9600 GT vs 9800 GTX+

[Spirrwell]

I've had these 9800 GTX+ cards sitting around for a while that were said to not work properly when I got them. I decided to replace my 9600 GT card with the faster 9800 GTX+ to see if it would work. After a few crashes, BSODs, downgrading drivers a bit, and reflashing the BIOS of the card it seems to be a bit more stable now.

I wanted some way to test it all, and for whatever reason oclHashcat popped in my head and I gave it a go. Logically speaking the 9800 GTX+ should outperform the 9600 GT I have in every aspect except it only has 512 MB of VRAM as opposed to my 9600 GT which has 1 GB.

I ran the cudaHashcat example 0 on both cards. The 9600 GT got near 200 M/s whereas the 9800 GTX+ only got a little over 100 M/s. Now perhaps I'm reading it backwards, but is that supposed to happen? What's making the difference, is it because my 9800 GTX+ is likely bad, or is it because of having less VRAM?

I'm sorry if this is kind of unrelated to the purpose of oclHashcat, but it seems like it would be a great program for testing a GPU. Any insight is greatly appreciated.

[epixoip]

less vram is inconsequential. the 9800 should be faster than the 9600, and even though cudaExample0 may not be enough to achieve full acceleration, the 9800 should still post higher numbers. so it could very well be that the card is bad.