Sign in with
Sign up | Sign in
Your question

IBM Dual Core Chips... Hyperthead?

Last response: in CPUs
Share
August 15, 2004 4:03:12 AM

IBM will implementing their Dual Core Processors.

I wondered if someone here could give an explination of this.

Does it compair to HyperThreading? or is this sometihng Completly new.

Looks like Apple might not be getting past 3.5 GHz for another year...

However it looks like there approach will be that of Dual Core Multi-Pocessors (quad processing, standard).
theroeticly... 12-14Ghz? in a year or so. hmm, I'll beleve it when i see it.



<b><A HREF="http://www.digitalgunfire.com" target="_new">DigitalGunfire-Industrial EBM</A></b>
ASUS P4S8X-P4 2.4B - 2x512M DDR333 - ATI 9500Pro - WD80G HD(8M) - SAMSUNG SV0844D 8G HD - LG 16X DVD - Yamaha F1 CDRW
August 15, 2004 4:43:43 AM

Who cares anymore. Apple is going downhill fast. Exspecially because some of the professional graphic studios are going the opteron/windows way, which was apple's edge.

Watch out for the <b><font color=red>bloody</font color=red></b> Fanboys!

AMD64 2800+ :: MSI Neo-Fis2r :: 512mb Kingmax ddr400 :: Sapphire 9800pro 128mb :: 10K WD Raptor
August 15, 2004 4:56:46 AM

i realize that... but im still curious about the dual core

<b><A HREF="http://www.digitalgunfire.com" target="_new">DigitalGunfire-Industrial EBM</A></b>
ASUS P4S8X-P4 2.4B - 2x512M DDR333 - ATI 9500Pro - WD80G HD(8M) - SAMSUNG SV0844D 8G HD - LG 16X DVD - Yamaha F1 CDRW
Related resources
August 15, 2004 8:24:25 AM

IBM chips have been dual-core for years now. From the Power4 and onwards I think. The recent Power5 is dual-core along with SMT (what Intel calls Hyperthreading). So it's pretty much 2 threads per core and 2 cores per processor.

Future revisions of the PPC architecture for Apple and others who use IBM's desktop chips will see a scaled down version of the Power5. Most likely with SMT enabled.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
August 15, 2004 2:13:54 PM

As far as I know, IBM's multithreading technique is quite useful. While Intel's Netburst, for instance, has a few execution units lying idle for a lot of clock cycles, IBM's power architecture has lots of execution units lying idle for a few clock cycles. Either way, it's wasted resources, which means multithreading can truly benefit either architecture.
August 17, 2004 3:23:47 AM

Quote:
2 threads per core and 2 cores per processor

so each CPU becomes a quad? interesting. since apple seems that they'l only be going with Dual Systems now, that'll make a dual/quad...let alone the possiblility of a quad/quad?


Looks like apple/IBM may have found a simple walk around for thise Hz barrier that chip makers are finding at the 90nm zone.

and although they will eventually plow threw the 90nm, they'll likely find just as many or more problems at 65nm... if that's the case, having gone 64Bit, dual Core, hyperthread(SMT), Dual CPU seems like a pretty safe road till then no?

<b><A HREF="http://www.digitalgunfire.com" target="_new">DigitalGunfire-Industrial EBM</A></b>
ASUS P4S8X-P4 2.4B - 2x512M DDR333 - ATI 9500Pro - WD80G HD(8M) - SAMSUNG SV0844D 8G HD - LG 16X DVD - Yamaha F1 CDRW
August 17, 2004 7:44:49 AM

There's really no problem at 90 nm (at least, Intel has gotten rid of them seemingly). Dothan and Montecito seem to be doing just fine scaling to 90 nm. Don't let a bad egg like Prescott ruin it for everyone.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
August 17, 2004 10:10:22 PM

true, that they are more or less past it(seemingly as you say). We have yet to truely find out...

eather way... there will be rad blocks none the less.



<b><A HREF="http://www.digitalgunfire.com" target="_new">DigitalGunfire-Industrial EBM</A></b>
ASUS P4S8X-P4 2.4B - 2x512M DDR333 - ATI 9500Pro - WD80G HD(8M) - SAMSUNG SV0844D 8G HD - LG 16X DVD - Yamaha F1 CDRW
August 17, 2004 11:09:06 PM

Current 90nm technology from Intel is not too bad at all.

Dothan had one or two problems which have been ironed out (Sonoma is ready to roll).

There's already silicon in testing for montecito, Intel's next 90nm dual-core, 24MB-cache, 4-way multithreaded, 1.7Billion transistor Itanium at speeds within 2-2.5Ghz and with 667Mhz FSB. In order to churn out these paper features (regardless of their real-life efficiency when dishing out performance), the 90nm process has got to be doing OK.

Prescott is the only bad egg here.

And right now it seems that the process doesn't like the (prescott) design. Or rather, the design is a badly chosen one; the process works, and actually probably gives us exactly what the prescott blueprint is about: crap.
August 18, 2004 1:33:26 AM

To me it just a others fail project from texas team.

i need to change useur name.
!