Anyone able to talk to J. Camrack and get a response? I g...
- Thread starter Socos
- Start date
Same here with UT2003. 10x7, 6x AA, 8x AF, all in-game settings maxed. Very playable.
Well lookie what I started here. 
Not bad for my second post.. hehehe
After looking at the linked images, I have to say it is rather obvious. But the reason why I asked if anyone could ask Camrack specifically is because in an earlier plan file he stated that the NV30 was slightly faster when running the NV30 ARB path.
What I would like to know is what settings is he running the cards at to compare. With AA? With AF? Without either? With both?
See what I mean? He never really says what settings and resoultions. It could be the NV30 is faster initally but when the eye candy is turned on it could be something totally different.
Anyways.. anyone have the inside line to Camrack?
Not bad for my second post.. heheheAfter looking at the linked images, I have to say it is rather obvious. But the reason why I asked if anyone could ask Camrack specifically is because in an earlier plan file he stated that the NV30 was slightly faster when running the NV30 ARB path.
What I would like to know is what settings is he running the cards at to compare. With AA? With AF? Without either? With both?
See what I mean? He never really says what settings and resoultions. It could be the NV30 is faster initally but when the eye candy is turned on it could be something totally different.
Anyways.. anyone have the inside line to Camrack?
gokickrocks
Regular
it is with no AA
I think benches show UT2K3 takes a bigger performance hit with AF than with AA.
It's also known that JC is a no-frills kinda guy--IIRC, he once said his base case was 6x4. Heck, I was surprised D3 was demoed with AA--or was it just the promo screens that were AA'ed? I'm sure he's since bumped up his base res, preferably to 10x7 (without AA or AF, obviously, tho it seems AF may help D3 more than AA b/c of the extreme use of bump-maps). At that res, the GF FX U damn well better outperform a 50% slower-clocked 9700 Pro, especially if it's running at lower precision (FP16 vs. FP24).
It looks like D3 may indeed be a multiple-E3 game. Wonder whose hardware it'll be running on come May?
It's also known that JC is a no-frills kinda guy--IIRC, he once said his base case was 6x4. Heck, I was surprised D3 was demoed with AA--or was it just the promo screens that were AA'ed? I'm sure he's since bumped up his base res, preferably to 10x7 (without AA or AF, obviously, tho it seems AF may help D3 more than AA b/c of the extreme use of bump-maps). At that res, the GF FX U damn well better outperform a 50% slower-clocked 9700 Pro, especially if it's running at lower precision (FP16 vs. FP24).
It looks like D3 may indeed be a multiple-E3 game. Wonder whose hardware it'll be running on come May?
IceKnight said:I wonder how why people care so much about carmacks opinion, what if he says he like the gf fx better? would you pay 100-200 bucks more for a lesser product just becouse carmack likes it?
Instead of thinking for yourself, it's easier to just promote someone as your personal icon, and use them to think for you.
Well he does offer sound advice and if you play games based on the Id engines, chances are his advice is relevent. Furthermore, he usually puts his advice in context. I believe he state the NV30 is better if you're a graphics engine coder out to do research and the like. I'm not sure if he explicitly stated which would be better if you're a gamer.
It's really hard to tell whose IQ is better because nobody has an FX. You really have to play both to tell which is better.
I might be wrong, but I think ATI produces 10/10/10/2 output like matrox and nVidia outputs 8/8/8/8 like the Geforce4.
This might make a difference on good quality monitors, but I am not sure. It wouldn't do squat on an LCD.
I might be wrong, but I think ATI produces 10/10/10/2 output like matrox and nVidia outputs 8/8/8/8 like the Geforce4.
This might make a difference on good quality monitors, but I am not sure. It wouldn't do squat on an LCD.
Last time I checked R300 only support 10/10/10/2 as a render target format, not as a back buffer format. Back buffer format is still restricted to 8/8/8/8. I don't know if this is a hardware limit or driver limit.
Furthermore, R300 does not support alpha blending for 10/10/10/2 render targets.
Furthermore, R300 does not support alpha blending for 10/10/10/2 render targets.
pcchen said:
Could you explain what render target format and back buffer format are?
From ATI's web site I was able to gather that DirectX 9.0 has a high precision 10-bit per color frame buffer format so you can see billions of colors instead of millions. (There are two bits for alpha channel blending instead of eight -- My own input). The idea hear is to allow you to see more levels of brightness for each color.
Also from ATI's website I read that pixel shaders 2.0 support multiple render target formats (4 - each is one color). Inside the pipeline each color(RGBA) has 128-bit percision (96-bit in the pixel shader). So I am not sure what you are talking about here.
I am not sure where the 8/8/8/8 "back buffer" comes from, but I am not an expert on the subject either.
Back buffer is a special render target which contains image that is going to be displayed on screen. Everything has to be put into back buffer to be displayed.
R300 can use 10/10/10/2 and floating point format render targets, but no floating point back buffer (DX9 does not support it anyway). I don't remember seeing 10/10/10/2 supported as back buffer format, either.
R300 can use 10/10/10/2 and floating point format render targets, but no floating point back buffer (DX9 does not support it anyway). I don't remember seeing 10/10/10/2 supported as back buffer format, either.
Similar threads
- Locked
