PDA

View Full Version : GeForce FX 5950 Ultra vs Radeon 9800XT


Pages : [1] 2 3 4 5

halduemilauno
10-08-03, 01:40 AM
hi.
http://www.anandtech.com/video/showdoc.html?i=1896&p=1
;)

Edge
10-08-03, 02:53 AM
It is important to note that anyone out there running the current 45.23 Detonators will get something along the lines of a 65% performance increase with no loss in image quality for the FX series of cards with the new 52 series of drivers due out late this month.

*waits for people to complain about it anyway*:rolleyes:

Miksu
10-08-03, 03:08 AM
http://www.nvnews.net/vbulletin/showthread.php?threadid=19182

halduemilauno
10-08-03, 03:25 AM
Originally posted by Miksu
http://www.nvnews.net/vbulletin/showthread.php?threadid=19182

hi.
sorry I haven't seen it.
bye bye.

:angel:

Sotos
10-08-03, 03:26 AM
According Anandtech Fx 5950 is the LOSER again!!
compared to Ati latest

But Nvidia has promised that there is a hope with 52.xx
for extra performance and better image quality from 51.xx
Does it??? or they just HYPING again :(

We will have to wait and see

I do not know but every promise that Nvidia makes lately does not make me so excited anymore and the Nvidia "promises result" is lately not so good
Anyway...

WarheadMM
10-08-03, 05:40 AM
wow....it loses with an average 10FPS. whats 10fps in a game anyway

Hanners
10-08-03, 05:54 AM
Originally posted by WarheadMM
wow....it loses with an average 10FPS. whats 10fps in a game anyway

That depends.

The difference between 30FPS and 20FPS, or 60FPS and 50FPS is a big deal - The difference between 100FPS and 90FPS isn't.

halduemilauno
10-08-03, 05:59 AM
Originally posted by Hanners
That depends.

The difference between 30FPS and 20FPS, or 60FPS and 50FPS is a big deal - The difference between 100FPS and 90FPS isn't.


:thumbsup:


but...seriously...(asking to all) when will the board go out?
thanks.

:jumping:

Ruined
10-08-03, 06:18 AM
If Anand's preview turns out to be the case across the board, which looks like a very real possibility, I must admit that Nvidia has astounded me once again. They took a card which was assumed to be "doomed," and made its performance on the level with the competition. Kudos to the drivers team if the final drivers are this good.

WarheadMM
10-08-03, 06:22 AM
Originally posted by Hanners
That depends.

The difference between 30FPS and 20FPS, or 60FPS and 50FPS is a big deal - The difference between 100FPS and 90FPS isn't.

Umm.....50 and 60fps diff is really nothing. 30 and 20 is big but 40 to 50 or 50 to 60 or 60 to 70 is really nothing

Hanners
10-08-03, 06:29 AM
Originally posted by WarheadMM
Umm.....50 and 60fps diff is really nothing.

It depends on the game, but sinking below 60FPs in someting like UT2003 is not much fun, the difference is still pretty clear.

CliveMilk
10-08-03, 06:50 AM
I have a GF4 4800SE ( aka 4200ti aka GF 4400 ) and was wondering whether to get a 9600Pro for Hl2, but as HL2 has been unfortunately delayed I think it might be wise to wait to see how nv40 and R420 turn out.

Another slightly worrying thing is the low frame rates being shown for the latest crop of games, after getting used to 300+ for Quake Arena these figures showing under 60 mean that you don't have much slack to play with (And he used an AMD FX51). Adding together the cpu and video card is not going to be cheap.

It's going to be difficult knowing where to spend the money on if there is not enough for both.

Sean P.
10-08-03, 10:44 AM
Not an easy question to answer given the info you've supplied. Given how well the game scales, you can esily run HL2 on your Ti4800...The deciding factor here is what quality level you want to run the game. If you want a full-DX9 eye-candy experience, you'll need to upgrade to something like the 9600 Pro ( or the new refresh coming ). Also, don't count out NV36...that might have a surprise or two...

Also, think about Doom3 as well as there are some interesting "early" comparisons which seem to illustrate a flipped-situation to the one seen in HL2...

Nobody ever said this stuff would be easy... :D

kruiser
10-08-03, 11:55 AM
actually it's reel easy for most people on these boards

digitalwanderer
10-08-03, 12:01 PM
Originally posted by Sean P.
Also, think about Doom3 as well as there are some interesting "early" comparisons which seem to illustrate a flipped-situation to the one seen in HL2...
The Dig smacks Pelly in the back of the head with a large halibut!

We do NOT point at the sham that was [H]'s D3 "bench" here, at least not in any serious terms to try and make performance comparisons.

It seems your de-programming hasn't all took yet.... :eek2:

Sean P.
10-08-03, 04:23 PM
LoL!

I'll keep my opinions and thoughts to myself on this...much safer that way... :D

Just keep in mind that the graphics card industry rivals any soap opera out there in terms of politics, drama, and controversy.

:fanboy: :D

SuLinUX
10-08-03, 08:13 PM
1. The 8xAF is much better in that shot of F1 challenge on the FX, I dont see them mension that.

2. The Radeon dont render the image correctly on the Gunmetal IQ test. again they dont mension that.

So you see even the Radeon has issue but we dont assume it's driver cheating.

ATI_Dude
10-08-03, 09:13 PM
Originally posted by SuLinUX
1. The 8xAF is much better in that shot of F1 challenge on the FX, I dont see them mension that.

I honestly can't see much difference. Sure the black line in front of the car is slightly more prominent on the FX cards, but the 52.14 driver hasn't got a smooth mipmap transition. You have to remember that ATI an nVidia handles AF very differentlly which sometimes yield different results.

Originally posted by SuLinUX
2. The Radeon dont render the image correctly on the Gunmetal IQ test. again they dont mension that.

Perhaps because they didn't notice any errors. I'm not much into Gunmetal, but recently nVidia seem to have had more problems with it than ATI.

Originally posted by SuLinUX
So you see even the Radeon has issue but we dont assume it's driver cheating.

This isn't cheating, but different approaches to AF and rendering errors.

But to be fair I did notice one difference in IQ in one of tests which, to my surprise, wasn't mentioned by the author. If you examine the Jedi Academy screenshots you'll notice that the there's a prominent glow around the lightsaber on the nVidia cards (both drivers) which isn't there on the ATI card. I've tested this on my ATI 9800 Pro and I cannot see any IQ difference with Glow turned on or off (besides the frame rate drop) so I believe this feature isn't working correctly. I suspect the Glow effect is partly rendered via the nVidia's GL_NV_point_sprite extension which ATI no longer support.

Behemoth
10-08-03, 09:31 PM
Originally posted by SuLinUX
1. The 8xAF is much better in that shot of F1 challenge on the FX, I dont see them mension that.

the F1 Challenge '99-'02 IQ 4xAA/8xAF shots look worlds better on the radeon, thanks to the superior AA.

Originally posted by SuLinUX

2. The Radeon dont render the image correctly on the Gunmetal IQ test. again they dont mension that.

So you see even the Radeon has issue but we dont assume it's driver cheating.
you just assumed radeon didnt render the image correctly. the truth probably is, again, 2 cards handle dx calls slightly differently.

SuLinUX
10-08-03, 09:32 PM
The AF on the FX is better on the F1 challenge test, it draws the yellow line in the distance, yet AF supposed to be superior on the Radeon.

You know one thing people overlooked is how much better the FX5900 is over the previous generation, I mean the AF performance and quality is miles ahead of the GF4 Ti.

I mean jesus give nvidia some credit for vastly improving there FX line over the GF4 and it seems the new dets are much better and even giving at one point a 65% boost for christ sake.

ATI_Dude
10-08-03, 09:39 PM
Originally posted by SuLinUX
The AF on the FX is better on the F1 challenge test, it draws the yellow line in the distance, yet AF supposed to be superior on the Radeon.

Like I said the difference is neglible. I like the 45.23 screenshot better than the 52.14 since the latter has some mipmap issues.

Actually the GeForceFX is supposed to handle AF better than the Radeon since it renders AF for the entire scene while the Radeon use an entirely different approach where AF is only applied to parts of the scene.

But take a look at the Jedi Academy screenshots and let me know what you think.

SuLinUX
10-08-03, 09:39 PM
Originally posted by Behemoth
the F1 Challenge '99-'02 IQ 4xAA/8xAF shots look worlds better on the radeon, thanks to the superior AA.


you just assumed radeon didnt render the image correctly. the truth probably is, again, 2 cards handle dx calls slightly differently.

The parts of grass are black and the hills are not the same colour at all in that shot.

Also like I said above the Radeon's AF does not draw(or blurs off) the yellow line in the distance, it's not a polygon it's a part of the track texture so the AF is not as sharp on the Radeon.

SuLinUX
10-08-03, 09:45 PM
Originally posted by ATI_Dude
But take a look at the Jedi Academy screenshots and let me know what you think.

It's obvious that the second glow is missing off the lightsaber on the Radeon shot.

Behemoth
10-08-03, 09:58 PM
Originally posted by SuLinUX
The parts of grass are black and the hills are not the same colour at all in that shot.

Also like I said above the Radeon's AF does not draw(or blurs off) the yellow line in the distance, it's not a polygon it's a part of the track texture so the AF is not as sharp on the Radeon.
it can be due to parts of the game codes dont follow dx specification strictly but nvidia driver tolerate that kind of coding practice. no one knows, except the ati driver teams and the game programmers.

and radeon doesnt draw tons of ugly aliasing edges as well.

theultimo
10-08-03, 10:03 PM
Heh, they're BOTH gonna be good cards, and miles ahead of my ti4200.....it all depends on how much you wanna pay, and what performance you expect.

IMO the 5950 can be better, but the 9800XT with cat 3.8s and Overdrive might just pull a bit farther ahead.