Massman

Yeah, I looked into your team as it surprised me too. I think the reason why you guys dropped so much is because you're mainly using the same hardware. Using identical hardware was very beneficial in R3, but not anymore in R4. //edit: actually, what you can see is the hardware sharing prevention kicking in. The more a team drops, the more this team would be prone to sharing accusations because the members use a pretty similar set of hardware.

You guys are slamming the server already! Enthusiast league = air/water only + requirement that the submission needs a picture of the setup. Deleting LN2 results will definitly raise your ranking in the enthusiast league. Just testing the effect. Not giving global results to CPU-Z makes sense, though. Comparing frequency has only value if you compare it to similar hardware. 7000MHz with Gulftown is much more impressive than 7000MHz with Cedar Mill. No, that's a typo. Team points = SUM(powerpoints) + [sUM(user points) / 10] with user points = SUM(all user points) (global, hardware, wr)

A bit later than planned, but here it is: http://uat.hwbot.org/ Some important notes. First, as you can see at the bottom of the site, we are still running a beta version of the R4 code (HWBOT v4.0.0-SNAPSHOT build 3115). In practical terms, it means that this public test server is to show you the concept of revision 4 (layout, etc) and the final hwboint algoritms. Some pages are still in development and some are finished but not added to this build. Secondly, the rankings. The algoritms you see on the test server are final. The rankings are not entirely up-to-date since we're not running a live version of the database, but a full backup of a couple of days ago. This means that submitting a result will have no effect. Important! Do not submit any benchmark results on our test server. These will not be added to the result database and will be lost once we install a new build on the server. The rankings: As for: the Overclockers League: the new algoritm has no hw cap. Instead of using a hardware cap, we've limited the hardware points contribution to your personal total to SUM(top-20 hw results). So, the maximum hw points you can reach is 20x50=1000p. But, to reach that, you'll have to be 1st in 20 highly competitive rankings! To check the powerteam ranking, just click on 'best of team' in each ranking. The results will then be filtered on unique team (instead of user). The point distribution of the powerteam rankings is not yet clearly visible in this version of the test server. The algoritm used to calculate the team powerpoints is the same like we use to calculate user points, but with different constants and offset points. As said, the weight between powerteam and user points is 10/1, but in absolute figures they are pretty much balanced at the moment. Some teams have big user contribution, but lower powerteam and other teams the other way around. Grinding will still help, but for each individual 10x slower than it does now. Fyi, what K404 did a few days ago (massive amount of top-5 results) is pretty much golden for revision 4 as those rankings are highly competitive powerteam rankings. As for the different user rankings: if you want to see the effect of switching from one league to another, you can choose the ranking you want to participate in from your user profile settings: http://uat.hwbot.org/profile/. As for those who wonder about the Pro OC League, here's a short explanation I gave to *censored* a few days ago. It pretty much sums up what it's all about. As a last not, I'm currently fiddling with the points for the different applications. Rev4 allows us to give benchmarks either no points, hardware+global points or either hardware or global points. A few minutes ago, we gave 5 benchmarks 'hardware points' (eg: PCmark vantage and 3dm11 performance) and I also removed global points from CPU-Z validation (hw points only). The engine needs to do a recalculation first, so i don't think the effect is already noticeable. Anyways, have fun and let us know your feedback! We'll keep you updated when installing new builds!

In the end I got 5710 running ok, but with annoying temp range of 7°C-8°C. That it took a week to get the setup running like I wanted was not planned . X58A-OC should be a bit more interesting, I think ...

You like?

I spend (only) one week trying to get 01 to pass at 5700 with the 4890. Just exchanged the P67 platform for the X58OC one. This week some 990X pre-testing .

Kinky!

- GA-P67A-UD4-B3 bios (F4): http://91.121.148.119/downloads/BIOS/p67aud43.f4 - GA-P67A-UD5-B3 bios (F4): http://91.121.148.119/downloads/BIOS/p67aud53.f4 - GA-P67A-UD7-B3 bios (F4): http://91.121.148.119/downloads/BIOS/p67aud73.f4 - P8P67 bios (1503): http://91.121.148.119/downloads/BIOS/P8P67-ASUS-1503.ROM (older than 1506) - P8P67 Le bios (1011): http://91.121.148.119/downloads/BIOS/P8P67-LE-ASUS-1011.ROM - P8P67-M bios (0901): http://91.121.148.119/downloads/BIOS/P8P67-M-ASUS-0901.ROM - P8P67-M Pro bios (0709): http://91.121.148.119/downloads/BIOS/P8P67-M-PRO-ASUS-0709.ROM

Auch. Took this to the staff center and awaiting the response of other moderators. Looks pretty clear.

Hooray!

You can use the "prepopulate fields by previous submission" option.

Hm. In your 124k run? [hwbot=2163352]submission[/hwbot] You're at 5850MHz with the CPU. I've only got 5710MHz, which makes quite a difference even in Nature. That being said, I did have this run on the UD4: [hwbot=2099979]submission[/hwbot] 1100FPS with only 1000 core. Not sure if it was part of the tweaking, but ever since I started working with the UD7, I felt nature was a tad slow.

This: From 56x and up, I'd test them under liquid nitrogen to see if they scale a bit. Remember, on Sandy Bridge the maximal operating frequency is function of temperature and voltage; some CPUs can only run highest frequencies at -10°C. And also this: GBT's UD7 is so much hard work. I'd say 'but eventually you get there', but so I'll do that when/if I eventually get there

First implementation (in R4.0) will be basic listing. We've moved all the neat stuff to R4.1 as we realised this new feature would be a lot more work than initially planned out. The idea is to have a custom page for every hardware you have tested or used to submit results with. This page will show your pictures, the OC results and of course has room for personal notes.

Elsa had some good GF2/3 cards, if I remember correctly. With the right kind of memory chips.

Product pages for processor and videocard families (like Core i7) are planned, but bot sure if it'll make 4.0 release. Maybe 4.1. User/team info extentions are for 4.1 or later too. It needs coding + debugging, which takes time. Time we don't have growing on trees

Product pages for processor and videocard families (like Core i7) are planned, but bot sure if it'll make 4.0 release. Maybe 4.1. User/team info extentions are for 4.1 or later too. It needs coding + debugging, which takes time. Time we don't have growing on trees

Hey Chris, can you please tell me how accurate this image represents your current situation?

That's like asking why the owner of a brewery has more free beers than you ...

BUT! You got 2 points for killing!

Cat is most active. Polar bear and Panda are ... well, still in training. So, not quite as productive.

Some small bits need a bit more work. Like this:

Don't expect any miracles from Rev4 guys. If you're not able to get a cup or medal in rev3, you won't magically get them in 4.

A blindfold? Mark, that's just creepy!