Massman

This idea was posted about 1,5y ago and kind of follows a similar line of thought. It came up during the discussion following the plans for the GTTP/HTTP Team Ranking (hence the reference too hw sharing). Quite a lot of positive replies to that idea, here's what I commented on that:

It also applies to CPU categories if you consider the 'discovery' of new tweaks. Closing the rankings is not an option, I think, but it would be interesting to brainstorm about this. I reckon there are a few 'out-of-the-box' ideas that might work well.

I understand what you mean and I agree. I don't really see any viable alternative for CPU-Z, though. Not many applications combine 'hard to hack' and 'neat, clean and easy online validation'.

It's an interesting idea, but I'm not sure how it would work practically. Does that mean that, at a certain point, you can no longer do anything with your old 3000+ Venice? One of the things that make hwbot so appealing, imho, is that you can basically have fun overclocking with nearly anything, no matter how old.

I suppose one of the problems is that the hardware/benchmark rankings at hwbot never finish. With any new tweak, any ranking could potentially be completely renewed.

The thing is, though ... Franck SHOULD not anything . Remember, we're just users of his free application. If he listens to our requests, that's very nice, but if he doesn't ... it's also fine.

Yeah. Whether or not to drop CPU-Z since the validation procedure is unreliable.

I'm pretty sure mobile will only work when there's an IGP (eg: SB/IB cpu) as well as a discrete graphics card present.

It's a CPU-Z problem, really. You should ask on their forums why it's so messed up. With LLano, I think it's because they read the HTT incorrectly.

CPU-Z + Llano = sigh

Dunno. I kinda like the fact that staff doesn't stand out visually.

Good news and bad news. Bad news: Linde doesn't seem to be an option, primarily because there does not seem to be a possibility to get a discount across countries (no way to pool LN2 usage to drive the prize down). Good news: Air Products (http://www.airproducts.com/) did respond sort of positively. So, now I'm going to gather the necessary data to check what the possibilities are. Stay tuned!

We're so close!

Hm, I don't see any bugs at first sight.

Yeah, I see your point. The thing is, though, that the argument can go either way. It's just a matter of what you regard as 'stock software environment'. If you consider "all OS tweaks applied" as stock software environment, then any further change in benchmark score is the effect of hardware tuning. I mean: when the software tweak is public knowledge and everyone applies it right off the bat, then it's just part of the 'software environment'. The fact that you can get beaten by lesser hardware isn't necessarily proof of a skewed bench. For example: copy-waza for SuperPI-32M or the previously mentioned vga driver updates.

That's pretty much 'max' speed. 200HT + highest memory divider, lol!

FYI, no benchmark is 'only' about hardware performance, because there's always software needed for the benchmark to communicate with the hardware. Just think about the driver of the graphics card: once in a while, a new driver (= new software) gives you a boost in benchmark score.

What about 12.3 preview ?

You disabled ULPS ?

Everyone has a limit

Wow, performance seems really low! Driver issue?

I think the efficiency is pretty okay given he's using a 2500K (2MB cache less than 2600K)

... sometimes it's nice not being allowed to participate in competitions ...

Whoah! Nice!

In my ear? Are you crazy?!