Massman

Just did the Vmem modification ... +0.1V (to 1.7V) makes the card artifact in 2D When removing it (back to 1.6V) solves this problem ...

I think it's on the back near the socket

I'm running a 950 D0 and all slots work ... sometimes. It's a known issue that sometimes the board 'forgets' the last bank and you can only address 4GB. 2 dimm slots ... that's probably a dead board.

Temp ~ -125°C Vgpu ~ 1.55V

For jokes, used this emoticon => <= On topic: still testing on air ... seems like this CPU is pretty decent: 4.7GHz 32M on air with 1.4V

How do you mean?

I just received this board and at first sight it looks quite good. First boot, went into bios and set 220MHz BCLK, perfectly stable . Let's see if it can meet the high expectations!

Hm. Laptops make use of desktop products ... that makes things a bit more complicated.

Socket 7 as well ... Socket A also.

Comparison.

Problem is on the to-do list. problem is unlocked CPU is determining the GPU

Link to score, please.

As for verification, I don't think it's good to change from our current guidelines. So, CPU-Z, GPU-Z and score will be the only mandatory information. Manufacturer, productgroup and model would be optional (just as it is now). Filters could be made, yes, but the database needs to be updated for this. As said before, there's no field indicating whether the CPU/GPU is mobile. Furthermore, we'll have to change something to the submission page, because it's not that easy to add 'notebook' information.

manufacturer = dell product group = inspiron product model = inspiron 17 manufacturer = asus product group = G70 series product model = G70S

So, 3D global ranking per benchmark: - Ranking ~ current global rankings, but all desktop components/systems filtered out - Submission details contain --- cpu information --- gpu information --- notebook manufactureru --- notebook product(group?)

The main point here is how to integrate these rankings into hwbot so that they are more valuable than just a plain ranking. A ranking in itself is not that difficult to generate, that's just a matter of making sure you filter out all the non-mobile cpus or gpus. But ... I suppose it would be interesting to see what kind of notebook you used to achieve a certain goal, right? (IIRC, the database model is not prepped yet to differentiate mobile GPUs from normal GPUs)

All of the Dothan, Merom and Yonahs, for instance .

A problem is that mobile CPUs can sometimes be used on desktop systems ... so we need a way to filter out those results.

Hm. It's certainly interesting as a concept, but need to think about his to implent this exactly. For sure, we cannot mix up the laptop benchers with the extreme overclockers, so global points (or just 'points' for that matter) for mobiles only are not up for debate. Mobile global rankings could work I guess ... give me some time to work it out so that it fits HWBOT

What would it be exactly?

Can you PM me your email?

No extra mods, not even Vmem

V1.50 Extreme

813FPS in nature with single 4850 - bugged!