DFI...Pls. send us new bios for Infinity 975X

[Dumo]

Just want to see how stable the board with 1.7Vcore and 2.6Vdimm w/out mods

Open multis still not work with stepping 5

[dinos22]

Just want to see how stable the board with 1.7Vcore and 2.6Vdimm w/out mods

Open multis still not work with stepping 5

you reckon we'll be able to have open multies for all CPUs

[Dumo]

400FSB possible, but with lower mem. strap...



Comparo to Asus P5W DH...

[maverik-sg1]

So if you lower the memory clocks you can clock CPU higher and over al is faster in superpi - be intersting to see how big the gap is in 32m

Thanks for the update Dumo

[sierra_bound]

Generally lower memory clocks will hurt you in 32M. Each of these boards (Bad Axe, P5W DH, and Infinity) will give different results at the same CPU speed. Also, the RAM you use can influence results.

All of the boards I mentioned use the same chipset. But the Infinity seems to have problems with 400+ FSB. I don't think a BIOS is going to fix that. I think it has more to do with the way the chipset was implemented on the board.

[Dumo]

Bios 0707 still won't unlock multis upward. It will be nice to run 420X12

[sierra_bound]

420FSB is fine if you have an X6800. But if you have an E6600, that only gets you up to 3780MHz. I would certainly want more than that.

I think kyosen said the Infinity had good memory performance. But he and some of the other Japanese moved on to other boards because of the FSB issue. The chipset should be capable of 450+, especially with the voltage options available.

[Nephilim]

Tony said on BE that the P5W DH may be soft-switching the strap from 1066 to 1333 once a certain FSB is reached (in the background) so the CPU thinks it's still running 1066 strap but the board is running 1333 strap.

This would account for the variances in people's testing results, showing the P5W ahead of boards like the DS3 or 975X/G at the same clocks one minute, and behind the next... just depends on the soft strap it's using. Since at 1333 strap the chipset will loosen latencies giving worse performance at the same fsb.

This is all heresay and conjecture until proof is found, but it's an interesting theory

[pcdoc1]

Why revive a 3 week old thread with old news…