PDA

View Full Version : GeForce FX 5900 not much better than GeForce 4 Ti4600


Pantherman
01-18-04, 09:21 PM
I recently bought a BFG GeForceFX 5900 nu for one of my systems. This is the card everyone's been buying because of its low price. However, it does not seem much better than my old Ti4600 in most of my games. I used to get 12344 in 3DMark 2001 with my old Ti 4600. Now I get 12403 with the 5900. The 5900 is better with AA and AF as well as in 3DMark 2003. However, most of my games seem to play the same without AA and AF. This is not a bad card for the price, but it's certainly disappointing.

I thought my system's CPU might be holding the 5900 back. I know it's definitely on the slow side with the 400MHz bus. Thanks for the response.

S.I.N
01-18-04, 10:53 PM
Finally someone in the same boat as me. No one will fully read or care what we say but who cares.

It's not your CPU..it's just not a big step up at all IMO. I went from a Ti4400 (oc'ed to near 4600 speeds) to 5900NU and did not see a big jump at all and infact some of my games ran worse (jaggyness was so much more apparent,......). I read all the reviews and took into account everyones talk here...but I still did not see a big difference to warrant the $200 investment. My 3DMark score was higher....but who the hell plays 3DMark for hours of enjoyment? I defragged, clean installed in case it was my system and still. I dont have a crappy system so I know it was not a bottleneck anywhere. I guess would have made a difference if I ran AA or AF but I was really dissapointed. I surely wont take into account anymore what those who have never played on these cards and go only on benchmarks say.

I felt so dissapointed I sold the card here for $150 almost brand new. No difference in image quality and yes a few games such as BF1942 ran better but most I did not see a big difference at all, Max Payne 2....Raven Shield....NOTHING. Either my system is masterfully tweaked that I can run almost every game now at 1280x1024 most times maxed our or alot of people here over rate the power of these cards over the last generation(9500-9700Pro-Ti4200-4600...58000).:confused:

bkswaney
01-19-04, 12:26 AM
Sure the Ti4600 is still a very fast card. "In DX8 games."

Why would you not use AA/AF owning a 5900?? :confused:
The more DX9 games come out the Ti owners will be left behind.
Sure they will play the game. But not the way it was meant to be played. :)

Pantherman
01-19-04, 12:30 AM
I agree that the biggest advantage of a GeForce FX 5900 or 5900 Ultra over a Ti4600 is that the 5900 fully supports DX 9 and is better with AA and AF. In today's games with normal settings, there isn't a huge difference.

fivefeet8
01-19-04, 05:26 AM
Originally posted by Pantherman
I agree that the biggest advantage of a GeForce FX 5900 or 5900 Ultra over a Ti4600 is that the 5900 fully supports DX 9 and is better with AA and AF. In today's games with normal settings, there isn't a huge difference.

You got the FX5900 to play with normal settings? I play my games at the highest settings with AA/AF and it runs way better on my FX5900u then my Ti4600 at those settings. In this case, it's a lot better than my Ti4600 was.

But, yeah, if you play at "Normal" settings without AA/AF then why'd you get the FX5900 in the first place? Your Geforce4 ti would have sufficed.

Johnmcl7
01-19-04, 05:44 AM
12403 sounds far too low for 3Dmark2001, even with your 400Mhz FSB P4. What's your score for 3Dmark2003?

John

DoS
01-19-04, 05:48 AM
I felt so dissapointed I sold the card here for $150 almost brand new. No difference in image quality and yes a few games such as BF1942 ran better but most I did not see a big difference at all, Max Payne 2....Raven Shield....NOTHING. Either my system is masterfully tweaked that I can run almost every game now at 1280x1024 most times maxed our or alot of people here over rate the power of these cards over the last generation(9500-9700Pro-Ti4200-4600...58000).

I am sorry but something is seriously wrong with your assertion. I swapped my 128 MB Ti4200 (which was oced ~ 4600) with a 9800 Pro, which i then replaced with an FX 5900 256 (due to bad luck with ATi - 2 RMAs) flashed to 5959 Ultra and the difference was huge. All of my games run smoother and i could use 4xAA with 8xAF and have smoother gameplay than the Ti. IMO the image quality differences are huge and if i play any game with AA and ANISO there is simply no turning back. During the second RMA of the 9800, i had to game with my old Ti and well, it honestly felt like gaming on a Geforce2 MX. I am not even going to bother mentioning pixel shaders and DX9 and even though there aren't any actual games yet (i mean good games) there are some around the corner. And what did you expect, run without AA/Aniso at 1024x768 and see the difference between 70 fps and 140 :rolleyes:

Gator
01-19-04, 07:38 AM
So far as I know, without AA/AF enabled the FX5900 is about the same as a TI4600 in DX8 games. A notable exception to that rule may be a game like UT2k3 which has a lot of details, so try that on for comparison. But otherwise, you're really not gonna appreciate the FX5900 until DX9 games come out or you turn up that AA/AF. That's why 3dmark2k3 is so much higher, it's pretty much the ONLY DX9 app available right now.

That's why I'm not upgrading from my GF4TI yet, I wanna wait for the new games to come out first like Doom3 and Stalker. At that time I'll grab a 5700U or 5900SE for a ridiculously low price hopefully because new hardware may be out by that time ;)

cthellis
01-19-04, 08:54 AM
The point of getting new cards is TO play with quality features, and to play new games with newer features (DX9, big-ass shaders, et al), p'raps better resolutions... If you're just going to use the card the same way as ever, on older games and benchmarks not made for made to take advantage of newer and better features, then yes--you're not going to see a heckuva lot of difference.

But hey, before you got your 4600 what was your card? Did you play the same way as ever and just enjoy more frames, or did you kick up resolutions, textures, features...? Same thing here.

DMA
01-19-04, 11:02 AM
Originally posted by cthellis
The point of getting new cards is TO play with quality features

The scary part is that most people buying these cards never touch the settings for AA and AF etc. They just leave it at "app pref".
Now thats a waste of money. :)

R300 set a new standard back in 2002 and gaming is more fun than ever. :cool:

cthellis
01-19-04, 11:09 AM
Well, it's much better to enjoy an increase from 110 fps to 140 fps, right? :p

Maverickman
01-19-04, 11:16 AM
I can get 5130 in 3DMark 2003. That's a lot better than the 1945 I was getting with the GeForce 4 Ti 4600. Of course, 3Dmark 2003 is a DX 9 benchmark. I like to use 2X AA and 4XAF in games. With those settings, the 5900 far surpasses the Ti 4600. Like I said, the 5900 and Ti 4600 are basically the same in DX 8 games without AA and AF. That's disappointing because you would think that a newer high-end card would be a lot faster.

DoS
01-19-04, 11:34 AM
Originally posted by Maverickman
Like I said, the 5900 and Ti 4600 are basically the same in DX 8 games without AA and AF. That's disappointing because you would think that a newer high-end card would be a lot faster.

Excuse me ? Could you please ellaborate further on the matter ? In which game you expected to see a difference but it turned out the FX5900 has the same speed with the Ti card ? Have you ever thought that there might be other reasons you are not seeing any differences as for example your CPU speed ? Or that maybe you are playing very CPU-limited games ? Or that the games you can't see any difference do not require a lot of GPU power so your Ti was providing more than enough FPS ? You won't be able to tell the difference between 90 fps and 150...
I am sorry if i sounded ironic, but what you are saying sounds like you are bitching about your new sports car which is twice as fast as your old one, just because you can't speed in residential areas...

i hope you all got the idea :D

Zod
01-19-04, 11:54 AM
They are not the same! I have a 5900U and a Ti4600 and the 5900U is much faster. In Unreal 2, UT2K3, RTCW, I can run higher resolutions and kick up the AA and AF modes. As for DX-IW and Halo, running them on the Ti4600 is painful (or rather even more painful than on the 5900U).

mrsabidji
01-19-04, 12:32 PM
I had a 4200 before. There IS a difference in speed. Come on, the memory is faster, the core is clocked higher (and has some features the nv25 hasn't by the way)... If some of your games ran at the same speed with the 4600 as they do with the 5900, then your CPU might be the bottleneck.

mrsabidji

S.I.N
01-19-04, 02:06 PM
I have a 2.53 so its not a CPU problem. And I am not saying it did not run fast...It was just not worth the $200 upgrade "NOW" IMO. I am sure it would be killer in DX.9 games but most of my games now did not benefit much from this upgrade or was not worth the $200 boost. I am not at all saying the card was crap and did not see any difference at all. I saw great differences in Black Hawk Dawn, BF1942 and America's Army fully maxed out at 1280x1024 with EAX on, but those out of my collection of some 30 games where the only notable exceptions.

Games on the Quake engine I saw virtually no difference.....and I guess this is why I am saying it was not worth it, because shooters on this engine is what I mostly play where FPS average 90 and over on the Ti. Call of Duty...RVS....ran just as good on my Ti with eye candy galore quality settings on. When HLF2, Doom III, Far Cry and BF Vietnam is comes out I will be upgrading to the next generation....nothing IMO is worth it now unless your card is running everything like dog poo.

fivefeet8
01-19-04, 02:22 PM
Hope you guys realize that the FX5900 NU's can be bios modded into a FX5950 ultra. Now there's a really good deal. $200 for the performance of a $400 dollar card. :p

Edge
01-19-04, 04:33 PM
Well if you're only playing at 1024x768 or so without AA or Anisotropic, you aren't putting a very big stress on the card at all. In this case, it IS CPU limited, simply because your videocard doesn't have much work to do. However, if you play at 1600x1200, there would be a very noticable difference, and with AA/Aniso there would be a HUGE difference. Even on my TI4200, with 2xAA and lvl8 aniso, I can get almost 100fps in many games. Now if I had an FX5900, in those games, at those settings, I probably wouldn't see ANY improvement. But even if you don't use AA/Aniso now, newer games like Doom 3 and HL2 will probably show a very noticable performence difference between a TI4600 and an FX5900.

The TI4600 is still a powerful card, and most games out there don't even put a big stress on the videocard, so I think this is simply a case of both cards being too powerful to show the difference between them (particularly in DX7 and DX8 games). But even going by raw clockrates, you can see the FX5900 is a decent amount faster (300/650 vs 400(?)/850), and that will be noticable once games start putting a bigger stress on the videocard.

The Baron
01-19-04, 04:36 PM
Try the Icetomb Demo (tonight, my child, TONIGHT!) on a Ti4600 and then on a 5900--the 5900 will spank the Ti4600, I promise.

Edge
01-19-04, 04:49 PM
Originally posted by S.I.N
Either my system is masterfully tweaked that I can run almost every game now at 1280x1024 most times maxed our or alot of people here over rate the power of these cards over the last generation:confused:

It's hardly difficult to get a TI4600 card to run well at 1280x1024 in most games. As I mentioned in my above post, I can get 100 FPS in many older games with 2xAA and lvl8 anisotropic on my TI4200 card. Of course I wouldn't notice a difference if I got an FX5900: my framerate is already high! But games like Halo or even Prince of Persia would show a big gap between the TI4600 and the FX5900, simply because they put a big stress on the videocard, and with those games you'll be GPU limited even at 1280x1024. But in most games without AA or Aniso at 1024x768, I think with a TI4200 you'd be CPU limited up to around 2 ghz. Only when you get into the 3-4 ghz range would you really notice a difference between a TI4200 or FX5900 using raw settings. Of course, most people got an FX or ATI 9x00 card to use the vastly superior AA and Aniso on it.

Cotita
01-19-04, 06:42 PM
I also tought the 4600 was not much faster than the 5900. And in some circumstances it performs very much the same.

I recently replaced my Visiontek 4600 for a bfg5900nu. The 4600 was not much of an overclocker but the 5900 goes past 5950 speeds with no problem. I can run it at 500/1ghz with no extra cooling.

I can see performance increases in most games, specially at high resolutions and aniso/AA.

with the 4600 I ran Unreal2 at 1024x768, anything higher there was stuttering. With the 5900 now I only play at 1600x1200.

Had I not got a 20" monitor maybe I would not see much of a difference, but now I wouldn't go back to the 4600 for nothing.

For me the upgrade was worth every penny.

Pantherman
01-19-04, 09:27 PM
I think the FX 5900 is the better card, but the increase in performance over the Ti 4600 in games like Elite Force II, XIII, Max Payne 2 and others is not that great unless you're using AA and AF. I would've thought that the 5900 would blow the socks off the Ti 4600 in everything, but that's certainly not the case.

AthlonXP1800
01-19-04, 10:07 PM
Originally posted by Pantherman
I thought my system's CPU might be holding the 5900 back. I know it's definitely on the slow side with the 400MHz bus.

Yep it either your CPU or slow bus, take a look at this 2004 Gamers' Hardware Upgrade Guide (http://www.aceshardware.com/read.jsp?id=60000307), you will see many games' performance are near doubled on Athlon 64 3400+ from Athlon XP 2100+ both with 5900 Ultra.

You see my Athlon XP 1800+ CPU is now 3 years old, it ran games fine before 2004. But now 2004 games like Far Cry, I tried the beta run on highest settings at 1024x768 with 8xAF on multiplayer LAN alone, it performed 18 fps minimum, 25 fps average and 35 fps maximum. It defintely need to upgrade my system, in next few months I probably planning to upgrade to Athlon 64 3700+ with more memory, probably 1.5Gb CAS 2 memory. When upgrade fully, Athlon 64 3700+ likely to perform more than 3 times with extra 1Gb memory that of Athlon XP 1800+, it could score like 64 fps minimum, 90 fps average and 126 fps maximum. ;)

StealthHawk
01-20-04, 06:37 AM
Originally posted by DMA
The scary part is that most people buying these cards never touch the settings for AA and AF etc. They just leave it at "app pref".
Now thats a waste of money. :)

R300 set a new standard back in 2002 and gaming is more fun than ever. :cool:

I know this wasn't your point, but...It's better to set AF from inside each app whenever possible!