POLL: cpu, who do you think will be whooping @ss in 6 months ??

[Soulkeeper]

yup the cpu's in apple comps

 

[Duvie]

Actually Apple just tapped the use of the IBM's new processor for its upcoming platform....

BUt since I would sooner eat bugs then run a platform that only has less then 8 percent of the market and thus less software and almost impossible to actually test comparatively against a pc to see real advantages and I mean more then a few filters in photoshop....


INtel becuase I truly belive in 6 months...paper launches aside because I wouldn't put it past amd to paper launch the cpu ahead of prescott even though it likley wont hit shelves until 08/03...that INtel will be releasing a 3.2ghz plus maybe even near 3.4ghz p[rescott chip with 666fsb dual channel pc2700 supported chipset and 1mb l2 cahe...and then ofcourse AMD better have something better then 3400+ pr rated chip cause I think just even with INtel isn't going to cut it anymore, around here and in the marketplace.....to many promises...too much hype...not much delivery in the last year....

 

[Soulkeeper]

yup.......
few days back IBM announced that
could be interesting

 

[Pabster]

ROFLMAO. VIA? Come on guys.

 

[Deskstar]

Killer apps? Killer hardware? Which comes first? Does it matter?

On the business side, easier videoconferencing is badly needed. Right now we gather in a videoconference room set aside for this purpose to talk to all affiliates and subsidiaries; too expensive and too much trouble. It takes an IS tech to even set it up. So, that is one area where a Killer App would drive hardware purchases in the business world.

On the home/small office side, easier video capture should follow in the footsteps of Napster. By that I mean I can see huge volumes of ripping of digital images by home users, again driving hardware purchases. If kids and yound adults want to preserve their home movies and rip Hollywood movies, then they will need user friendly software and hardware that is quick and easy. Those are Killer products for the next 5 year cycle. DVD burners, encoders (probably hardware to get it done fast), file systems for the video clips. Then inclusion of these videos in PowerPoint presentations at school and work, replacing the current animations and "still life" PowerPoints that predominate in the business world.

Want to be on the cutting edge....become a video guru.

IMHO only.

 

[merlocka]

Only one "l" in Motorola.

Maybe if they had a pair they could make a decent CPU.

 

[merlocka]

Originally posted by: dullard
"Well, by all means, if you are all about speed, why don't you get one of those spiffy new Celeron 2ghz CPU's that overclcok to the 3ghz level...of course they can't even match a 1.6A P4....but speed is all that matters right?"
I defined what I call performance above: performance = frequency * IPC. The Celeron has a very low average calculations per clock. Thus even with a high frequency the performance isn't very good. You are ignoring the IPC part. Is that a difficult concept for you?

The Celeron's have the same IPC as the P4. It's the differences in cache which cripple their performance.

 

[dullard]

The Celeron's have the same IPC as the P4. It's the differences in cache which cripple their performance.

Twice in my posts (even in the post you linked) I used the definition:

Performance = frequency * the average number of calculations per clock

But since that is really awkward I used the shorter

Performance = frequency * IPC

If that is the wrong definiton of IPC, I'm sorry. But as it stands the Celeron does have a lower "average number of calculations per clock". It does less work with each clock tick because of its exceedingly low cache. It doesn't change my point at all - performance isn't frequency only. Frequency matter, but it also really matters how much is done per clock tick.

Is there a better acronym for "average number of calculations per clock"? I'll use that instead next time.