budget card woes!
polycounter lvl 18
Since the R9800 and R9500 falis in my machine (Athlon 950), I'll get a "budget" card in the $40 range, because I know those'll work in this crap machine I have. Unfortunately there's only two, and they both suck:
Powercolor ATI Radeon 9250 128mb $38
The 9250 only has DX8 support, and everyone talks about radeons having graphics glitches and bad driver stories, also Blender i've heard doesn't work well on Radeons (unless they fixed that, but I lost track) plus you don't get pixel shader 2.0
eVGA GeforceFX 5200 128mb $39
The GFFX has dx9 support, but supposedly has bad performance overall, even many customers saying "DON'T ATTEMPT DOOM3 ON THIS CARD OMG"
Both cards are AGP 4x/8x of course.
I CANT DECIDE
Powercolor ATI Radeon 9250 128mb $38
The 9250 only has DX8 support, and everyone talks about radeons having graphics glitches and bad driver stories, also Blender i've heard doesn't work well on Radeons (unless they fixed that, but I lost track) plus you don't get pixel shader 2.0
eVGA GeforceFX 5200 128mb $39
The GFFX has dx9 support, but supposedly has bad performance overall, even many customers saying "DON'T ATTEMPT DOOM3 ON THIS CARD OMG"
Both cards are AGP 4x/8x of course.
I CANT DECIDE
Replies
Get the 5200.
Otherwise, get the Nvidia... the 5200 pretty much sucks, but anything from Nvidia is better than that Ati crap
anything from Nvidia is better than that Ati crap
[/ QUOTE ]
*insert cruel and degrading remarks about JK's mother here*
Of course why in gods good name would you want to even attempt doom3 on a 950 athlon anyway.
[/ QUOTE ]
It's already playable with the Geforce2 I have, perhaps the gffx makes d3 performance worse? I've also heard horror stories of low framerates in Farcry with the card as well, making it seem like the GF2 beats the GFFX in performance.
The FX 5200 supports DX9 shaders 2.0 and it is really shitty at handling them.
So if you turn off shaders and shadows the card runs great.
I ran StarWars galaxies at max detail fine as long as i had shaders turned off (didnt make the game look any different anyway) and increased the framerate considerably!
It's already playable with the Geforce2 I have, perhaps the gffx makes d3 performance worse? I've also heard horror stories of low framerates in Farcry with the card as well, making it seem like the GF2 beats the GFFX in performance.
[/ QUOTE ]
HELLO wake up, GF2 has no shader support so it doesnt do normal maps etc. so you might aswell play Q3 and it looks the same as D3 with GF2...
Oh wait maybe we could run D3 on 486 in text mode, WEEEE so playable.
Anyways i have Radeon 9600 and it runs D3 fine (high detail at 800x600) and i havnt had any of this "omg ati drivers suck so bad you cant use it atall" crap, actually i think i had more problems with my GF3ti200.
HELLO wake up, GF2 has no shader support so it doesnt do normal maps etc. so you might aswell play Q3 and it looks the same as D3 with GF2...
[/ QUOTE ]
hrm
[ QUOTE ]
HELLO wake up, GF2 has no shader support so it doesnt do normal maps etc. so you might aswell play Q3 and it looks the same as D3 with GF2...
[/ QUOTE ]
hrm
[/ QUOTE ]
8fps. It may handle them, but not well
[ QUOTE ]
hrm
[/ QUOTE ]
8fps. It may handle them, but not well
[/ QUOTE ]
The framerate doesn't get faster if you disable bumpmaps or specular, so they're not the problem, just the stupid stencil shadows (stencil shadows eat fill-rate no doubt), note i'm playing in a hi-res too
Delta Force XTREEMEMEME!!! (Barf, at least it's the demo) runs well online on a 32 player server and just about everyone is bumpmapped.
[ QUOTE ]
HELLO wake up, GF2 has no shader support so it doesnt do normal maps etc. so you might aswell play Q3 and it looks the same as D3 with GF2...
[/ QUOTE ]
hrm
[/ QUOTE ]
The D3 i played looked about 100x better
meh anyway I bough a Power Color 9250 256mb PCI a while back for my digital 2d art PC. It's ok, but it's still slow running and of todays flash programs and high poly 3Dsmax. It's decient enough to run Painter and Photoshop. I only mainly bought it so I can run multiple monitors. (I probably blame the outdated PCI slot limiting performance, don't know why I didn't go with AGP instead...)
I did that for roughly 600 Euros, you can probably get away with less by not buying that expensive a graphics cards (I got a GF6800 when they were new, I just like how much mileage you can get out of an expensive card, no need to adjust to new cards for a few years). After all, there's no need to replace the drives, case, peripherials or monitor. I think in some custom vs. Dell thread here we managed to build a very good gaming machine for 400-500$.
Is it the shader support that sucks, forcing you have to turn it off and take huge step backwards to the Geforce2-level GFX? What about the Radeon 9250, how is it's shader performance, even though it only supports pixel shader 1.1?
And no, i'm not getting a new PSU and CPU. My Athlon works great and i'm not getting a new CPU/mobo/case till that breaks. Besides, wouldn't all these newer cards be bottlenecked by the 950?
I've only made up this thread to hear opinions on which card is obviously better for a polycounter, and most point to the GFFX5200, so i'll just get that then.
Oh and finally, I don't plan to run many "next-generation" games - so yeah, not really looking forward to playing Fear anytime soon.
TOO LATE I ORDERED HAHAHAHA
The problem with the low-end GeForce and Doom 3 is actually the Fillrate. Budget GeForce FX/6 have horrible small fillrates, so they can't run shaders very well.
That's where the majority of Doom 3's speed goes. It's also quite a damn sight faster than the budget Radeon. So again I wouldn't worry about the whole 'speed difference'
Personally I would've gone for the 6200 AGP8x or 9600XT as they both run one hell of a lot quicker at only a slightly higher cost, like 2-3x the speed.
Something I don't like about the 5200 is it's vertex throughput.
5200 - ~63 Million / Second
9200 - ~80 Million / Second
6200 - ~260 Million / Second
You won't notice much of a difference in Doom 3, but in Half-Life 2 the speed will seriously improve.