PDA

View Full Version : GeForce FX sweet spot - 1600x1200 with 2X AA/8X AF?


Nv40
03-10-03, 02:37 AM
wow! take a look a this review..


http://www.theregister.co.uk/content/54/29606.html


that site shows the Geforcefx(ultra) is more faster
than the Radeon9800pro at raw performance and
2xaa/8x AF.. in UNREALT2003,SS2,AQUANOX,and
COdecreatures.

using DET42.72 drivers...

now i understand why people dont trust in any review anymore,
LOL! :D IMO , the diferences between WEb sites reviews
of the radeon9800pro in benchamrks between TOms ,ANAD, ,hardOCP and now the register are as greater as the diference between directx8 and directx9..

i knew that the GeforceFx is a fast card in raw performance
and with high Aniso , but not that fast . :D
INTEL the way its mean played.. :)

BTW. to be fair ,i still think that ATI radeon9800pro
have the edge in performance when using 1600x +AA/AF
to its maximun levels possible ,but still is interesting to see how top cards from Nvidia/ATI perform at more realistic playable settings :)

AngelGraves13
03-10-03, 02:57 AM
it's not about speed. Do you need the fastest car?? no, I'll take features and quality over a lambo any day. you get more leg room too! Anways, instead of hearing people complain like little babies in the 4th grade "mine is better than yours!!" why don't they run a few "new" games with the card and then talk. For those that have seen it running, you know what I'm talking about.

ChrisW
03-10-03, 03:07 AM
LOL! That review is a joke. Everyone knows 2xfsaa does very little on the GFFX. He also fails to state exactly what aniso settings he used. Looks like he hand picked the benchmarks just to benefit nVidia.

Hellbinder
03-10-03, 03:08 AM
Get Serious...

They bench at 2x AA adn 0 and 8x AF... Its completely Ridiculous. and then to post a conclusion like that??? Plus i have SERIOUS Doubts about the validity of their game settings. As these results dont line up with any of the other numbers people are posting. To single out 2x AA and claim the 9800pro is inferior to the FX is just as underhanded as it comes. I mean. They HAVE to know that at 4x+8x the 9800pro is 30 FPS faster than the Fx right?? How could you not even mention that? Further i know that the 9800pro does not lose to the FX at 2x+8x in the first place. There is simply NO other explanation that they are either PAID OFF by Nvidia. Or have some personal Agenda. This BS is INEXCUSABLE!!!!

Look at these Rediculous charts...
http://www.theregister.co.uk/media/1015.jpg

http://www.theregister.co.uk/media/1017.jpg

http://www.theregister.co.uk/media/1016.jpg

Can you Fing BELIEVE these underhanded Pieces of ****??? Seriously. Its CRIMINAL.

Look at anandtechs Serious sam numbers
http://www.anandtech.com/video/showdoc.html?i=1794&p=8

Give me a break..

Nv40
03-10-03, 03:08 AM
Originally posted by 99 to Life
it's not about speed. Do you need the fastest car?? no, I'll take features and quality over a lambo any day. you get more leg room too! Anways, instead of hearing people complain like little babies in the 4th grade "mine is better than yours!!" why don't they run a few "new" games with the card and then talk. For those that have seen it running, you know what I'm talking about.

indeed.
all ATI and Nvidia top cards goes beyond 300 frames p/s
in quake3,and perform really good in most Directx7 games .
what we need is new games bechmarks ,
to really undertand the future potential of each card
in latest directx8 games and directx9. :)

Lezmaka
03-10-03, 03:14 AM
What we need is reviews that show some reasonable numbers, not crap that they pulled out of their asses.

Those numbers are sooo messed up it's not even funny. If you actually compare those results to almost any other review, you should see that, unless your BS detector (also known as common sense) is just as messed up.

John Reynolds
03-10-03, 07:04 AM
Originally posted by Nv40
indeed.
all ATI and Nvidia top cards goes beyond 300 frames p/s
in quake3,and perform really good in most Directx7 games .
what we need is new games bechmarks ,
to really undertand the future potential of each card
in latest directx8 games and directx9. :)

This is the first thing I've ever seen you post that makes sense.

However, we do have a forward looking new benchmark to use, but Nvidia is refusing to send out review/preview boards to sites unless they agree to not use it. It's called 3DMark03.

:nono:

Nutty
03-10-03, 07:21 AM
3dmark03 is not forward looking. Its a hack of small amount of dx9 parts bolted onto a crap set of engines that fail to achieve the desired effect with any remoteness of efficiency or similarity to future game engines.

Kruno
03-10-03, 07:22 AM
I don't need numbers, I need features. Ati and nVidia will have to work ultra hard to get me to buy their next chips.

zakelwe
03-10-03, 07:42 AM
Originally posted by John Reynolds
This is the first thing I've ever seen you post that makes sense.

However, we do have a forward looking new benchmark to use, but Nvidia is refusing to send out review/preview boards to sites unless they agree to not use it. It's called 3DMark03.

:nono:

It's interesting to note that the FX does better than the 9700 Pro in the DX7 and DX9 tests but loses out in DX8 compared to the 9700 Pro.

If you ginore AA/AF ( I don't know why you'd want to though ) then the FX does appear best for current DX7 games, if you can afford one and put up with the other limitations of slots used and noise.

Kruno
03-10-03, 07:50 AM
My Voodoo 3 runs DX7 games fine. ;)

Moose
03-10-03, 09:03 AM
I don't think I'm the only one, but, I don't even look at benchmark numbers for anything less than 4xAA with at least 8xAF. This is what I play at with my R9700 and I don't see a reason to ever lower the IQ to anything less. Once you get used to the quality and speed there is no going back. People who are buying $400 video cards don't care if a card gets 200FPS in Quake 3 or 230FPS. It is all about speed at the high quality settings and features.

That said, any review that doesn't use those settings and show IQ screenies for comparison are a complete waste of my time.

PreservedSwine
03-10-03, 11:19 AM
Originally posted by Moose
I don't think I'm the only one, but, I don't even look at benchmark numbers for anything less than 4xAA with at least 8xAF. This is what I play at with my R9700 and I don't see a reason to ever lower the IQ to anything less.

But, to complicate things even further, ATI's 4XFSAA is more comparable to NV30's 6xFSAA.......

NV30's 4XFSAA is more like ATI's 2XFSAA.......

I have heard many times that the Radeon series have better FSAA quality than the GeForce line has. These tests seem to justify those opinions. Both 2x and 4x settings look much smoother on the Radeon 9700, and shows that these settings are not comparable.

From NV305600ULTRA review here (http://www.3dgpu.com/reviews/5600u_4.php)

Moose
03-10-03, 12:21 PM
Originally posted by PreservedSwine
But, to complicate things even further, ATI's 4XFSAA is more comparable to NV30's 6xFSAA.......

NV30's 4XFSAA is more like ATI's 2XFSAA.......

From NV305600ULTRA review here (http://www.3dgpu.com/reviews/5600u_4.php)

That's why a good review must include screenshots. The R9700's 2xAA is almost as good quality wise as the GFFX's 4xAA and ATI's 4xAA is as good if not better than the GFFX's 6xAA.

ATI's 6xAA is untouchable (quality wise) compared to anything that the GFFX can do.

Nv40
03-10-03, 12:56 PM
Originally posted by Moose
That's why a good review must include screenshots. The R9700's 2xAA is almost as good quality wise as the GFFX's 4xAA and ATI's 4xAA is as good if not better than the GFFX's 6xAA.

ATI's 6xAA is untouchable (quality wise) compared to anything that the GFFX can do.

are you really sure?

just look at this screenshots

http://www.hardocp.com/image.html?image=MTA0NjgwOTI1NEdRU0hrZlkyeDRfMTJfM TFfbC5qcGc=

http://www.hardocp.com/image.html?image=MTA0NjgwOTI1NEdRU0hrZlkyeDRfMTJfM TJfbC5qcGc=

and before anyone reply saying that anisoF was
activated in the Geforcefx or the screenshots are wrong
look at what HARDOCP reviewer there have told..
he re-checked ,the screenshots and they show exactly
how both cards looks in NFS2 at 6xaa .
and ANiso in fact was disable on both cards..

http://www.rage3d.com/board/showthread.php?threadid=33670711&perpage=20&pagenumber=3

at least for me a good AA should be able
*not only* remove jagies but also to not blur so much
the textures. and the Geforcefx clearly does much better *overall* Job in NFS2.

John Reynolds
03-10-03, 01:06 PM
Originally posted by Nv40
are you really sure?



Yes, we're really sure. You can keep harking back to one set of screenshots from one game (while of course ignoring other sets of screenshots from the very same review of other games that show a huge disparity between the AA quality) all you want, but I've used both products for hours 'n hours and Nvidia's AA modes are a joke. In fact, the only one worth the bandwidth of actually using is 2x. The rest are just a waste of bandwidth. Better to use 2x and go for higher resolutions.

Nv40
03-10-03, 01:24 PM
i have seen hundreds!! of screenshots of ATI 6x AA
and yes it smooth better the jaggies ,
ofcourse that it shoud do a "better" job , its bluring the
textures . it is more noticeable in NFS2, funny how the
blury thing is the one that most people disliked so much
in NV QUincux , but its now great on ATI. :rolleyes:

John Reynolds
03-10-03, 02:00 PM
Originally posted by Nv40
i have seen hundreds!! of screenshots of ATI 6x AA
and yes it smooth better the jaggies ,
ofcourse that it shoud do a "better" job , its bluring the
textures . it is more noticeable in NFS2, funny how the
blury thing is the one that most people disliked so much
in NV QUincux , but its now great on ATI. :rolleyes:

No, it's not blurring textures. And one set of screenshots do not a case make.

kyleb
03-10-03, 02:17 PM
the fact is that the fx sharpens the textures due to useing multiple samples in the anti-alising method in question. granted the 9700pro has more than enough power to use anisotropic filtering to compate for that while still geting twice the fps:

9700pro (http://www.hardocp.com/image.html?image=MTA0NjgwOTI1NEdRU0hrZlkyeDRfMTJfM TNfbC5qcGc=)

fx5800 (http://www.hardocp.com/image.html? image=MTA0NjgwOTI1NEdRU0hrZlkyeDRfMTJfMTJfbC5qcGc=<br />)

and the 9800pro gets another 5 fps on top of the 9700pro:


9800pro (http://www.hardocp.com/image.html?image=MTA0NjgwOTI1NEdRU0hrZlkyeDRfMTJfM TRfbC5qcGc=)

Moose
03-10-03, 02:49 PM
Originally posted by kyleb
the fact is that the fx sharpens the textures due to useing multiple samples in the anti-alising method in question. granted the 9700pro has more than enough power to use anisotropic filtering to compate for that while still geting twice the fps:

9700pro (http://www.hardocp.com/image.html?image=MTA0NjgwOTI1NEdRU0hrZlkyeDRfMTJfM TNfbC5qcGc=)

fx5800 (http://www.hardocp.com/image.html? image=MTA0NjgwOTI1NEdRU0hrZlkyeDRfMTJfMTJfbC5qcGc=<br />)

and the 9800pro gets another 5 fps on top of the 9700pro:


9800pro (http://www.hardocp.com/image.html?image=MTA0NjgwOTI1NEdRU0hrZlkyeDRfMTJfM TRfbC5qcGc=)

hey those pics look very similar in quality. The only difference I see is the FPS in the upper left corner.

GFFX - 15FPS
R9700 - 30FPS
R9800 - 35FPS

BIG difference there!!! :D

ReDeeMeR
03-10-03, 03:09 PM
Originally posted by Nutty
3dmark03 is not forward looking. Its a hack of small amount of dx9 parts bolted onto a crap set of engines that fail to achieve the desired effect with any remoteness of efficiency or similarity to future game engines.


But then if you would step back and think, in the near future ALL games wuld be DX8-DX9 hybrid, THERE WILL BE NO pure DX9 for another year or two, forget that.