I.nfraR.ed

I tried it and it is sloooow. Wasn't that supposed to make my AMD better than 4770K in Pi32M? Because it doesn't...

That's his strategy - bin chips, then keep 'em until they are not popular anymore. Then run at max and with his eff the chances the score gets beaten by someone in the future is small

So HDMI is not required and I can get away with the connector. Thanks! That's what a friend told me too, so I will definitely try it next time. I've never touched 680, newest nvidia I've benched before 980 was GTX580 and that was not so long time ago.

That's great, thanks! Very useful for those of us who can't get unlimited LN2 or different 980's. Can you tell something about VGA connectors? I've tried mine just once, but was not able to bench at lower than -80 or so without blinking, artifacts or lockups. Low mem freq didn't change anything. Maybe memory temp was getting too low, maybe something else, but I had no HDMI cable to try or DVI -> DSUB convertor.

So, does it mean only valid DDR4 subs would be from those participating in the handful live events during the year or records by vendors with oscilloscopes? How would you know some random guy didn't tweakthe registers if he didn't bench live or has no oscilloscope?

I don't want your rejects Seriosuly though, not bad at all.

http://webap1.dfi.com.tw/DFI-Lanparty/;jsessionid=2F07D94D600994C77FAA53603E5C305D Select either bios or beta bios. Official bioses for 790FXB-M2RS: http://webap1.dfi.com.tw/DFI-Lanparty/support/MBBiosDownload.jsp?showselectvalue=30LP_DK&selectBiosItemId=539 Beta bioses: http://webap1.dfi.com.tw/DFI-Lanparty/support/MBBetaBiosDownload.jsp?selectedBetaBiosItemId=539

I haven't seen 890FX board that can't do at least 400MHz HTRef, but I've only tried Crosshairs and M4A series, plus one MSI GD80. I 've also had M4A89GTD Pro doing 450+ on stock cooling, but that submission was removed long time ago, due to "red" validation caused by HTRef fluctuation. Don't remember trying 2GHz DDR3 on my Semprons, but I believe it is not that hard, since most AM3 (excluding Thuban) DDR3 "top scores" are done with Semprons. Highest benchmark stable I've ran on X2 555 was 966 6-6-6-17 dual channel and I guess higher was possible with slackier timings and single-channel/Ganged mode. Pretty sure it would run higher even on these timings, but I suspect the CPU freq was the limiting factor in that bench. PS: I was wrong about Semprons, but there are certainly some scores above 1000 with different AM3 CPUs, including Sempron.

Yes, last week, but we had a lot of problems - first time 5960X on cold and first time 980 on cold too high CB on both. Will try again, but will have to use 3770K.

Btw, what is the voltage tolerance of these chips? I'm sure someone tested it I went up to 1.7V on my Strix (on LN2 ofc), for a short time though, because I had other issues. Would be good to know, especially for us who bought the card retail and have no chance to RMA it, since the warranty was void in the moment I removed the seal of one of the screws on the stock cooler.

Yes, but it doesn't make sense to not allow at least 800-based devices. Always wonder why some comps are limited in this way, basically "compete with specific hardware to win more of the same".

Compete with your Snapdragon 801 device to win another 801-based smartphone. Nice.

I thought ES are not allowed in competitions?

PSC on Sandy are easier to work with, but properly tuned Hypers should be faster.

I've convinced MaxOCer to join my country. We're going to win...

Just finished arranging mine

Not 100% sure, but I think it is between 15 and 20.

Nice placeholder

Thanks for the info. I don't want to go above that as well, since I plan to freeze it anyway.

Btw, what was the vGPU? I'm hitting 1740+ with 1.45V on a single Strix and chilled water.

These Classies are excellent on air/water cooling. Haven't seen any other brand match their clocks on ambient temps.

Sorry, but that looks like a bugged Nature subtest.

Geometric and shader domains run at different frequencies. You can control that with the delta value in bios, so you need to edit the bios and reflash. ROP domain also works at the shader domain frequency. Basically the geometric delta value in bios determines what would be the difference bewteen geometric domain freq and shader/ROP. What you set in e.g. RivaTuner is your shader/ROP, then Geometric domain clock is that X freq + delta value in MHz. For 7900GTX normally it is 40MHz, I think, but not sure about 7950GT. So if you hit the limit of your geometric domain, you can decrease the delta. That would allow higher ROP/shader frequency at the same geometric clock.

Dino, ES is not allowed in competitions. I'm afraid you need to bench again with retail cpu. And yeah, bring back the good old logo, this one is too generic.

The one I tried for the team cup had a bad CBB, so I don't think LN2 will help much if all are like it. Plus it didn't raise the max stable bclk frequency, so I ended up running it on air. It was with an external VGA though, so it might make some difference.