Sorry to return to this so late! Everyone from johngalt on down I agree with. The original question was "Is there any way at all to increase my GPU performance of my ATI Radeon 3100M 256 MB integrated?" I responded. I failed to take care of the possible risks. Be aware in my posts of this kind that I've cooled single-phase Fluorinerts Fluorinert - Wikipedia, the free encyclopedia
to near gel and immersed an entire working computer (minus case and power supply).
I suspect that from your description etc. that the best you'll get is about 10% o/c. That typically raises GPU scores less than 3% due to bottleneck issues with RAM. Is it worth the risk to try an extreme o/c w/o knowing the signs of impending mobo failure or the right monitoring tools to have running when you try it? Prolly NO.
I will tell you that getting an Eee PC Netbook 1000HEB up from 1.66MHz to 2.2 is doable with remarkable stability -- but only on the internal batteries. I suspect the transformer brick can't feed the instantaneous power requirements. So there's another issue.
As my friends here have warned, there are so many variable to take into account that it might not be worth it. But in the spirit of a guy who's spent his life treating silicon as though it were elastic what's a burnt board than a stepping stone to the next level of expertise?