BFG10K
Lifer
First of all, I'd like to clarify that I'm not an anti-ATi zealot. I think the Radeon is a great board and Hyper-Z is really cool. What I'm arguing here is that while the Radeon's scores are consistent with other sites, the GTS scores are much lower.
Right, now onto the points.
Hawk:
First I'll answer the two points you asked me to look at.
There's no hit with a Radeon (well, really small), even on the 32 MB. Besides, just lower your resolution
I agree. I never implied the Radeon couldn't handle 32 bit colour. All I was saying was that the GTS sees a big boost from moving to 16 bit colour. The guy in the article made it sound like both boards receive no boost in 16 bot colour.
I cannot believe that a Geforce GTS is incapable of running Q3A in 32 bit at 1024 though
Your disbelief is completely founded. That website had GTS scores that were too low and that's what I'm trying to point out. A GTS can easily handle Quake 3 at 1024 x 768 x 32.
Or that higher than 55 fps at 32 bit is very playable?
55 FPS is not playable and I was rolling my eyes because you are yet another person to dream up a magical FPS limit and claim the human eye can't see any difference when the scores are higher. I can easily see the difference between 60 FPS and 120 FPS and any half-decent 3D gamer can as well. Going back to 60 FPS seems like a slide show after seeing 120 FPS.
In fact my old Celeron 500 was getting 55 FPS in Quake 3 and it was far too slow and jerky for me. After I ugraded my CPU and video card I'm almost getting 110 FPS at the same settings, and the difference in smoothness is astounding. If I try and play at 55 FPS now it seems like I'm browsing through a photo album.
And btw, if the drivers for Nvidia (6.16) are so fast and awesome and stable and perfect, I wonder why Nvidia is releasing more drivers? Cause lots of people are having problems? There's something left to be desired?
I never claimed nVidia's driver were perfect. Far from it. If you ever take the time to read some of my posts you'll notice I sometimes mention nVidia's drivers have bugs in them, and even the latest 6.72 drivers that I'm using have certain glitches when you're doing 2D stuff. You also may remember that I initially had problems with getting my board to work with my system.
However nVidia's OpenGL and Direct 3D drivers are superb. Every game I've tried has worked beautifully with them.
Taz:
Not on this gentleman's machine. And to reiterate Hawk's point if these were perfect drivers there'd be no need to release new ones.
To reiterate my point, that has no relevance on this discussion. Tom was using the old 6.16 drivers and he was getting better scores than HardCOP or whetever the heck they're called. The newer drivers are bound to improve performance even more.
Now the 70* ATI drivers are quite a bit faster than the 3056s he used. Would have made for some VERY interesting numbers if he'd used them.
Great. I have nothing against that at all. A fair benchmark is a fair benchmark. Howver in this case the benchmark isn't fair because the GTS numbers are too low.
Other websites are obviously NOT using the same settings or they'd have scored lower I guess.😉.
All websites start off with a default Quake 3 config file. Then they change the graphics settings to the preset fastest, fast, normal or high quality, and some even put the settings on max quality, which is simply all the settings on full-bore.
From HardCOPS benchmarks they were quite clearly using the same method of testing (eg HQ etc). If they weren't their benchmarks are meaningless because they never said what settings they used so no-one can reproduce their tests.
And don't forget it's very possible he was using Quake 1.27 and it's attendant lower FPS when benching.
Arrgghhhh! <Kicks Taz in the shins>
Don't you get it? The Radeon scores were consistent with the other benchmarks done on the web at the coresponding settings. The GTS scores were not consistent.
Right, now onto the points.
Hawk:
First I'll answer the two points you asked me to look at.
There's no hit with a Radeon (well, really small), even on the 32 MB. Besides, just lower your resolution
I agree. I never implied the Radeon couldn't handle 32 bit colour. All I was saying was that the GTS sees a big boost from moving to 16 bit colour. The guy in the article made it sound like both boards receive no boost in 16 bot colour.
I cannot believe that a Geforce GTS is incapable of running Q3A in 32 bit at 1024 though
Your disbelief is completely founded. That website had GTS scores that were too low and that's what I'm trying to point out. A GTS can easily handle Quake 3 at 1024 x 768 x 32.
Or that higher than 55 fps at 32 bit is very playable?
55 FPS is not playable and I was rolling my eyes because you are yet another person to dream up a magical FPS limit and claim the human eye can't see any difference when the scores are higher. I can easily see the difference between 60 FPS and 120 FPS and any half-decent 3D gamer can as well. Going back to 60 FPS seems like a slide show after seeing 120 FPS.
In fact my old Celeron 500 was getting 55 FPS in Quake 3 and it was far too slow and jerky for me. After I ugraded my CPU and video card I'm almost getting 110 FPS at the same settings, and the difference in smoothness is astounding. If I try and play at 55 FPS now it seems like I'm browsing through a photo album.
And btw, if the drivers for Nvidia (6.16) are so fast and awesome and stable and perfect, I wonder why Nvidia is releasing more drivers? Cause lots of people are having problems? There's something left to be desired?
I never claimed nVidia's driver were perfect. Far from it. If you ever take the time to read some of my posts you'll notice I sometimes mention nVidia's drivers have bugs in them, and even the latest 6.72 drivers that I'm using have certain glitches when you're doing 2D stuff. You also may remember that I initially had problems with getting my board to work with my system.
However nVidia's OpenGL and Direct 3D drivers are superb. Every game I've tried has worked beautifully with them.
Taz:
Not on this gentleman's machine. And to reiterate Hawk's point if these were perfect drivers there'd be no need to release new ones.
To reiterate my point, that has no relevance on this discussion. Tom was using the old 6.16 drivers and he was getting better scores than HardCOP or whetever the heck they're called. The newer drivers are bound to improve performance even more.
Now the 70* ATI drivers are quite a bit faster than the 3056s he used. Would have made for some VERY interesting numbers if he'd used them.
Great. I have nothing against that at all. A fair benchmark is a fair benchmark. Howver in this case the benchmark isn't fair because the GTS numbers are too low.
Other websites are obviously NOT using the same settings or they'd have scored lower I guess.😉.
All websites start off with a default Quake 3 config file. Then they change the graphics settings to the preset fastest, fast, normal or high quality, and some even put the settings on max quality, which is simply all the settings on full-bore.
From HardCOPS benchmarks they were quite clearly using the same method of testing (eg HQ etc). If they weren't their benchmarks are meaningless because they never said what settings they used so no-one can reproduce their tests.
And don't forget it's very possible he was using Quake 1.27 and it's attendant lower FPS when benching.
Arrgghhhh! <Kicks Taz in the shins>
Don't you get it? The Radeon scores were consistent with the other benchmarks done on the web at the coresponding settings. The GTS scores were not consistent.
