Anand must be smoking crack or has his head stuffed into his anus. The reason is that he's not taking into effect things discussed in the past...Pixel Shader 2.0 and Floating point precision and why it is part of DX9 standard. The review that calls this in to play is http://www.anandtech.com/video/showdoc.html?i=1896.
Here's the problems I have with it:
1. Tomb raider didnt have fps graph it had some dodgy frame lost while using PS2.0 graph. IMO, anandtech will mislead a lot of people by using those graphs
2. they used unreleased driver for benchmark(which i remember they said they never will after that GF4 fisco)
3. the images were not captured in a way that truely reflect cards' rendering within game, eg. they capture those dark and outdoor area...Even worse, some of the most important FULL SIZE images had BROKEN links, halo was the one stand out.
4. All the screenshots are identical. I mean every one is like its been produced from the same card. No differences in lighting or anything. Surely there should be at least a difference in lighting on some of the screenshots?
5. the MS Flight Simulator game has been removed???? That was the game where the ATI cards got 125fps and the 5900 got 27fps???? Why remove a game and not test I.Q on it when its one of the games Nvidia suffers most on. Surely we want to see if the ATI cards have lower I.Q in this game ??
6. Nvidia still cannot do Floating Point 2.0 Textures (as noted first line of this page in Anands article). Most people do not know what pixel shading (2.0) and FP is...So here's some info on what they do: The algorithms implemented at the pixel processing stage (where the DX9 Standard includes PS2.0/Floating Points) include bump mapping, shadowing, environment mapping, and lighting (e.g. reflection, fog, etc.)...IN FACT, they (PS2.0/FP) don't really effect AA or AF directly...they use shading techniques (utilizing color pixels) to provide depth of field, textures, bumps, blips, ridges, etc. So, interestingly enough...Nvidia still can't do this even with the 52.14 Dets. What up with that? They no likey the DX9 STANDARD?.
7. Why is Anand even comparing benchmarks if PS 2.0 can't be enabled? DX9 MUST HAVE PS 2.0 enabled to be DX9 at all. I mean, it's like Carl Lewis running a 100 meter dash against a person in a wheelchair. One doesn't even use his legs (or in our case, PS 2.0). It is damn ridiculous. Someone over at Anands should have their head examined. In fact, Nvidia cards shouldn't even be included when speaking of DX9...because they ARE NOT COMPATIBLE WITH THE STANDARD. PERIOD. When Nvidia finally sprouts a pair and steps up to the DX9 plate...then lets talk benchmarks. Until then, they are sacrificing quality in order to perform. I'd feel lied to if I had an FX series card...no matter how hard Nvidia tries to polish it up, it's still a turd...it's only going to get buffed and become a shinier turd. I wonder how long they will have people snowed over in believing that they are DX9 capable. It seems they've been able to for a few months now. WAKE UP EVERYONE....we should all be performance junkies not ATI or NVIDIA junkies...I know I am.
8. The bottom line of things are that Anand is trying to say that (and I quote):
If I were Microsoft, I'd prosecute for altering the DX9.0 code (which Nvidia's drivers do...they 'enable' mixed mode instead of fully rendering PS2.0 stuffs). Anyone that can't see the error of Nvidia needs to first, pry the rock off themselves cause they've been living under it for too long and second, pull their head outta their arses...cuz the facts stick. If I hold up a red colored pen and tell you it's red...but you say it's blue...you come off sounding and looking like a total moron...if you continue to support a video card that doesn't do what they've claimed it could do and doesn't meet the DX9 standard...you come off sounding and looking the same.
You want a true test for Floating Point? Go here and use it. This test ignores Nvidia's drivers' request to enable only partial (12-16bit) FP precision and uses DX9 24bit The source of the download is digit-life; it was an article with John Carmack talking about FP and PS 2.0 and how Nvidia isn't up to microsoft DX9 specs regarding the issue. Good read but a bit technical...you can pick out the simple stuff though as they are good to give in to laymans terms.
Enough is enough people. Nvidia is still lying to us. As of today...Nvidia tech data still states PS 2.0+ compliant with the FX 5900. It isn't. They're lying and the proof is purely evident. If they make it so today that 24-32bit comes into play with a new driver, I'll go ahead and clap for them and say good job guys...you're DX9 compliant now and then I'll go purchase an FX 5950...but until then, we cannot ignore things and pretend that there is no fault here. What they are doing is robbing people by selling a product that is handicapped. It isn't right and they shouldn't be getting away with it.
Here's the problems I have with it:
1. Tomb raider didnt have fps graph it had some dodgy frame lost while using PS2.0 graph. IMO, anandtech will mislead a lot of people by using those graphs
2. they used unreleased driver for benchmark(which i remember they said they never will after that GF4 fisco)
3. the images were not captured in a way that truely reflect cards' rendering within game, eg. they capture those dark and outdoor area...Even worse, some of the most important FULL SIZE images had BROKEN links, halo was the one stand out.
4. All the screenshots are identical. I mean every one is like its been produced from the same card. No differences in lighting or anything. Surely there should be at least a difference in lighting on some of the screenshots?
5. the MS Flight Simulator game has been removed???? That was the game where the ATI cards got 125fps and the 5900 got 27fps???? Why remove a game and not test I.Q on it when its one of the games Nvidia suffers most on. Surely we want to see if the ATI cards have lower I.Q in this game ??
6. Nvidia still cannot do Floating Point 2.0 Textures (as noted first line of this page in Anands article). Most people do not know what pixel shading (2.0) and FP is...So here's some info on what they do: The algorithms implemented at the pixel processing stage (where the DX9 Standard includes PS2.0/Floating Points) include bump mapping, shadowing, environment mapping, and lighting (e.g. reflection, fog, etc.)...IN FACT, they (PS2.0/FP) don't really effect AA or AF directly...they use shading techniques (utilizing color pixels) to provide depth of field, textures, bumps, blips, ridges, etc. So, interestingly enough...Nvidia still can't do this even with the 52.14 Dets. What up with that? They no likey the DX9 STANDARD?.
7. Why is Anand even comparing benchmarks if PS 2.0 can't be enabled? DX9 MUST HAVE PS 2.0 enabled to be DX9 at all. I mean, it's like Carl Lewis running a 100 meter dash against a person in a wheelchair. One doesn't even use his legs (or in our case, PS 2.0). It is damn ridiculous. Someone over at Anands should have their head examined. In fact, Nvidia cards shouldn't even be included when speaking of DX9...because they ARE NOT COMPATIBLE WITH THE STANDARD. PERIOD. When Nvidia finally sprouts a pair and steps up to the DX9 plate...then lets talk benchmarks. Until then, they are sacrificing quality in order to perform. I'd feel lied to if I had an FX series card...no matter how hard Nvidia tries to polish it up, it's still a turd...it's only going to get buffed and become a shinier turd. I wonder how long they will have people snowed over in believing that they are DX9 capable. It seems they've been able to for a few months now. WAKE UP EVERYONE....we should all be performance junkies not ATI or NVIDIA junkies...I know I am.
8. The bottom line of things are that Anand is trying to say that (and I quote):
Impressed with the image quality improvement? Where? There is no improvement period from a PS 2.0/FP precision standpoint...and this is where Nvidia is lacking...this is what is causing all the fuss...yet they've improved? That's like saying, hey, I know you didn't beat Carl Lewis in the 100 meters Mr. Handicap person...but you shaved .5 seconds off your time!! woo hoo! It just doesn't make any sense...Nvidia is still on DX8.1 and IS NOT up to par with the standard.
If I were Microsoft, I'd prosecute for altering the DX9.0 code (which Nvidia's drivers do...they 'enable' mixed mode instead of fully rendering PS2.0 stuffs). Anyone that can't see the error of Nvidia needs to first, pry the rock off themselves cause they've been living under it for too long and second, pull their head outta their arses...cuz the facts stick. If I hold up a red colored pen and tell you it's red...but you say it's blue...you come off sounding and looking like a total moron...if you continue to support a video card that doesn't do what they've claimed it could do and doesn't meet the DX9 standard...you come off sounding and looking the same.
You want a true test for Floating Point? Go here and use it. This test ignores Nvidia's drivers' request to enable only partial (12-16bit) FP precision and uses DX9 24bit The source of the download is digit-life; it was an article with John Carmack talking about FP and PS 2.0 and how Nvidia isn't up to microsoft DX9 specs regarding the issue. Good read but a bit technical...you can pick out the simple stuff though as they are good to give in to laymans terms.
Enough is enough people. Nvidia is still lying to us. As of today...Nvidia tech data still states PS 2.0+ compliant with the FX 5900. It isn't. They're lying and the proof is purely evident. If they make it so today that 24-32bit comes into play with a new driver, I'll go ahead and clap for them and say good job guys...you're DX9 compliant now and then I'll go purchase an FX 5950...but until then, we cannot ignore things and pretend that there is no fault here. What they are doing is robbing people by selling a product that is handicapped. It isn't right and they shouldn't be getting away with it.
Anychance you can edit that for us TKS.
