I.nfraR.ed

Someone in the US will have an advantage Just passing by to say it is a very clean work and the price is low for what you offer. Good luck with the sale.

Nailed it!

That is correct. All results in single-threaded benchmarks like superpi, pifast, cpu frequency go to the unified ranking and only your best counts, no matter how many cores you had enabled during the benchmark. Btw, your 3952MHz result is invalid, because the validation file only shows x36 multi, not x38.

Congrats! Almost total domination. Germany and Australia pulled some great scores, too. I could probably compete with the AMD stuff, but not with the rest. Hopefully things will be better for us next year and we meet again

Nice one, lads.

THIS! It is giving a chance to countries with very few active overclockers actually compete without cheating. For example I can do several stages with several different cards if I have the time and money to invest in it. But when we are 3 active overclockers, even if gather together, we will still need to hand over some results to other people. And I don't want to do that, although we all know it's a reality. Let's say my team benches together and all members come to the party. The problem is half of them are inactive most of the time. So some of the active overclockers (or we as a group) will bench the hw and someone else will submit it. So what's the point of disallowing more than one result per overclocker? The only difference is it is uploaded from the profile of someone else. Doesn't really matter he was in the same room watching or drinking beer. It only adds to the fun factor, but nothing stops us get together even if more than 1 result per overclocker is allowed. It even makes it better, because we would know we ACTUALLY have a chance. When we submit 3 out of 5 required, even if they are WRs it doesn't matter. Absolutely no motivation for that. It is practiacally impossible for a single guy to cover all the stages with top of the line LN2 results, but if someone does, then why not allow it? Let's make it official.

If for some reason @Razwar gives up on #3, I'll take it.

You didn't think the result is strange? Especially when your previous run is ~59K with (almost) the same clocks?

Shaved 1 minute off. Still ~10 sec slower than Linux, but I'm on a very old Win 7 installation. It might be better on e.g. 10. So I guess the official Linux build is newer than official Win build. PS: Checked both timestamps and they match: both are compiled on 24.10.2016. Wonder if there will be such an improvement if someone compiles the linux version (update: building it right now). v2.78.4 for Linux (UNOFFICIAL, compiled from source) blender-2.78.4-unofficial-linux_x86_64.tar.bz2

FX-8320E@5117MHz running Linux Mint 18 x64. On Windows 7 SP1 x64 it was more than 2 times slower!

1 from second group, thanks.

Get an universal one - it's a better investment.

Bad, bad Michael! I will be a rebel then - save the quadro bios, then flash it again

And what about the other way around. Are we allowed to flash a quadro with a geforce bios?

This is impressive for a KT600 chipset.

Keep 'em coming, man I also have some good cards that need to be benched some day.

Yes, true. All/most custom Gainward and Asus PCB use one of these NCP controllers and the only way is VID mod. Probably others too, but that's what I had. It's not that hard and in most cases you don't want to support all possible VIDs anyway. Sometimes not all VID pins have exposed resistor, they are just grounded. Maybe there's another way, but this is the most reliable one. After VID-modding the card there's no OVP anymore on the ones I've modded.

Can the memory go higher? My best card (reference Asus) did 1120/1340 on water before it died mysteriously All my other 4890's are in the 1250-1290 range.

Oh, you're back . Excellent result and great cooling. Keep up the good work!

SM3 test 2 (Deep Freeze) seems to be bugged.

Was looking back to these ranking and this is one of the few proper (aka efficient) Pi32M scores at the high frequency range. The top scores are fine, but when you go down the table there are so many awful scores. The SandyBridge effect - people getting lucky chips with easy overclock without actually knowing how to properly bench them.

I wanted to try mod the bios, but the one cold slow bios for 9800 ref card linked on hwbot is exactly the same as one of the "stock" bioses on TPU. I can't find the "before" and "after" of this exact bios, so can't understand what is changed exactly, comparing with HEX editor.

Great score, guys.