I.nfraR.ed

Change PCMark05 with wprime and all would be fine

What happened with the "Core 2 9000 series" stage?

So, after all...only der8auer got an increase?

Guys who glued back the ihs, did you left a small part not glued or you placed glue on the whole outer surface of the ihs? I can't explain it right, but look at the photos with removed ihs. The stock glue don't go all the way, but have one small "hole" for pressure equalization and gas escape. And like other said - did you apply correct pressure before the glue went stiff? Someone tried Arctic Ceramique? I don't think MX-2/3/4 are good for subzero, they use to freeze easily, while Ceramique is rated for -150.

I guess you tried it on the C5F with 1402 bios?

Here's the 7GHz validation. The link was removed quickly http://valid.canardpc.com/show_oc.php?id=2386564 The other one is for memory validation. 4 slots populated, 4x4GB: http://valid.canardpc.com/show_oc.php?id=2385230 If you think it's not appropriate sharing these, you can delete my post (on behalf of hicookie).

You should easily get to 6 - 6.5GHz range on air, I believe I can even give you my new batch FX - 10GHz candidate chip with 1.79V.

What new batch? It's (1204PGT) older than my second 8150 (1209 PGN) that I got from RMA and it's total crap on water. However not tested on LN2 yet and yes I know you can't make conclusions over different samples. But 600MHz over the previous record isn't something to believe easily, plus all his big validations are rejected. Someone should test this 1402 bios, I don't have the board. Refused to buy 9xx chipset board from Asus when the Crosshair IV didn't get full-working bios.

So changing your nickname from ksin to ksin_reliable makes the score valid?

Have the same card, gonna beat you sometime

Hm, noticed from a picture he's running the latest bios 1402, which was released to public 3 days ago. Could it be this bios is bugged? The difference with the previous record is so big and he's still going up. 10GHz soon, anyone? Windows 7 gadgets are happily running, while he's validating. Yeah, I know I sound like hater, but I still can't believe it. Wonder if he's increasing the HTref/multi in windows or directly boots with the dialed settings and doesn't touch anything while in windows. Voltage reported by cpuz is still 1.89V.

It's not instant validation anymore or had never been?

Yep, pi at that frequency is super slow. Could that frequency be replicated with lower HTref and higher multi or at least get close to it? Voltage seems low as well, but everything is possible. Pi result might be bad if other cores are clocked much lower and no affinity is set?

Wow, probably I can't do that with the DFI, even though I have some good Microns. Should ask you a couple of questions some time...

Yes, afaik Latency Boundary deals with tertiary timings only. It's a set of profiles/working values. I usually set it to the lowest which memory could take after I've done with the other timings, then try to tighten tertiary timings further from that point. I think Shamino said LB is there just to save us time. PS: But yeah... I might be wrong though, bought that board before 3 days

L212B228 #1700 3770K on M5G 5000MHz 2c/2t takes 1.4V to load XP and run 32M on water. 4700MHz 4c/8t 1.28 LinX + AVX stable

Ok, you win!

Aha, got it. But you still need to change the frequency, it's higher than stated

FX-57

This one and pi (1M & 32M) are fine, but I can't get pifast right

You should understand this is a problem with cpu-z and K7 platform. It's always showing real FSB = Rated FSB. I don't remember which version broke it, but I believe it was something like 1.52. And we can't validate with older versions... If you divide the cpu speed by the multiplier, you will see cpu-z Bus Speed reading is false. And it does it only for validations. When opened in Windows it's perfectly fine. But you can't force to have a screenshot when the rules for validations don't require it. Edit: Seems 1.55 or 1.56 was the last validating correctly. 1.55 is ok: http://hwbot.org/submission/1060468_i.nfrar.ed_cpu_frequency_athlon_xp_2000_thorton_2853.73_mhz 1.57 and up are broken: http://hwbot.org/submission/2185457_i.nfrar.ed_cpu_frequency_athlon_xp_m_2600_3188.46_mhz

Advanced button next to the command line.

Yep, I know we can't compare chips this way, but it's somehow indicative. If you're doing 5.2 1.4V on what you call air, then mine should be total crap, considering my cooling - 1x triple 140 rad (PA140.3) with YateLoons at 2300rpm and Supreme-HF block

Are you pre-testing on air with 2/2 and some specific voltage for all chips or just quickly finding the max 32M on air regardless voltage? Mine does 32M 5.2GHz @ 1.55+ V, which seems pretty bad, compared to what I read here. My M4G is broken and M5G is not yet available here. I'm crying badly...

Wow, you found it. Great CPU \m/