P4EE @ 7.2GHZ

[arbytech]

If it actually boots and is even somewhat stable then that's something!

 

[TomKazansky]

he got it boot at 6.2 ghz....i'm not sure about 7.0 ghz

i say he can't get it stable even at 6.2, more like 6.0-ish

either way, the high clock speed will now make up for the flaw in the prescott core =P

 

[VirtualLarry]

Originally posted by: gobucks
isn't there like a hardware limit to how fast a given architecture can physically be?

Yes. They are called "critical paths" in the CPU, and are the absolute upper-bound in terms of the freqency that the CPU be clocked at, and still operate reliably according to its design. That was one of the reasons why the original Intel 1.13Ghz P3 had to be recalled. Intel managed to fabricate a chip that would clock faster than the critical paths in the chip's design would allow it to operate reliably. Hence they had to re-design it.

Originally posted by: gobucks
Anyways, I don't give a crap what he can boot into BIOS with, or even what he can boot windows with - if it isn't stable, then there is no point in bragging about it. It's about as useful as a car that can hit 500MPH, but you can't shift it out of park.

Exactly. It's like boasting about your "3000W" (P.M.P.O.) speaker set, that will operate at that level for about 0.00000001 seconds before it goes pop. In this case, it won't go "pop" unless it exceeds certain thermal or electrical limits with respect to the silicon, but there's no way in heck that it will be stable at that speed, given the design. You are correct, sir.

 

[VirtualLarry]

Originally posted by: infestedgh0st
he got it boot at 6.2 ghz....i'm not sure about 7.0 ghz
i say he can't get it stable even at 6.2, more like 6.0-ish
either way, the high clock speed will now make up for the flaw in the prescott core =P

That's about the extreme upper-end of what the re-designed Prescott was supposed to be able to hit (~6Ghz), power/cooling issues notwithstanding. There may be some opcodes that will not operate reliably at those speeds either.

If someone were really hardcore, they would create a custom microcode patch to replace those hardwired functions with critical-path circuit violations with microcoded implementations instead, thus slowing down those operations, but allowed the remaining ones to run at those insane maximum clock speeds. (Kind of like relaxing the timings on DRAM, in order to ramp the raw clock speeds. It's hard to say which would offer better performance.)

The other question is whether you have to disable the L2 cache to be able to clock that high, and what kind of real-world performance would you see with it disabled? Back in the P2 days, with even half-speed L2 cache, disabling it would often allow you to go past the limit of the stock 450Mhz chips, but real-world performance with no L2 was actually far lower.

 

[amol]

Originally posted by: ShadowBlade
too bad its intel

at THAT speed, it will DEFINITELY PWN AMD!