, with its larger cache, had better performance in server environments.
chips desktop distance intel trying
What Intel is trying to do is put some distance between multiprocessor configurations, which will be Xeon-based, and Pentium 4 chips for the desktop and low-end workstation environments,
battery cooling desktop facing kinds life physics similar small tight
When they did the Pentium M, they were under tight constraints on power. Now, desktop and server are facing similar kinds of constraints. It's not so much battery life as it is noise, just the physics of cooling a really hot, small chip.
exploit multiple performance silicon single winning
Henceforth, those who need more performance must exploit parallelism -- multiple processors, often on a single silicon die. AMD is winning the dual-core server race.
advantage hearing lose market mean narrow next performance six today
What I'm hearing today is that AMD's performance advantage is going to narrow in the next six months, or AMD may even lose its performance advantage. That doesn't necessarily mean they're going to lose market share.
based december forecast four intel mid quarter shock slam weeks
Based on what Intel said in December in their mid quarter update, they should know what's going on, so it's a real shock when they can't forecast four weeks in advance. That should be a slam dunk.