Massman

- Raid0 configuration, preferably with iRam, Acard or multiple SSDs - Single R5870 - Vista + Aero theme Never got further than that

Hey, you found one! :-D

No. Tickets are personal and cannot be passed on. If you can't join, the next in list will.

It shows in hour format when <24h; in minute format when <1h. Perhaps we indeed need to extend this

2010-11-21 23:59:00, GMT+1 There's always a timer on the competition page. And that time is always right, even if the above doesn't comply with it.

If you do, make sure you video tape it. Word on the street goes damaging hardware on tape is the new marketing.

If only you would have ended the 'this is the last ...'-statement with the word 'period' for humorous effect Good luck debugging, though!

What does the BSOD say?

Ok, got it! Thanks for the reply

Quick question: In the past, free versions of the benchmarks allowed upto 5 published verification links and unlimited storage of submissions. Does this mean that the free version only allows 1 saved submission, which can be made public or kept private? Can we save benchmark results offline with the free version to upload in the future (eg: benching on system not connected to internet and upload from other PC)? In addition, do we see the score output on the basic version without having to check online?

If only there would be some kind of medium through which you can communicate and discuss your problems with fellow overclockers. Perhaps I should give you my morse machine number so you can pass me your system details. DOT-BLEEP-DOT-DOT BLEEP-BLEEP-BLEEP DOT-BLEEP-DOT-DOT

Being 1st or 2nd doesn't matter as they both equal winning. In Brian's case, he'd be taking a ticket from stage3 (higher weight) and #3 of stage 2 grabs the second ticket in that ranking.

Do note that I didn't tweak any of the subtimings for optimal performance. High values do not directly mean I had to set them that high for stability, rather those came standard. VTT isn't necessary for all this; it's only this high to make QPI at 5.3G stable. TRFC being this high is due to GBT autotables. For boot, I set the main timings: - tCL=7 - tRCD=8 - tRP=7 - tWR=10 - tWTP=1 When in OS, before clocking up, I changed the following timings: - tRCD_WR = tRP+1/2 (for tRP=7 I needed 8; for tRP=8 needed 10) - tRD_WR = 12/13 - tWR-RD = 12/13 This bios is fucking some things up in the memory timings, though. I need a hard reboot to change tRP. When running 7-8-7, I could never adjust tRP in OS above 7 as it resulted in an instant reboot. In fact, I think in general I was never able to increase tRP in OS to anything higher than boot value.

Actually, submission links are not really useful to solve this problem, at least not when we're speaking of the current practical solution of removing the row with the empty field. The engine doesn't even store the result_id that needs to be recalculated. Since points is function of rank and weight, every submisison affects the point scoring of any submission within that category. Therefore, the engine stores the information of the specific ranking that needs to be recalculated. The engine recalculates the points for each score in that ranking.

For a team that traditionally doesn't do much 3D, I must say this is an impressive result. Well done!

The HWBOT engine has a backlog of all the rankings that need to be recalculated. When submitting a new scores, the backlog notices this, saves the necessary information (eg: what ranking) and does all the necessary calculations. The backlog is flushed everytime the recalculation is triggered: each time a new score is submit or when someone pushes the 'recalculate now' button. In some cases, the engine incorrectly scans the info provided by the submission, or the data of the submission is provided incorrectly and one of the table values is empty. In this case, the engine tries to find the right ranking to recalculate, but will never find it and is stuck with this one entry. You can see this very clearly: when this happens, a new thread is opened on the forums by someone who wonders when he will get points. I just have to manually delete that faulty backlog entry and points are awarded again.

I was stuck at 210MHz BCLK when I first started with the UD6 (just like I was on M3F and GD80). Fiddling with settings suddenly unlocked frequencies I didn't reach before. IMC is good, but I'm not confident enough to say it's IMC only that dictates here. This is, as a matter of fact, just a random chip.

Should be fixed in a couple of minutes

VTT at 1.52V (depends on board). Temps in bios -16°C idle; according to M3F readouts, it's in the positive range whil load.

Improvement of personal best!

DDR3-2280 baby!

DDR3-2200 baby!

Maybe I should've also noted down how much results are worse

Consistency-wise I didn't see any big problems. Re-running the same config over and over again resulted in a pretty stable result output.

Looking cool! Can you drop the big red-PI image and move the Pifast image to the lower right corner with background still black? That allows us to have all the buttons in the black background.