Opteron, look at the UT2003 Benches

[St3v3]

By the looks of it, The Athlon 64 "can" go far if the memory controller is disabled. It all depends on how much heat the memory controller will generate. It's at times like this you wish rambus was the standard and not DDR-RAM as the rambus memory controller didnt emit so much heat.

[vortigern_red]

Agent: Seriously have you seen results for anything?

Trickle: You are getting confused betweeen AMD64 and Athlon64.
Also I hate to tell you but even Intel cannot do 5gig whenever they feel like it! The new prescot core is already whacking out the heat, far more then the current chips and its 0.9u and *only*3.4. When did Intel release the 3.06 chip Nov last year?

Thats almost a year and they have released only one higher clocked chip it would only take a 3400 or 3600 to take the performance crown undisputedly, but they are stuck just as AMD is stuck at 2.2 with current cores.

Unfortunatley for Intel they seem to be having problems with their next die shrink as well. Amd however seem to have got thieir act together and finally delivered on Athlon64.

But, and its a big but, it does not rely on a die shrinkl. Traditionally Intel seems to transition to a smaller process better than AMD. AMD needs to produce 0.9u chips in early 2004 to stay competitive (Intel needs them sooner!).

In short Intel is having problems with the move to 0.9u, AMD does not need it yet but may struggle when the time comes.

Time will tell, but its naive to think Intel can just kick out a 5gig chip if needs be!

BTW I think the core will scale well past 3gig! the athlon has had over a 400% speed increase in mhz alone never mind FSB and cache, instructions ect. Nobody develops a X86 CPU core with just a 50 or even 100% speed increase in mind like you said the P4 "netburst" core is expected to hit 5gig and started off at 1.4, also the P6 core started off at about 100Mhz(?) and scaled to 1.4gig!


Has anybody even run UT2003 in software mode?

Edit: st3v3: where the hell did you get that memory controller stuff from?

The Hammer allows the chipset to take over the memory if required but as far as I know no ones bothered to do so as the on die controller has such low latency?

As a side note if rambus was the dominant memory type no way would I be running 1gig of the stuff I'd be lucky to afford 128mb :-)

[Stoo]

AMD isn't stuck at 2.2Ghz with the current cores if the o/c results of the latest uber barton2500 are anything to go by, from one of the other threads they're regularly hitting 2.3-2.6GHz..

I'd imagine that there's still life in the old dog yet

[vortigern_red]

As I've said in other threads AMD may release a faster 0.13u barton.

But look at what people are overclocking the P4 to, does it mean Intel are confident to release a 4gig chip?

FYI I run an AthlonXP 2500@2.3 (in that other thread!)myself but there is a diffreence between what overclockers and chip manufactures are happy with. As the same applies to both AMD and Intel its fair to say that Intel have released only one chip above last novembers high of 3.06, and (I hate to say it) in terms of clock speed AMD have never released a chip higher than last years 2.25 gig athlonXP 2800.

Both companies have been pretty static for almost 12 months!!

[Agent]

Originally posted by vortigern_red
Agent: Seriously have you seen results for anything?

All i said was id like to see the banchmarks of UT2003 run in software mode on all 3 of the chips ?

[Nemeliza]

Impressive, i didnt realise the opertron would big enuff balls to get a look by intel. Only time will tell.

So which one of u wants to buy me one, and a board to go with it

[d_fresh1]

y does the 3200 barton get beat so easily

[Agent]

Originally posted by d_fresh1
y does the 3200 barton get beat so easily

Because IMO, i think its over rated as a 3200.

[Trickle]

Aye, most places agree a 28/2900+ name would be better. It might have been able to justify the name had Intel not released the 865/875 chipset.

Remember, whilst it has a little less cache, a 2.2 Ghz 1700+ is virtually the same speed as it.

[spikegifted]

Originally posted by Agent
Because IMO, i think its over rated as a 3200.

I think AMD rate their CPUs using a broad spectrum of applications relative to whatever benchmark system they've chosen. The increased bandwidth due to raising of the FSB allows AMD to justify its decision to rate the processor at 3200. In most situation, I doubt there'll be much with XPs with same clock speeds but slower FSB; but benefit is immediately obvious when memory bandwidth become bottleneck.

[Agent]

Originally posted by spikegifted
I think AMD rate their CPUs using a broad spectrum of applications relative to whatever benchmark system they've chosen. The increased bandwidth due to raising of the FSB allows AMD to justify its decision to rate the processor at 3200. In most situation, I doubt there'll be much with XPs with same clock speeds but slower FSB; but benefit is immediately obvious when memory bandwidth become bottleneck.

Dont get me wrong, i totaly agree with most of the rating system, with the exclusion of th 3200 and 3000 xp's.

[dgr]

http://www.sudhian.com/showdocs.cfm?aid=426&pid=1509

the 3dmark2k1se benchmark for the barton3200 seems a tad low... I benchmark that high with a 2800/rad9700pro!!!

-and I'm only running a single dimm at the moment :\

dgr

[dadofsam]

WOW! how much r these goin to be?