Dual Tualatin outperforming dual Foster?

[Shagga]

<< The Inquirer is reporting that the new P3 Tualatin processors running in dual configuration, are topping the performance of the much more expensive Foster XEON processors. This of course isn't good news for Intel, as they are trying to phase out their P3 line for the Pentium 4 and Pentium 4 XEON line.

Mike Magee goes on to say that because of this &quot;...the dream configuration will never see the light of day.&quot; Which I guess I don't understand. Tualatin processors are already being reported to be selling in some parts of the world and motherboard companies such as IWill have already shown off dual processor boards capable of running Tualatin processors. Unless of course Intel physically does something to the processor to make it not run in SMP. Maybe I am missing something here, but I'm not buying this story. >>

 

[Duvie]

I still have my suspicions from other reviews I have read and little dropped comments here and there that this little engine that could (tualitin) is going to be crippled or at least delayed long enough to allow the p4 to ramp up to higher numbers so the p3 doesn't beat it....It would be a black-eye in marketing if the so called older technology came out faster then the new technology.

You saw what the dual palominos did to the dual 1.7 xeons...

 

[Shagga]

I'm possibly thinking of SMP next time around. Although Tualatin to me is an option, I believe I will end up going P4 anyway. Has Tualatin got the SSE2 instructions?

If Tualatin is faster than the P4, won't Intel just cease manufacturing Tualatin sooner rather than later.

 

[Duvie]

I suspect (IMO) they may wait until p4 is up to 2.2ghz and then release the 1.13 and 1.2 tualitin and then they should be safe...by then northwood is out and I think 2 p3's at those lower speeds will not beat the p4's...SAVE FACE!!!At that point they won't care if they sacrifice the lower end of the p4 line.

as for sse2 instruction...I didn't think so, just an enhanced sse fomat...could be wrong though.

 

[MadRat]

Don't even compare the 2.2gHz Socket-478 P4 to the Taulitan. The bandwidth of the nexgen P4 will so outclass the Taulitan's DDR setup as to make it a mute point. We are talking about over a true 9gb/sec of data flow compared to ideally 2gb/sec.

No contest.

 

[Duvie]

That is exactly what I am saying...They will get to the new socket and all but abandon the old p4 to be ravaged by the overperforming tualitin...

I wonder how the earlt p4 adopters with the 1.3to1.7 are going to feel about this...they buy &quot;latest and greatest architecture chip&quot; expensive rdram and get beat by older chip with a slim down design...

Doesn't make much sense!!! maybe Intel should have held off of the p4 line until they switched to the .13micron northwoods...AMD no doubtedly forced their hand in the ghz race.

 

[DynaOne]

You can't look just at the straight bandwidth numbers. P4 needs 3X the bandwith just to perform equally. Only the bandwidth above 3X is incremental.
DynaOne

 

[Wingznut]

&quot;I wonder how the earlt p4 adopters with the 1.3to1.7 are going to feel about this...&quot;

I can only speak for myself, but I feel pretty good. Pretty tired, and not really looking forward to donning that bunny suit again tonight, but I feel pretty good overall. Thanks for your concern.


&quot;AMD no doubtedly forced their hand in the ghz race.&quot;

It takes a significant amount of time to design a cpu, and then manufacture enough. It's not like Intel saw AMD's 1.0ghz Athlon and said &quot;We need more mhz by the next quarter!&quot;, and then whipped one up. Or the more popular theory that they ceased R&amp;D mid-flight, and shipped the P4 as it was.

 

[Duvie]

I am not saying they whipped up a batch of new p4's...I am saying maybe they pushed them into the market a bit sooner then they might have first wished or planned...

And I think the factor of the p3 rampping issue at 1.13ghz might have been another factor that the p4's time had come.

Wingznut, being honest...does the good numbers on the p3 tualitin surprise you or since the p3 is relatively close to the athlon in same mhz performance this is basically no surprise since the lower mhz athlons 1.333 had already outperforming the p4 1.7 in many areas?

As a marketing stance do you think pushing this chip makes sense when it shows it will outperform the lower end line of the p4's right out of the gate?

 

[Sunner]

<< We are talking about over a true 9gb/sec of data flow compared to ideally 2gb/sec. >>


Do you know something we dont?
To achieve 9+ GB/sec you'd need 4 channels of 1GHz+ RDRAM.

Now, while RDRAM has a very low pincount making dual channel configs like the case of i850 rather cheap, 4 channel 1+ GHz RDRAM would most likely be ridiculously expensive.

This sort of bandwidth will probabaly be reserved for more high end boxes than the P4 will ever reside in.

 

[dowxp]

haha, take with a grain of salt. the inquirer is no longer the &quot;source&quot; for news... they arent that reliable. i heard one instance they used their power to raise amd's stock. something like that./

 

[Wingznut]

&quot;the inquirer is no longer the &quot;source&quot; for news&quot;

&quot;no longer&quot;??? When were they EVER?

But you have to admit, they are a good source for rumors to post on msg boards.

Speaking of, wasn't Dell supposed to start carrying AMD this month?

 

[MadRat]

<<Do you know something we dont? To achieve 9+ GB/sec you'd need 4 channels of 1GHz+ RDRAM.>>

The news lately has been RAMBUS is poising manufacturers to produce PC1066 and PC1200 modules. They also have devised new standards beyond todays 16-bit standard to 32-bit and 64-bit modules. With RAMBUS toting their 32-bit and 64-bit future, why would that be unrealistic? I'd think this 9gb/sec figure would actually be lowballing the future of RDRAM.

 

[dowxp]

<< &quot;the inquirer is no longer the &quot;source&quot; for news&quot;

&quot;no longer&quot;??? When were they EVER?

But you have to admit, they are a good source for rumors to post on msg boards.

Speaking of, wasn't Dell supposed to start carrying AMD this month? >>





hehe, they are good rumor posts. but they dont sound very credible. comp1 speaking, they have no ethos..

dell = AMD? i dunt think so..

 

[MadRat]

I'm sure Intel holding onto Dell is more of a prestige move than a business decision. Dell, with them leading the OEM scene, is an important moniker to possess. Wouldn't want anyone in the Fortune 100 to buy non-Intel, eh? Seems that these big spenders want to buy from the &quot;best&quot;, not necessarily the low bidder.

 

[christoph83]

Duvie in Tom's Review the P4 1.7ghz beats the overclocked 1.419 tualatin in a majority of the tests except in Sandra's benchmarks. A 1.5ghz tualatin and 1.5 P4 maybe a little different. And actually intel said they are already shipping out the 1.13 and 1.2ghz versions of the tualatin for desktops and notebooks to OEM's.When that hits the market,who knows,but fairly soon.I do agree with you Duvie on pushing the P4 a little too early.They knew the P3 would be running out of juice near the 1ghz mark and needed something else out to control the ghz race. I really think the weakness in the P4 will change soon with the release of the northwood. That bandwith is amazing.

 

[Wingznut]

&quot;I am not saying they whipped up a batch of new p4's...I am saying maybe they pushed them into the market a bit sooner then they might have first wished or planned...&quot;

&quot;I do agree with you Duvie on pushing the P4 a little too early.&quot;

But see, that's simply not possible.