5200 & 5600 benchmarks at ET

[lifeguard1999]

Link

Looks like the 9500 Pro beats the 5200 and 5600, probably because it has 8 pipes. Too bad ATI is discontinuing it.

 

[Glitchny]

Acording to some of those benchmarks the 5200 cant even run some games at a playable framrate at low resolutions (commanche4). If wanted to play games and didnt have much moeny i dont think i would spring for the 5200 with that lackluster performance.

 

[rogue1979]

The 9500 Pro was faster than the 5600 Ultra at stock speeds. However I am wonder how it will fare at overclocked speeds since the 5600 has a .13 micron core vs .15 for the Radeon, and it has faster rated memory chips. I have a Radeon 9500 Pro that hit 350/300, much faster than the 275/270 it comes clocked at, but I am thinking the .13 micron 5600 might hit 450MHz or better.

 

[bunnyfubbles]

Originally posted by: rogue1979
The 9500 Pro was faster than the 5600 Ultra at stock speeds. However I am wonder how it will fare at overclocked speeds since the 5600 has a .13 micron core vs .15 for the Radeon, and it has faster rated memory chips. I have a Radeon 9500 Pro that hit 350/300, much faster than the 275/270 it comes clocked at, but I am thinking the .13 micron 5600 might hit 450MHz or better.

Is there any reason to believe the 5600/Ultra will be able to hit 400+MHz considering the 5800 Ultra @ 500MHz requires an insane cooling solution? In that sense you could probably get 400MHz+ out of a 9500 Pro should you use cooling that's more agressive than a simple quiet fan and small heat sink.

I dunno, a 5600 Ultra is stated to be clocked at 350MHz while the 9600 is to be around 400MHz, I just don't think nVidia was wise on striking at .13 as quickly as they did.

I don't have much doubt that a 9500 Pro will be superior to a 5600 Ultra when there are cases where the 9500 Pro can keep up with a GF FX 5800 in some cases, especially so once you crank up the resoluion and/or FSAA. In terms of AF, you only need to crank that up to bring the FX to it's knees, with the 9500 Pro keeping up with a 5800 even at 1024x764.

the 9500/9700/9800 might be on a .15 micron process, but do not forget that they are on a very mature .15 process.

 

[Orbius]

The 5200 ultra is a $149 card and in most non-FSAA situations especially hi-res it gets clobbered by the Ti4200. Nvidia really screwed up across the board this time.

 

[Rand]

3DGPU has some overclocked results of the 5600U.... they got it to 410/407.5 (815) up from the stock 350/350.
Linky

Their benchmarks leave much to be desired however.

 

[BenSkywalker]

Is there any reason to believe the 5600/Ultra will be able to hit 400+MHz considering the 5800 Ultra @ 500MHz requires an insane cooling solution?

It has ~40Million less transistors then the 5800, so it should be able to OC considerably better.

The 5200 ultra is a $149 card and in most non-FSAA situations especially hi-res it gets clobbered by the Ti4200. Nvidia really screwed up across the board this time.

The 128MB 5200Ultra has a MSRP of $149, I would expect the 64MB to undercut that by a decent margin, likely being well within the range of the 9200 with the 128MB part falling within the price range of the 9600 non pro which it should be quite competitive with(actually, it could surpass it given the serious memory bandwith limitations of the 9600NP). When looking at the current performance levels, given that this board has a MSRP of $50 less then the Ti4200 and is a replacement for the GF4MX, I don't see what the big issue it. It bests the MX boards in some cases by x00% and has a MSRP that is in line with it. Replacing a part with another that is an enormous amount faster has a significantly expanded feature set and costs the same is screwing up? Would it have been better if they took a Ti4200, cut off half of its pixel pipes and clocked it slightly higher from your perspective?