VOGONS


GPU Duels of 2000 - 2005 Video Cards

Topic actions

First post, by havli

User metadata
Rank Oldbie
Rank
Oldbie

This test is work in progress since 2012. And as the methodology is perhaps too wide and time consuming (10+ hours to test one card)... it is time to reconsider my goals with this project. To be honest I believe not many people are willing to read long article similar to this http://hw-museum.cz/article/2/benchmark-vga-1 … 2011-edition-/1 But in this case it would be even larger - around twice as many numbers. 😵 Of course now I have tools to make nice interactive chars even with frame-times detailed information and such...

Well maybe one day I will write the article, but not anytime soon. Today video presentations are getting more and more popular... so I developed an application to create the GPU Duels. It is more suitable for my pci-e GPUs benchmarking project, but it should work here as well. Here is how such GPU Duel looks https://www.youtube.com/watch?v=HRQzPmFBnNE

Unfortunately in this case I won't be able to measure power consumption - I have only one PSU that provides this information, the Corsair RM1000i... which is located in the pci-e test rig and will remain there. Also I'm not sure how MSI Afterburner will work on GPUs like GeForce 4, Radeon 9000 series and such. Everything except fps and frametimes is measured by AB. Well, even if only fps and frametimes would remain, I think it is still nice and quite unique. Not many people bench old HW with frametimes analysis.

The test system:

Pentium E5700 @ 3733 MHz (14*266)
AsRock 4coreDualSata2
2GB DDR2 533 MHz CL4-4-4-12 CR1
Adaptec 2940UW + 80 GB SCSI HDD
Win XP SP3

And the list of possible combinations. These are just a quick suggestion.
Feel free to propose some more if you like. I have almost every GPU of this time frame, take a look at my website - what is there, I can bench. With one exception - Matrox Parhelia. That one should be here somewhere but I haven't seen it in years. 🤣

R9700 Pro vs FX 5800 Ultra R9800 XT vs GF 6600 R8500 vs GF3 TI500 R7500 vs GF2 Ultra GF4 Ti4600 vs Chrome S25 GF4 Ti4600 vs GF 6 […]
Show full quote

R9700 Pro vs FX 5800 Ultra
R9800 XT vs GF 6600
R8500 vs GF3 TI500
R7500 vs GF2 Ultra
GF4 Ti4600 vs Chrome S25
GF4 Ti4600 vs GF 6200
V4 4500 vs GF 256 SDR
V5 5500 vs GF2 GTS
GF4 Ti4600 vs FX 5800 Ultra
Radeon DDR vs GF2 GTS
R9700 Pro vs GF4 Ti4600
GF4 Ti 4200 vs GF3 Ti500
GF2 MX vs Radeon SDR
R8500 LE vs GF4 MX 460
R9000 Pro vs FX 5200
R9000 Pro vs R9500
GF 6200 Unlock vs GF 6600
R9800 XT vs X700 Pro
GF 6800 vs X850 Pro
GF 6800 Ultra vs X800 XT PE
FX 5600 vs R9500
GF 6800 vs GF 6800 GS OC
X1800 XL vs GF 7800 GT
GF 6800 Ultra vs GF 7800 GT
X1300 Pro vs GF 6600
X1300 Pro vs X1600 XT
GF 6800 vs X1600 XT
GF 6800 Ultra vs FX 5800 Ultra
R9800 XT vs X800 XT PE
GF4 Ti 4200 vs GF FX 5900 XT

HW museum.cz - my collection of PC hardware

Reply 1 of 37, by Garrett W

User metadata
Rank Member
Rank
Member

Took a look at the GPU Duel video, I have to say it is astonishing work! If you manage to pull this off or something similar with DX7-8 cards, it will be spectacular! It's no Parhelia 256, but I have a Matrox P650 which is essentially a Parhelia cut in half AFAIR, which I can donate to the cause if you want. Perhaps it'd be interesting to compare against GF2MX or something.

Reply 3 of 37, by havli

User metadata
Rank Oldbie
Rank
Oldbie

Ok, the first video - Radeon X1300 Pro vs S3 Chrome S25.
https://www.youtube.com/watch?v=tnVyFbzpFBo

HW museum.cz - my collection of PC hardware

Reply 7 of 37, by havli

User metadata
Rank Oldbie
Rank
Oldbie

Yeah, stuttering is really strong here. And while Fraps frametimes measuring may not be completely accurate, in this case it matches the actual smoothness well. Also in real-time the stuttering is much worse than on the video, which is strange as it is 60 fps capture, so everything should be there.

@Arctic
All I have is this driver number provided by GPU-Z.
chrome_s25_gpu-z5zjb0.png

HW museum.cz - my collection of PC hardware

Reply 8 of 37, by SPBHM

User metadata
Rank Oldbie
Rank
Oldbie

I noticed the Chrome was close in Doom 3, and not so much in the other two, I wonder if it's because of MSAA4x or it's just that much slower overall.
in any case, it's not looking too bad for such uncommon/unsuccessful card, well, in Doom 3

Reply 9 of 37, by havli

User metadata
Rank Oldbie
Rank
Oldbie

I didn't use MSAA, that is only Anisotropic Filtering, which shouldn't be too performance demanding.

HW museum.cz - my collection of PC hardware

Reply 10 of 37, by appiah4

User metadata
Rank l33t
Rank
l33t

How about a Radeon 7500 vs GeForce 2 Ti battle instead? The 7500 is a refreshed Radeon DDR, the GF2 Ti is a refreshed GF2, both are released as companion cards to 8500/GF3 DX8 cards which got all the attention and comparisons.. Radeon 7500 was a GREAT card at the time that got way too overlooked.

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 11 of 37, by havli

User metadata
Rank Oldbie
Rank
Oldbie

I can do that, no problem. There is not much difference between GF2U and GF2Ti, just the 15% higher memory clock for the Ultra. Will be interesting to see how R7500 can perform.

HW museum.cz - my collection of PC hardware

Reply 12 of 37, by SPBHM

User metadata
Rank Oldbie
Rank
Oldbie
havli wrote:

I didn't use MSAA, that is only Anisotropic Filtering, which shouldn't be too performance demanding.

right, not sure how I read it wrong 😵

in any case, with the x1k I think AF is almost free, I have no clue if it's the same with the Chrome, with older cards like Geforce FX AF was expensive.

Reply 13 of 37, by havli

User metadata
Rank Oldbie
Rank
Oldbie

I don't have the exact numbers for X1300 Pro, but X1600 XT could be similar enough. In Call of Duty (OGL) AF performance hit is very similar for both S25 and X1600. In Far Cry (DX8) S25 loses more performance. AA on S3 doesn't work at all when running OGL and with D3D the performance hit is massive. But in the video only AF 4x is used.

aa-af-speed0ij0t.png

HW museum.cz - my collection of PC hardware

Reply 14 of 37, by havli

User metadata
Rank Oldbie
Rank
Oldbie

Next one is online, this time Radeon 9800 XT vs GeFroce 6600.

https://www.youtube.com/watch?v=1oHcXmNeWmQ

HW museum.cz - my collection of PC hardware

Reply 16 of 37, by SPBHM

User metadata
Rank Oldbie
Rank
Oldbie
Garrett W wrote:

Wow, that's interesting, wonder how much far ahead the 6600 GT is, I was lead to believe for years that the GT variant was actually quite competitive with 9800 Pro cards.

the difference between this 6600 and a 6600gt is actually huge, 300 vs 500MHz core, 550 vs 1000Mhz ram
if you look at the specs you could assume this 6600 is a better match for a 9600XT actually (well it should be a little better than that, but..)

but yes, the 6600 had no hope in this one, still I kind of expected it to be a tiny bit better on D3 but oh well, the 9800XT was a beast for its time, no doubt, as for the 6600 at least the power usage looks good!

Reply 17 of 37, by Garrett W

User metadata
Rank Member
Rank
Member

Oh crap, I had forgotten that. I thought the only difference between non GT and GT was GDDR3 on the GT part which meant immensely increased memory bandwidth. I had completely forgotten that the 6600GT ran at 500MHz.

Reply 19 of 37, by havli

User metadata
Rank Oldbie
Rank
Oldbie

It is time to go more old school 🤣

Voodoo4 4500 AGP vs GeForce 256 SDR
https://www.youtube.com/watch?v=r8SqUYMNtdE

With some effort V4 4500 could have been ready at the end of 1999 (there are fully working prototypes with final clocks and PCB design). However the performance would be too low to counter GF 256... so probably for that reason the first VSA-100 card was the V5 5500 introduced 6 months later.

HW museum.cz - my collection of PC hardware