Maybe this is old news, or more than likely, no news at all, but I was adjusting the BIOS settings for my Radeon 9700 Pro (using RadEdit 1.1d) and I happened to notice a memory timing box in the editor. The default setting in my original BIOS was CAS2, but if it can be adjusted, how would I ever know what benefit or harm would come from changing it if I didn't try. 
I figured it could be bad news if I changed the timing to CAS1 and left the memory overclocked, so I lowered it back down to the default clock speed (~310) and moved the timing to CAS1, saved the BIOS file and rebooted so I could flash my card. I ran 3DMark 2003 with the CAS1 timing at default memory speed and then flashed the card again to change it back to CAS2 at the default memory speed and ran 3DMark 2003 for a comparison. The scores were almost identical, 5160 vs. 5161, with the higher score coming from the CAS2 run, so I almost concluded that changing the timing was not doing anything at all, and then I saw a strange difference between the 2 online results summaries. In the results browser, the summary for the normal run at CAS2 shows me:
GPU: ATI RADEON 9700/9500 Series
378 MHz / 311 MHz
But the summary for the run at CAS1 shows me:
GPU: ATI RADEON 9700/9500 Series
378 MHz / 621 MHz
Not being one that jumps to conclusions, I performed these tests again, 2 more times at CAS1 and 2 more times at CAS2. The results were always the same: when CAS1 timing is set in the video cards BIOS, 3DMark reports the actual data transfer rate, but at CAS2 it reports the memory clock setting, not the transfer rate.
Anyone else ever monkeyed with their Radeon 9700 Pro memory timing? I haven't pushed it yet with CAS1 timing and any overclocked speeds to see if it performs any better or locks up, but I found this to be as mysterious as Area 51.... well maybe not that weird, but almost.
Comments???
I figured it could be bad news if I changed the timing to CAS1 and left the memory overclocked, so I lowered it back down to the default clock speed (~310) and moved the timing to CAS1, saved the BIOS file and rebooted so I could flash my card. I ran 3DMark 2003 with the CAS1 timing at default memory speed and then flashed the card again to change it back to CAS2 at the default memory speed and ran 3DMark 2003 for a comparison. The scores were almost identical, 5160 vs. 5161, with the higher score coming from the CAS2 run, so I almost concluded that changing the timing was not doing anything at all, and then I saw a strange difference between the 2 online results summaries. In the results browser, the summary for the normal run at CAS2 shows me:
GPU: ATI RADEON 9700/9500 Series
378 MHz / 311 MHz
But the summary for the run at CAS1 shows me:
GPU: ATI RADEON 9700/9500 Series
378 MHz / 621 MHz
Not being one that jumps to conclusions, I performed these tests again, 2 more times at CAS1 and 2 more times at CAS2. The results were always the same: when CAS1 timing is set in the video cards BIOS, 3DMark reports the actual data transfer rate, but at CAS2 it reports the memory clock setting, not the transfer rate.
Anyone else ever monkeyed with their Radeon 9700 Pro memory timing? I haven't pushed it yet with CAS1 timing and any overclocked speeds to see if it performs any better or locks up, but I found this to be as mysterious as Area 51.... well maybe not that weird, but almost.
Comments???
Facebook
Twitter