I.nfraR.ed

Or just disable points and interest will fade away by itself PCMark lovers, don't tell me you all bench it because of the global points...if you're so passionate about the bench, then you will continue even without points, am I right? Don't think so . There are some real PCMark addicts, but I don't see them posting scores lately. Except 1 or 2. They all stopped benching when the scores got so skewed that you can't tell anymore if it's a bug, cheat, tweak or something else... E.g. I bench a lot of old stuff, just because I love it and have fun. I don't get a lot of points, sometimes just 1pt or so, but I don't care about it. These points don't even add up for my global ranking.

The question is why HDTune doesn't work with it, but PCMark05 works and how PCMark is indicative for system performance If a 3DMark run bugs and I see a black screen through the whole subtest, then when my subtest score is a lot higher than comparable systems, my score is invalid (for obvious reasons). When someone discovers another loophole/exploit in PCMark, then it's a tweak. But yeah..I'm a hater, you know .

Congrats to the winners. Very tough battle. Next year Eastern Europe should fight back

That card is sick. I can see the winner...

Go Joanna, go!

Greeks couldn't make it after all? Too bad, voted for them.

That's only valid for online purchase here, but I'm not sure if any shop here will accept it more than once. You have to deal with an "inside-man", otherwise no chance for such tactics. And the place with affordable price has no online shop. Plus that's the only place I can ask for a specific batch. On other places I just get random cpu.

Unfortunately I don't have that much money. Lowest price here is around 310euro, going up to 330-340 on some places. I can't afford more than one chip/month. Tested these on an OC show, hoping for 7GHz one. And it was a one-time thing, so this was my chance for some easy binning. I have no vendor/site/retailer support . At least found a CBB-free 6.9GHz chip that will replace mine, otherwise no real difference. Now I have to buy it on a normal retail price, but still happy I have that chance.

I was lucky with my second chip, but now we found 3 more cpus that are identical to mine - 6.9GHz. I don't call this luck anymore, lol. The goal was 7GHz...

I vote for Costa Rica. Even though Malay might have better IMC, it's harder to utilize it, e.g. you need really high-clocking memory. You'll have higher gains with better clocking cores.

5 3770K chips 3218Bxxxx tested on LN2 out of 25. 3 max clocks ~6.9GHz on one core, 2 - 6.6GHz All of them can handle 1.9V or higher. Various voltage for 5GHz on air, but in the 1.28 - 1.32V range for the tested 5 chips. Some cpus from this batch didn't even boot at 5GHz on air, however I tested only 4 on LN2 myself.

CPU can run much higher than this, I just clocked down with low vcore to save LN2. Dry ice is 3 times more expensive than the LN2 from my supplier, lol. But you're right.

Yeah, I have a lot of good cards that are waiting for their turn The problem is the cpu eats a lot of LN2...

Tested HD3870 yesterday and the cpu worked flawlessly full pot through 03, 05, 06, aquamark. For example - 03 Batch is 3218B987. 2 cores, 4 cores, 4c/8t - no problem. But had strange issues with my first one, however not that bad, but still from certain point I couldn't finish a single run with some cards. Haven't tried with PCI-E 1.1.

Most of the recent good chips are 3218 and 3219 I think.

Similar thing happened to me with the hd3870, but after switching the slots 2-3 times it finally got back to x16 and I forced Gen2 from bios. Not sure if it's cpu related or just a bad contact (I thought so). And I had very odd behavior for all tested cards during the TeamCup with the cpu on cold, however it was with the old one. The new one runs fine I think, except the x2 issue I had once. Unfortunately can't test it on other boards, because I have only one.

The efficiency king!

Ok, haven't dig inside, but the values are defined in the xml, so I've missed that. Like this: <CustomSettingValue> <UserfriendlyName>14.250</UserfriendlyName> <HexValue>0x00000073</HexValue> </CustomSettingValue>

That's not considered a good joke here in Bulgaria .

nvidiainspector sets it no problem. There's even a modified version with LOD maximum > 3. You can use the standard version as well, but you need to input the LOD offset as a HEX value. The values in the dropdown are just predefined values with labels for convenience. The step is 0.125 and looking at the first value in the inspector, hex value for it is 0x0000001. LOD = 5 would be 5/0.125 = 40 which converted in HEX is 28. So you need to input 0x00000028 in the LOD box. As simple as that

Yes, rivatuner config file, it's called rivatuner.cfg or something.

You need to add device string in the config file in order to get it working. Open the config file located in the riva folder and then search for nvidia section. Then open gpuz and see what is your device id. Add that device id in the appropriate section, then search for the second section where the device string is specified, e.g. (Nvidia GeForce GTX480) which will correspond to the graphics core added in the previous section. If it's not clear enough, search in google, there should be some guides, or I can help you further when back at home. This is the way it is done for every unsupported cards, regardless ATI or nVidia.

Sam is right, this should be the issue if the board sets the NB at x10 by default, however Bulldozer NB should be 2200. Not sure how my giga sets it though. PS: Uhm, excuse me, the 2200MHz is not valid for FX-6xxx and FX-4xxx, so yeah...this should be your problem. They have it at 2000MHz stock.

A lot of good chips lately, but it's unusual to see Aristidis behind me, lol. Push more

Hey guys, Anyone of you tried the Noctua NT-H1 under LN2 (ihs not removed)? I tried it now and had the worst experience, but not sure if the TIM or some other factor. Don't know if I'll have enough LN2 for a second test, so I'm asking here. If it's bad, then I'll save the rest of LN2 for something else.