Massman

It's possible that the readouts are incorrect? What is your room temperature?

Yes, it used to be worse. Like I said ... ASUS claimed that Lucid was better than native SLI/CF on their AMD board.

Done!

In the meantime, please throw me your dump. (hmm)

Yes. I'll post a dump of the GTX2 SPDs here in a couple of minutes (perhaps 1h).

P386? How is that possible?

That's how incompatible SPD would show

Yes.

If I can find identical ATI cards, I'll give this a spin. As far as I know, ASUS claimed better results with Hydra multi-GPU than a native SLI/CF solution.

... and the quest goes on! After a message from Chew* who mentioned he had the exact same issue on the H57 board, I had to re-test this behavior on a different branded mainboard: GIGABYTE H55M-UD2H. This is my first ever H55 mainboard ... nostalgia FTW! To make a long story short, I recreated the problem on the GBT board: 180MHz 2:10 made the screen go bezerk. However, this was after setting the same frequencies I used on the MSI board ... in my first couple of runs, I went over 185MHz 2:10 smoothly. So, something else than the memory is bottlenecking the system. For now, it looks like the QPI has something to do with it ... but it's definitly not linear. Lowering the QPI multiplier 1 setting does not magically solve this issue; the 180MHz limitation is still there. Conducting further research ...

Isn't Lucid multi-GPU performance > SLI multi-gpu performance?

Added to competition

Dumping an SPD can be done using the Taiphoon Burner app: http://www.softnology.biz/

Can the people who have problems submitting with unlocked cores post the URLs of the scores here, please? I will manually link them to the competition untill the bug is solved coding-wise.

None of those results are in the database; seems like they are just hard-deleted :-/ We have backups from our database, so it should be possible to find them back ... but it's going to take a few days (might not be this week). The best thing you can do is try to re-submit as much as possible.

More details on the 661 scores? Superpi? what frequency? what score?

http://www.passmark.com/products/pt.htm The people of Passmark have contacted us to check if we'd be interested in hosting their benchmark application. I've Been testing this PCMark-a-like benchmark this morning and I have to say it looks quite good. It's definitly not ready to be launched in the next couple of weeks, but with a bit of fine-tuning I think this could be a worthy PCMark alternative. Tests: - CPU - VGA 2D/3D - MEM - HDD - CDR (wtf!) Few remarks after testing: - Both CD-Rom as HDD tests are vulnarable to virtual drives - No DX11 3D test Good thing about this benchmark is that the storage benchmarks are not affecting the end result too much. With a 4GHz 1090T and no virtual drives I get around 2450, with 2.8GHz 1090T and virtual drives roughly the same. I think it should be possible to even disable these virtual drives (like FM did with Vantage), so I'll ask them. Things I like to see added: - dedicated audio/video encoding? - USB test - DX11 test - A lot more security - Direct submission to HWBOT database Let me know what you think about the benchmark

Can you give an example of a score that you submitted?

I had a peek in the database and couldn't find scores linked to your profile (apart from those showing in your profile). This problem is too complex for me, we'll have to wait for Richba5tard to return from vacation.

One thing I noticed is that this board sometimes randomly disables cores when unstable.

Still quite a big group that prefers to keep it the way it is. Arguments against it are more than valid, I think.

You can only go to some extend to make the competition more fair to less fortunate countries, but you need to keep in mind that this is an extreme overclocking competition where the final rounds will be featuring the latest and greatest high-end hardware. I know it's a pity that sometimes hardware to qualify isn't available or really expensive as well as the liquid nitrogen, but ... I don't think it's a good idea to leave pre-qualifiers open without extreme cooling required. It's nice to see one or two relative newcomers at competitions like this, but neither MSI nor Gigabyte (for GOOC) wants a final filled with people who have limited LN2 experience. Besides, I think MSI did a very good job switching from a must-play qualifier set-up (where overclockers were FORCED to play with a specific type of hardware) and open it up to a 5-stage-choose-your-battle kind of competition where you can qualify in 5 difference ways.

Impressive

Yeah, looks like all scores are gone? How did you change team?

Yep, yep, yep ...