TerraRaptor

Sweet cat. What a shame I made it hate me)

dat msi power)

switch to 7sp1. Without service pack results are bad

yes, just point monitoring line with mouse

Good guide Michal!

Use different version of os. Сборка хп может корявая?

Установи .net 2.0 SP2

That was a great idea to make weekly prizes during contest - loads of strong results every Friday make it interesting to check AOOC page every day.

And I wasn't able to go below -105 with foil - checked with thermocouples attached to each stick's heatspreader.

Thanks all for answers. I have used liquid tape to insulate my new M7G. Will advise if it works better for me next week when I get ram pot (I was using aluminium foil all this time).

Those goodruns are making me mad. Well done Andrey!

Hi all please share some pics and tips about DIMM insulation. I'm able to run it 1-1.5 hours and then game is over - random reboots, errors, 55 post code. According to my experience it is better to avoid vaseline in empty slots and it is better to avoid Bostik/Blutaс for board insulation. Do you use fans for this area ot let it be frozen? How to avoid short circuits? Thanks in advance for sharing your experience.

Probably you need to install drivers for unknown device in device manager so it become "Intel device" - the driver is part of OC pack.

*snip*

Strong. Run aqua as well

The best I did with this setup is 1314 so far - just a slightly tweaked win7. 1350 is too much for the clocks/timings anyway. And it could be 1340, which is worse for cheat identification... Now you know, the tweak works as it should and gives some boost but benchmark itself can't handle this most of the times.

aaand:

I saw some results with bs like 458345395395035 there and they are in line with others)

Win7.

No it isn't. Will post a result with win7 within 20 minutes. Description of the tweak sent to der8auer as he is a part of hwbot team.

Benchmark compromised - I pass validation and, again, the thing I do is not against the rules. I will try to qualify. You see better options than deny xtu from qualification? What if I share tweak with other guys but not with you?)

Let others decide. I won't use it for qualifiers but the tweak is simple enough to be discovered by others. And it allows runs with deviations small enough (I.e. +10 points to an average 1320) to be accepted by most moderators. Me personally vote for 2 benchmarks (pifast and fsx) with equal point distribution (50%/50%).

http://hwbot.org/submission/2610576_ I beleive 7.9 is not possible? This is still 5ghz run. I can cold boot to windows and show you same results running xtu without need of any actions prior to running xtu.

I don't think it is legal for current version of benchmark anyway. Want me to post overall global wr to start discussion, lol?)

no, bugged benchmark. With this tweak i can get different results from run to run - 1310 to >2000. And the tweak fully complies with the rules.