Every ATi mobile card using 3DMark05, READ this!

Hi,

I wanted to tell you about a little glitch with drivers/3dmark05 that took me a long time to figure out. This may just be with Omega drivers, but I have seen it in Omegas 6.3 and 6.7 (my last 2 sets).

If you have a mobile ATI card, you may be aware of the powerplay option in the ATI control panel. There you have 3 sliders for when powerplay IS ENABLED, for AC, battery, and low battery. There are 3 options for each, "optimal battery life", "balanced", and "optimal performance".

You would expect that no matter what settings you have for the above, if you have powerplay DISABLED, your score in 3Dmark is the same. IT'S NOT!

Even if you have powerplay disabled (no checkmark), but if even one of those sliders is at anything BUT optimal performance, 3Dmark will give a lower score. I don't know how it does this. Clocks are still what they should be with powerplay disabled, so it does some kind of weird half-powerplay mode.

I don't believe this performance issue also happens in games, so feel free to leave the sliders at whatever you wish except for when using 3Dmark05.

Just wanted to warn you guys.

 

Interesting. Props on the 2k score with the X300. That is pretty amazing.

 

That is interesting I always have powerplay disabled, even on battery. When I do 3Dmark05 I get accurate scores though. That is how I got the one in my sig. When I open up my powerplay tab though, even though I do have powerplay disabled, it says me active setting is optimal performance. Interesting none the less. I have read about some people getting insanely low scores and maybe that is why.

 

Whoa, yea, that is a nice 3Dmark05 score, got me beat. How did you get your mem clocks so high? That is incredible. Mine would most likely become unstable. Do you use actually use those clocks, or was that just to get a score.

 

The clocks I used for scores would never be past stable/artifacting. However, the clocks in my sig, the core I have lowered a bit since, but I forget why. I think it was crashing some game but I forget which. I will bench again with my current clocks once I get omega's 6.3 installed.

Remember, the mem clock you see is doubled b/c I left it as DDR. I also messed around with my memory timings and that put me up ~50 points. But I almost killed my card a few times doing it. (Even though that might've been cause I accidentally overclocked by 100Mhz more than I was supposed to, search my past threads if you wanna read up on that lol)

 

Flav_cool, is that 444.60/651.84 Core/Memory, or Memory/Clock?

My 256MB x300 only gets around 1330 3DMarks05 at 370Mhz/270Mhz Core/Mem with Omega drivers. Maybe I'll give that powerplay slider thing trick a try.

 

It's core/mem. Yes you should. Is your x300 the shared memory version? That may be causing the low score. I haven't heard of a 256mb version. Mine gets around 1400 stock clocks @ 297.00/229.50 (or 297/460DDR) and stock mem timings.

 

Flav cool when I am at home I am gonna try it mate. So you say:

Set everything Maximal Performance and then deselect Powerplay. Is that correct?

Charlie

 

My mobility radeon x300 has hypermemory, i.e. 128mb dedicated, 128mb shared.

At the default clocks of 300/230, I get 1150 or so. I only get >1400 when I bump the clocks up to 410/330Mhz or so.

And I tried the powerplay slider thingy, and no improvement there.

 

Yea. Don't forget Idon't know if this is a bug with all ati cards or just mine (X300)