Well, I purchased an ATI 9700 gold for my latest system. I went this route due to the fact that the 5800's seem to pretty much suck, and they were discontinued so fast that if I wanted one, I would be hard pressed to find one. Also, The 5600ultra v2 is not out yet, and the current crop of 5600/5600u are not quite up to par on performance. The 5900 ultras just hit the streets so are still demanding a premium, and I have not seen a 5900 or 5900value yet. I wanted a 5900 (non-ultra) which just seems to be the same as the ultra without 256 mb ram. But if I had waited, the first release would have been demanding premium prices again as nVidia fans will be flocking to this card when the supply is low, driving the prices up. I didnt want to wait, so I settled on an ATI 9700 gold for $220 shipped.
Unfortunately, timing dictated that I had to swap the whole system (cpu, mobo, ram, video) rather than getting a chance to try the card in the existing system. (I upgraded from a gf4 4400).
After fighting the ASUS p4p800 (article in the troubleshooting forum), I got the system to run stable, and started testing. The ATI drivers installed with no problem, although the readme said it would run a series of tests upon reboot, which I never noticed. I started with the CAT 2.4 on the disc that came with the card. I immediately upped to the 3.4 drivers. I ran a dx9 game (NDA.. cant disclose) and 2 minutes in I got a hard lock. Thinking this was still a memory issue, I worked on some memory settings, and loaded up 3dmark03 (latest build) to test.
First time I fired it up, I got to frame 556 of the first test and got a hard lock. After rebooting, and resetting some memory settings (went to 3-3-3-3-5 from 2.5-3-2-2-5), I ran 3dmark03 again and got to fram 556 and hard locked. I suspected something other than hardware at this point, but wanted to be sure. I rebooted, set memory to 320mhz from 400mhz. Frame 556 it hard locked again. Now I was sure it was not memory. I scoured the web looking for catalyst 3.2 drivers and found some in europe. After installing them, which went perfectly smooth, although I was going from 3.4 down to 3.2. I fired up 3dmark03 again, and it ran perfect. I scored a 4089 with the slow memory settings. I fired up the beta I am playing and it ran perfect for 2 hours. I had gone back into bios and set my memory back to 400mhz, 2-3-2-2-5.
Tonight I will play with the PAT control, as well as seeing if I can run 2-2-2-2-5 without a glitch. The mobo was trying to overclock the system by itself early on, but I put a stop to it before I got to the testing.
First Impression: seems to run well, 3.4 drivers are obviously garbage. DX9 apps oviously are calling a function that the 3.4 drivers cant do. BTW, in the game I am beta testing, the 3.4 drivers are causing all kinds of havoc with lighting. I had no probs with the 3.2. Overall the card seems solid at this point, and I have yet to try out AA and AF. I increased my CPU horsepower by about 25% overall, and my ram is running about 80% faster than before, and is dual channel now. I wish I knew exactly what factor I could subtract to get an accurate comparison, but oh well. The game I am beta testing ran better by about 25% overall, and did not really look better. The framerates in the game went from around 22-26 avg in light areas to about 23-29 and feel smoother. In hard areas I went from 1-18 fps avg to 4-24 fps avg, and once again feels smoother. The rough areas before saw frame rates dropping below 5 about 30% of the time, and with the new setup, frame rates dropped below 9 about 25% of the time. I actually had worse performance in one area, but certain objects in the terrain were enabled automatically with the new card, and dialing them down regained the performance loss. Overall, the card seemed to perform well, about as expected. I am still going to upgrade to the 5900 when it is reasonably priced (read: 375 dollars). ATI has come a long way in a short time with their hardware, and even though the latest offerings of drivers are screwed up, they are better at building drivers than they were.
Perp
Unfortunately, timing dictated that I had to swap the whole system (cpu, mobo, ram, video) rather than getting a chance to try the card in the existing system. (I upgraded from a gf4 4400).
After fighting the ASUS p4p800 (article in the troubleshooting forum), I got the system to run stable, and started testing. The ATI drivers installed with no problem, although the readme said it would run a series of tests upon reboot, which I never noticed. I started with the CAT 2.4 on the disc that came with the card. I immediately upped to the 3.4 drivers. I ran a dx9 game (NDA.. cant disclose) and 2 minutes in I got a hard lock. Thinking this was still a memory issue, I worked on some memory settings, and loaded up 3dmark03 (latest build) to test.
First time I fired it up, I got to frame 556 of the first test and got a hard lock. After rebooting, and resetting some memory settings (went to 3-3-3-3-5 from 2.5-3-2-2-5), I ran 3dmark03 again and got to fram 556 and hard locked. I suspected something other than hardware at this point, but wanted to be sure. I rebooted, set memory to 320mhz from 400mhz. Frame 556 it hard locked again. Now I was sure it was not memory. I scoured the web looking for catalyst 3.2 drivers and found some in europe. After installing them, which went perfectly smooth, although I was going from 3.4 down to 3.2. I fired up 3dmark03 again, and it ran perfect. I scored a 4089 with the slow memory settings. I fired up the beta I am playing and it ran perfect for 2 hours. I had gone back into bios and set my memory back to 400mhz, 2-3-2-2-5.
Tonight I will play with the PAT control, as well as seeing if I can run 2-2-2-2-5 without a glitch. The mobo was trying to overclock the system by itself early on, but I put a stop to it before I got to the testing.
First Impression: seems to run well, 3.4 drivers are obviously garbage. DX9 apps oviously are calling a function that the 3.4 drivers cant do. BTW, in the game I am beta testing, the 3.4 drivers are causing all kinds of havoc with lighting. I had no probs with the 3.2. Overall the card seems solid at this point, and I have yet to try out AA and AF. I increased my CPU horsepower by about 25% overall, and my ram is running about 80% faster than before, and is dual channel now. I wish I knew exactly what factor I could subtract to get an accurate comparison, but oh well. The game I am beta testing ran better by about 25% overall, and did not really look better. The framerates in the game went from around 22-26 avg in light areas to about 23-29 and feel smoother. In hard areas I went from 1-18 fps avg to 4-24 fps avg, and once again feels smoother. The rough areas before saw frame rates dropping below 5 about 30% of the time, and with the new setup, frame rates dropped below 9 about 25% of the time. I actually had worse performance in one area, but certain objects in the terrain were enabled automatically with the new card, and dialing them down regained the performance loss. Overall, the card seemed to perform well, about as expected. I am still going to upgrade to the 5900 when it is reasonably priced (read: 375 dollars). ATI has come a long way in a short time with their hardware, and even though the latest offerings of drivers are screwed up, they are better at building drivers than they were.
Perp
Facebook
Twitter