nVidia's response to Valve
- Thread starter Joe DeFuria
- Start date
Bouncing Zabaglione Bros.
Legend
DaveBaumann said:
Hmm, if it's a next gen benchmark that needs PS 2.0 in order to render correctly (in this case for HDR), then surely there is an argument for saying cards that can't render properly should be penalized in the final score? Dare I say it, much in the same way that cards were penalized in 3DMark2003 for not being able to run the more advanced shaders or not being "enough" DX9 to run correctly/quickly enough.
Otherwise, by the same logic, shouldn't the benchmark be reduced down to DX7 levels because a GF4MX can't render equally to the DX9 standards that the game is actually targeted for?
Welcome back Dave, I appreciate the above insite and look forward to more of your take on this....but go get some rest so I can go crash before you 'plain it all and I miss the fun, I mean get some rest and hit it in the morning.DaveBaumann said:
Pardon me, but what should we think about the Half-life 2 benchmarks now, in light of the above?
It seems that an important feature of the final engine has been disabled in the benchmark version. Wouldn't this affect the scores of different video cards? Wouldn't this make any current comparisions between various video cards, especially the NV3x/R3xx line, invalid?
Also, a lot of screenshots have been released from the benchmark version, including the image of the cave + pool of water DX 8/8.1/9 comparision. If HDR is not activated, then the DX9 version can't be what it should. Let me also note that many people reported that they see no big difference between DX8 and 9, so they don't understand what's the big fuss is about...
And Valve has disabled HDR just because NVIDIA has a driver problem (which should not even be there in a WHQL driver) - it is strange that now they decide to favor them...
The point of the DX9 benchmark was to create like for like conditions. The standard DX9 path does this, as both NVIDIA and ATI are doing the same amount of work. HDR is actually a very new addition to the game, however if it were enabled by default in the benchmark then either, due to a lack of float buffers, the FX would fail to post any scores, or this effect would be disabled and there wouldn't an equal comparison.
IT has to be said, though, that I don't think everyone there appreciated this point and with HDR enabled there will be visually dramatic differences between the DX8 and DX9 paths.
nelg said:
This feature all comes down to driver support. Currently Nvidia's drivers do not offer float buffers. It is unknown if/when Nvidia's may support this.
My own opinion on this is if they don't have support for this now, close to a year after DX9 release, then they will never support it on drivers for NV30/NV35 cards. So much for their golden drivers...
Microsoft did mention they might do a point-release of Dx9 to support a new conditional floating point texture/render target formats.
Similar to conditional non-pow2 limitations, it should allow cards that can't do full Dx9 floating point texture enough usage for HDRI etc.
Similar to conditional non-pow2 limitations, it should allow cards that can't do full Dx9 floating point texture enough usage for HDRI etc.
Question >
"mshdribl" is a HDR demo that was released earlier this year. GF3 (DX8) users can run the demo. If DX8.1 can't support HDR due to the lack of 'range' > how can the GF3's run that demo? If it is software/CPU emulated, can't that be driver implemented for all DX8.1 & lesser cards?
:?
Hmmm, http://www.4gamer.net/specials/hl2int/hl2int_e1.html HDR was discussed @ SIGGRAPH 2003 in July, according to that interview w/Valve. Seems Splinter Cell also uses HDR for DX8. Hmmm.
Edit: correction on the demo & date.
"mshdribl" is a HDR demo that was released earlier this year. GF3 (DX8) users can run the demo. If DX8.1 can't support HDR due to the lack of 'range' > how can the GF3's run that demo? If it is software/CPU emulated, can't that be driver implemented for all DX8.1 & lesser cards?
:?Hmmm, http://www.4gamer.net/specials/hl2int/hl2int_e1.html HDR was discussed @ SIGGRAPH 2003 in July, according to that interview w/Valve. Seems Splinter Cell also uses HDR for DX8. Hmmm.
Edit: correction on the demo & date.
Freak'n Big Panda
Regular
Wasn't mshdribl originally a DX9 only demo then a DX8 compatable demo was released some time after the original release? I'm pretty sure that is the case and I'm pretty sure I remember GFFX 5800 ultra onwer running that DX9 demo... if i'm right then why can the GF FX run this demo and not the HDR in HL2?
Well, now this has me concerned for nVidia's future. I'm just a gamer with an above average interest in vidcards, yet it appears for most of this year my grasp of the situation has been better than nVidia's. :?Uttar said:3. The worse is that with all this crap, it seems to me management is actualy believing these e-mails. And that quite simply because, well, they still got to trust their employees... And it's higher-end employees sending e-mails to all employees...
So, what do we have here?
A mess, a complete mess for multiple things, that management is barely aware of.
And no, comparing this to 3DFX ain't original
Somebody at nVidia hasn't been earning their salary lately...
Do you think this hl 2 flack will wake them all up? If this doesn't do it, nothing will.
Thanks for that, but could someone please explain what that means to a thicky like me?DeanoC said:
Would someone please dumb that down a level for me? I'd really like to understand it.
While on the topic of HDR, here is another program that can be used as a benchmark.
http://www.daionet.gr.jp/~masa/rthdribl/index.html
Some results.... http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=18027
Guess which card performs better
http://www.daionet.gr.jp/~masa/rthdribl/index.html
Some results.... http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=18027
Guess which card performs better
Too confusing to judge in there, nelg. Seems like a particularly interesting bench to spread around and let hardware reviewers check out and post about, though.
nelg said:While on the topic of HDR, here is another program that can be used as a benchmark.
http://www.daionet.gr.jp/~masa/rthdribl/index.html
Some results.... http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=18027
Guess which card performs better
I ran this on my Geforce FX 5900 before I requested a return-material authorization for a Radeon 9800 a few weeks ago. There is no need to explicitly run a benchmark - the performance difference is quite evident...
Interestingly enough, the program does not appear to operate correctly on a Geforce FX 5900 (at least not with the 45.23 drivers - depth-of-field does not function, and the glare types do not appear to be applied correctly).
Over 1200 views since I posted my question & no answers?
Ok then, lets try it this way:
Dave,
Wanna explain why you think HDR needs DX9 & PS2.0 when it evidently doesn't or why you feel the nV drivers don't support HDR when many ppl can run both the ms & rt HDR demos w/nV DX9( :? ) cards & drivers? 8)
Thnx,
Ok then, lets try it this way:
Dave,
Wanna explain why you think HDR needs DX9 & PS2.0 when it evidently doesn't or why you feel the nV drivers don't support HDR when many ppl can run both the ms & rt HDR demos w/nV DX9( :? ) cards & drivers? 8)
Thnx,
