Respect P4 2800@4400 Review !!!!

[Hardware]

http://www.muropaketti.com/artikkelit/cpu/nw2800/index.phtml


Kudos its a fresh and cold review!

 

[MSantiago]

If you look at the pics of the CPU you can tell it's an engineering sample from the "intel confidential" etched on the heat plate. hans007 has a P4 2.4GHz engineering sample and I've seen it do 2.93GHz at default voltage and stock cooling (it's multiplier unlocked too). His board doesn't have voltage tweaks, but with enough cooling and running @1.75v, I'm sure that thing can get to the mid to upper 3000s. I don't doubt that the P4 in that review is even more awesome. Engineering samples rock.

 

[WarCon]

Did you see the difference in Super PI between the B and C versions of the 2800? BTW, what is the C stepping? Is that just the normal 2800 with the improvements that Intel stuck in to surprise us?

 

[Czar]

thanks for the link Hardware

only if it were in English

 

[hans007]

what a freak. too bad they dont usually give unlocked units for review.


those n02 guys are crazy. c stepping has a 128 entry tlb, and the b step has 64 entries.

it helps with L2 cache, entry indexing .

 

[fluxquantum]

wow

 

[MadRat]

Hans007-

Anand said the C-stepping of the P4 has the EXACT same TLB entries as the other steppings.

See his message here and here. Where did he get his information? From Intel of course!

Kyle of [ H ]ardOCP was WRONG.

 

[Deeko]

hehe...I thought Hardware had mended his ways, no *confirmed* and a link..I was about to congratulate him...then I read czar's post lol...good work anyway hardware.

 

[Hardware]

this review is far more fun to read then the usual reviews

 

[Hardware]

Anand said the C-stepping of the P4 has the EXACT same TLB entries as the other steppings.


I guess Anand was wrong!

 

[hans007]

i hadnt read that new blurb, i guess there is a bug in the old chip that misreports the old b steppings chips number of tlb entries.


i was going off what aceshardware said.


besides that, anand can be wrong... he is not god. he wasnt wrong this time, but i dont take his word as truth and neither should all of you.

 

[MadRat]

If it hadn't been something that he aggressively pursued then I'd agree that Anand could have been wrong. But he did pursue his information directly from the source.

 

[Sunner]

There's an update at Ace's in case you haven't seen it.
Apparently there was a bug in the P4 core, not WCPUID, which reported the incorrect number of TLB entries.

Oh well, read it all over at Ace's