• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

New 800fsb Xeons.

Joined
Jun 16, 2004
Messages
50
So....

How do you think a dual rig with the new 2.8Ghz Xeons (800fsb, 1MB L2 cache, DDR2 support, E64MT) would compare to a dual Opteron 246 rig? They're approximately the same price...

According to the promo blurbs at Dell, the E64MT will allow these new Xeons to run 64-bit XP and other OS's:

...All Xeon processors available on the Dell PrecisionTM 470 and 670 workstations are Intel® EM64T-enabled and will support 64-bit extended operating systems, including Red Hat® Enterprise Linux WS v3 for Intel® EM64T at the time of Dell Precision Xeon platform launch and the 64-bit version of Microsoft XP when it becomes available in 2005.

So that takes away one upgrade factor. There's still the upgrade path to dual-core opterons coming down the pike...

-Dan
 
bored-sarcasm said:
it might, if the intel's would be 64 bit. i'd rather see dual fx vs dual EE
You cant run FX's or EE's in a dual config. Only Opterons and Xeons.
 
I read yesterday that Nocona based EMT64 Xeons are the same core as Prescott with the exception of the extra cache and it retained the 20 stage pipeline that Northwood has. Making it Intel's best performer by far. Combining everything that was good in Northwood, Gallatin and Prescott.
 
Sir-Fragalot said:
I read yesterday that Nocona based EMT64 Xeons are the same core as Prescott with the exception of the extra cache and it retained the 20 stage pipeline that Northwood has. Making it Intel's best performer by far. Combining everything that was good in Northwood, Gallatin and Prescott.

No shit? That might just make these a nice upgrade path then, since they didn't include that absurdly long 31 stage pipe in them. Do you have some links talking about this? :)
 
Lemme see if I can find the link. I actually found the link in another thread.
 
Sir-Fragalot said:
I read yesterday that Nocona based EMT64 Xeons are the same core as Prescott with the exception of the extra cache and it retained the 20 stage pipeline that Northwood has. Making it Intel's best performer by far. Combining everything that was good in Northwood, Gallatin and Prescott.

If this is true, then the Nocona's are the obvoiuse choice.
 
Well I am not sure now. I can't find the article I found yesterday. But I did find another website saying that it had the 31 stage pipeline of the Prescott. Which would suck but make sense if what Intel says about it being necessary to ramp up clock speeds.

Even so a Multi-Processor Nacona with EMT64, and larger cache sizes than Prescott and an 800MHz FSB, it would still outperform EVERYTHING in Intel's lineup.

Even in Uni-processor form a single Nacona would win out by having an 800MHz FSB like Northwood and Prescott, but would have the cache sizes at least as good as Prescott on the lower end and on the higher end would be like Gallatin. So even so it's a great performer and is cheaper than the Pentium 4 EE.

Personally I think a 3.2GHz Nacona and the Asus PC-DL Deluxe would be an excellent gaming machine. If they release a BIOS update for the PC-DL to support Nacona. As the PC-DL has the i875P chipset. So natively it already has 800MHz FSB support and DDR400 support.
 
Sir-Fragalot said:
Well I am not sure now. I can't find the article I found yesterday. But I did find another website saying that it had the 31 stage pipeline of the Prescott. Which would suck but make sense if what Intel says about it being necessary to ramp up clock speeds.

Even so a Multi-Processor Nacona with EMT64, and larger cache sizes than Prescott and an 800MHz FSB, it would still outperform EVERYTHING in Intel's lineup.

Even in Uni-processor form a single Nacona would win out by having an 800MHz FSB like Northwood and Prescott, but would have the cache sizes at least as good as Prescott on the lower end and on the higher end would be like Gallatin. So even so it's a great performer and is cheaper than the Pentium 4 EE.

Personally I think a 3.2GHz Nacona and the Asus PC-DL Deluxe would be an excellent gaming machine. If they release a BIOS update for the PC-DL to support Nacona. As the PC-DL has the i875P chipset. So natively it already has 800MHz FSB support and DDR400 support.
It has the 20 stage pipeline. Think about it. The Prescott tweaks made it about 33% or around 30% more efficient then the northwood core. Then intel lengthened the pipeline to 31 stages. Increasing the FSB wont make any groudbreaking improvements. While i dont know how much of an increased trace cahce will do for a CPU, Intel giving the Xeons a bigger cache should make the CPU around 30% more efficient. That wouldnt leave any room for an increased pipeline. Also logically, it would make sense. The Opteron beats the Xeons in some tests and Intel needs something to bring destruction to the Opteron lead. Also, having an 800 MHz isnt that groundbreaking when moving up from the 533 Mhz FSB. Thats because Intel used a shared bus. Thats basically the FSB ot the slowest CPU divided by the number of CPUs there are = the bandwith the CPU gets. So moving up from the 533 to 800 Mhz FSB gave the processor a bandwth of 3.2 GB/sec as opposed to 2.7 GB/sec. This gets worse in the XeonMP when each CPU gets 800 MB/sec of bandwith in a quad processor setup. Thats why i think a company should release a motherboard that has dual Northbridges. One for each xeon. That way, each xeon can take full advantage of its 800 MHz bus. Also, the last time i checked, the Asus PC-DL Deluxe only supported DDR333.
 
I believe nocona will continue the tradition of the previous xeons, a slower performer compared to hammer.

The strong of the hammer architecture is surely not in the AMD64 IA, and EM64T can't change the situation. Not talking about the 31-stages...

Hope to read soon a review on aceshardware on the matter. Am I wrong or there isn't a single nocona review out?








subliminal msg: trash IE and replace with firefox
 
xxriversxx said:
I believe nocona will continue the tradition of the previous xeons, a slower performer compared to hammer.

The strong of the hammer architecture is surely not in the AMD64 IA, and EM64T can't change the situation. Not talking about the 31-stages...

Hope to read soon a review on aceshardware on the matter. Am I wrong or there isn't a single nocona review out?
there is. THe INQ said that while it was ok, the 64 bit extensions werent perfact. Microsoft even admitted that Intel's 64 bit extensions werent as good as AMDs, but will continue to mature and eventually be better:http://www.theinquirer.net/?article=16879
 
Well all I have to say on the subject is that Intel's extensions may or may not be superior to AMD's. But AMD had thiers out first. So Windows XP 64bit was actually somewhat tested on the AMD platform first.

It will take time, driver patches and maybe even a newer version of the Xeon to make EMT64 work. But eventually it will.

As far as the 20 stage vs. 31 stage pipeline, I looked through the Intel Data sheet on the new Xeon and it didn't say. Just because a 20 stage pipeline makes sense doesn't mean that's what Intel did. If what they say about processor scaling is true anyway.
 
Well I hope the Nocona has a 20 stage pipeline. That would be great. I would definatley get one over Prescott is that's the deal. I'm contemplating that anyway.
 
Sir-Fragalot said:
Well all I have to say on the subject is that Intel's extensions may or may not be superior to AMD's. But AMD had thiers out first. So Windows XP 64bit was actually somewhat tested on the AMD platform first.

It will take time, driver patches and maybe even a newer version of the Xeon to make EMT64 work. But eventually it will.

As far as the 20 stage vs. 31 stage pipeline, I looked through the Intel Data sheet on the new Xeon and it didn't say. Just because a 20 stage pipeline makes sense doesn't mean that's what Intel did. If what they say about processor scaling is true anyway.
Overclockers.com states that the increased pipeline was a result of an increased heat load. They seem to be right.
 
The fact that they hae paper launched last week and nobody has none for a simple review tells a lot - everything that i have heard says they are the prescotts with more cache AND 31 stages. basically a prescott EE with AMD64 extentions. They probably suck, if they didnt we would see some leaks by now.
 
beyoku said:
The fact that they hae paper launched last week and nobody has none for a simple review tells a lot - everything that i have heard says they are the prescotts with more cache AND 31 stages. basically a prescott EE with AMD64 extentions. They probably suck, if they didnt we would see some leaks by now.

uhm, xeons are always p4 with more cache.
and wasn't a paper launch, the same day akiba pc hotline started selling them.
 
I found several web sites that had them for sale. They are also not that much more expensive than a Pentium 4. So really the cost increase is in the motherboards and the registered ram if the board used requires it.
 
beyoku said:
The fact that they hae paper launched last week and nobody has none for a simple review tells a lot - everything that i have heard says they are the prescotts with more cache AND 31 stages. basically a prescott EE with AMD64 extentions. They probably suck, if they didnt we would see some leaks by now.
How would you explain a 30% increase in performance? Thats hard to reproduce when you have the same exact core (supposidly) and the Prescott core is less efficient while the Nacona is more efficient.
 
Back
Top