PDA

View Full Version : Official GeForce FX 5200 & 5600 review thread


Pages : [1] 2 3 4 5 6 7 8 9 10 11

Cotita
03-10-03, 10:13 AM
Reviews for the NV31/nv34 are comming up (3dgpu and bjorn3d), and as I expected, their performance is nice.

The thing I really like is the overclocking potencial, with core speeds for the nv31 as high as 400mhz and ram speed of >700.

John Reynolds
03-10-03, 10:26 AM
Actually, I think the performance is very disappointing. Barely faster than a Ti4200, crap AA modes, probably benched with Aggressive or Balanced AF enabled, and tested on very high-end CPUs. Vertex shading is sloooow, too, so testing on slower CPUs could be important (how many 3Ghz CPU owners are going to run out and buy one of these cards?).

Cotita
03-10-03, 10:35 AM
The nv34 does beat the geforce4mx and that speaks well for it.
The thing is that I don't think it can beat the Radeon9200, so its only "marketing" advantage will be DX9 support.

As for the nv31 I also think it will not perform as well as the Radeon9600, altough the radeon9600 is supposed to be slower than the radeon9500.

gravioli
03-10-03, 10:37 AM
I have to disagree that their performance is nice. The Ti4600 spanks (w/out AA/Aniso) or is essentially equal (except 3DMark) to the 5600 Ultra in virtually every test at Bjorn3D (4600 wins 8, NV31 wins 7, they both tie 1). Without AA/Aniso, the Ti4600 beats the overclocked NV31. The 5200 Ultra is better than the MX460, but it doesn't take much to be better than that card. Anyway, they're somewhat nice with AA/Aniso enabled and with overclocking potential, but other than that.....

If the NV31 is cheaper than the Ti4600, which is available for ~180 now in some places, when it comes out, then it will be a great deal.

Slappi
03-10-03, 10:37 AM
The 5600 ultra holds it's own against the 9500pro so it should be right with tthe 9600.

Cotita
03-10-03, 10:40 AM
Originally posted by Slappi
The 5600 ultra holds it's own against the 9500pro so it should be right with tthe 9600.

Don't think so, the 9500pro beats the ti4600 in almost every test I've seen, so the geforcefx5600 won't even touch it.

Slappi
03-10-03, 10:44 AM
Originally posted by Cotita
Don't think so, the 9500pro beats the ti4600 in almost every test I've seen, so the geforcefx5600 won't even touch it.

What are you talking about? Read a decent review. The 5600 ultra beats the 9500pro in about 40% of the benches with 4XAA and 8XAF. That is what counts the AA and AF.

John Reynolds
03-10-03, 10:48 AM
Originally posted by Slappi
What are you talking about? Read a decent review. The 5600 ultra beats the 9500pro in about 40% of the benches with 4XAA and 8XAF. That is what counts the AA and AF.

If this is what counts then bear in mind that the AA modes are not at all comparable at equal samples and that the 5200 and 5600 are both being benched with Aggressive AF enabled (can you say bilinear filtering?).

Take that 5600, run it with 4/8x Application AF and 4x AA and compare its score against a 9500 with 2x AA and 4/8x Quality AF and then see who wins. The 9500 will eat it for breakfast, lunch, and dinner, and it's no one's fault but Nvidia's that they haven't improved their AA in years and that their AF is slooow unless running in bilinear or 2.5linear (aggressive and balanced, respectively) mode.

slickfloater
03-10-03, 10:52 AM
WTF??!! I just looked at extremetech's review of the Geforce
FX line and besides all 5200 and 5600 FX's getting thoroughly thrashed I noticed that in their table of GPU features the FX 5200 has a lower transistor count than the FX 5600 which has a lower count than the mythic FX 5800. What gives?? I thought all Geforce
FX's use the same core and have the same transistor count. It
looks to me like the 5600 and 5200 don't have the same functionality as the FX 5800.

Dazz
03-10-03, 10:59 AM
Originally posted by Slappi
What are you talking about? Read a decent review. The 5600 ultra beats the 9500pro in about 40% of the benches with 4XAA and 8XAF. That is what counts the AA and AF. You should note that was in benchmarks that are highly optimised for NVIDIA cards, SS SE for example. Everyone knows ATi can't even touch NVIDIA in that game. Me I was greatly disappointed and think I will just stick with my Ti4400. I thought the Ti5600 was pathetic in Comanche 4 where it got completly trashed by the Radeon 9500Pro; hell it's doesn't even seem playable on that card! Iíve got a bad feeling about the 9600Pro now :(

Mainstream my @ss!!!
It SU><S big

Slappi
03-10-03, 11:01 AM
Originally posted by slickfloater
WTF??!! I just looked at extremetech's review of the Geforce
FX line and besides all 5200 and 5600 FX's getting thoroughly thrashed I noticed that in their table of GPU features the FX 5200 has a lower transistor count than the FX 5600 which has a lower count than the mythic FX 5800. What gives?? I thought all Geforce
FX's use the same core and have the same transistor count. It
looks to me like the 5600 and 5200 don't have the same functionality as the FX 5800.


You need to read that review again. Why would you lie like this? The 5600 does not get trounced. It loses to the 9500pro in about 60% of the benches and beats it at the extremes AA and AF in about 40%. Yah that is a real trouncing.

SilentSnipeR
03-10-03, 11:18 AM
http://www.extremetech.com/article2/0,3973,922652,00.asp

Probably the best review atm.

digitalwanderer
03-10-03, 11:20 AM
I heard that ExtremeTech compared these to the 9500 Pro (http://www.extremetech.com/article2/0,3973,922652,00.asp), but I'm only just now starting to read it.

Thought it might be helpful. See, no flamage. :)

Myrmecophagavir
03-10-03, 11:25 AM
Originally posted by slickfloater
I thought all GeforceFX's use the same core and have the same transistor count.Where have you been hiding? There are 3 different cores so far, NV30, NV31 & NV34! You are joking, right?

Dazz
03-10-03, 11:30 AM
I thought that was the best review myself, after all the Ti5600Ultra is targeting the Radeon 9500Pro. BUt they used alot more tests which gives a better picture :)

slickfloater
03-10-03, 11:41 AM
I thought the different cards would just have different core clocks
and memory speeds. Now I know that my 2 yr old Geforce 3 has a feature (and a rather important one) that the Gefarce FX 5200 doesn't, that being lossless Z-compression. It appears that the FX 5600 has 45 million transistors less than the mythic FX 5800
and doesn't even have 2 TMU's per pipeline (of course we still can't figure out if the non-existent FX 5800 has 4 or 8 pixel pipelines). Anyway, all Geforce FX's (let's not even mention the 5800) get their ass kicked by a mid-range 9500 pro ATI part so nVidia really has their work cut out for them.

jbirney
03-10-03, 11:43 AM
Originally posted by Slappi
What are you talking about? Read a decent review. The 5600 ultra beats the 9500pro in about 40% of the benches with 4XAA and 8XAF. That is what counts the AA and AF.

And just what reviews are you reading? After seeing the one at extrem tech it begs to differ with you point...

SilentSnipeR
03-10-03, 11:45 AM
Originally posted by Slappi
What are you talking about? Read a decent review. The 5600 ultra beats the 9500pro in about 40% of the benches with 4XAA and 8XAF. That is what counts the AA and AF.

wtf have you been smoking/snortting/injecting? cause whatever it is, you need to get some rest and read the reviews again.

digitalwanderer
03-10-03, 12:12 PM
Originally posted by Slappi
What are you talking about? Read a decent review. The 5600 ultra beats the 9500pro in about 40% of the benches with 4XAA and 8XAF. That is what counts the AA and AF.

Methinks you're misreading them graphs and getting your results confused....either that or you're reading Peter Glaskowskie's review... :p

silence
03-10-03, 12:26 PM
i read few first reviews and it seems it isn't too bad.
i would like to get more info on 5600 ultra by gainward with 256 ram if any1 has any.

sebazve
03-10-03, 12:52 PM
muhahhaahahahahah:D

sebazve
03-10-03, 12:54 PM
they look ok although their ps/vs power sucks....big time:rolleyes:

MuFu
03-10-03, 01:02 PM
I think considering how much they cost to manufacture (5200 Ultra especially) they are BEYOND CRAP. I doubt nVidia are in the slightest bit satisfied with them, moreover relieved that they have been able to bump speed up to reletively acceptable levels.

From an EE point of view they are horrible, horrible products and it is a victory of marketing/management (yet again!) that they are available at fairly reasonable prices. "Aggressive" aniso is a plague and their FSAA quality still sucks balls.

That is all I have to say. :)

MuFu.

sebazve
03-10-03, 01:03 PM
WHY THE **** THE NON ULTRA COST A LOT LESS WHILE THE ULTRA WHICH JUST AN OVERCLOCKED NON ULTRA COST TWICE THE PRICE?

im talkin about the 5200

c4c
03-10-03, 01:06 PM
Originally posted by slickfloater
I thought the different cards would just have different core clocks
and memory speeds. Now I know that my 2 yr old Geforce 3 has a feature (and a rather important one) that the Gefarce FX 5200 doesn't, that being lossless Z-compression. It appears that the FX 5600 has 45 million transistors less than the mythic FX 5800
and doesn't even have 2 TMU's per pipeline (of course we still can't figure out if the non-existent FX 5800 has 4 or 8 pixel pipelines). Anyway, all Geforce FX's (let's not even mention the 5800) get their ass kicked by a mid-range 9500 pro ATI part so nVidia really has their work cut out for them.

How is it you have enough knowledge to even look for a feature like z-compression, yet you don't know about the different cores for nv34, nv31, and nv30?

Also, the 9700 and 9500 do not have 2 TMUs per pipe.

Aaaaand :) the extremetech review also mentions that the 9500 is being discontinued in place of the 9600 (which will also be a 4x1 pipeline part just like the 5200 and 5600). Apparently the 9500 costs Ati too much to make.

The moral of the story is: Buy a 9500 before theyre all gone :o