Half-Life 2 Stress Test Article

Chris D

Administrator
Staff member
Joined
Jul 1, 2003
Messages
11,927
Reaction score
231
Paul Jastrzebski from VR-Zone Hardware has just notified me that he has just posted an article on the performance of the Source engine, based on the results of Valve's benchmark software included in the Counter-Strike: Source BETA.[br]
[br]
It's a very well written article and once again shows that in the realm of Half-Life 2, at least, ATi cards reign supreme. Here's a snip:
The ATI Radeon X800XT Platinum Edition reigns supreme in our first series of Stress Test benchmarks. At all resolutions, the card is able to stay around 25 FPS faster than the NVIDIA GeForce 6800 Ultra. The same is true with the ATI Radeon X800 Pro and the NVIDIA GeForce 6800 GT, with the X800 Pro able to run 10-15 FPS faster than the GeForce 6800 GT.
You can catch the rest of the article here.
 
Can anybody tell me how the ATI Radeon 9700 Pro will be placed on the charts on that site?
 
I assume somewhere between the 9600XT and the 9800PRO which averages out at 80fps but that's not an exact science of course.
 
The ATI Radeon X800XT Platinum Edition costs over five hundred dollars.. my whole PC didnt cost nearly that much.. I couldnt even begin to imagine paying that much for a video card. One would hope that it'd reign supreme.
 
I wish I could have taken part in the CS:S BETA and thus have been able to test my system. :(
Something tells me that my Radeon 9100 won't come through for me.

Anyone know if ATI are still doing the bundle with Half-Life 2 once there is a release date again? Seeing as ATI is the manufacturer of choice this would help me a lot.

NVIDIA doesn't look that bad either after those tests. I was sure that ATI would blow them out of the water.
 
My 6800 can get 85fps at 1024x768 4xAA/8xAF?


Not bad for $280 :D
 
hey guys I wrote that article, I'll be using old generation cards and todays budget based cards in part III of my source engine benchmarks, just wait a few days :)
 
What is next ?

A test how a 1600 * 1200 * 32 hmmmm oh brother
Is this another atemt to find way the can postpone this again.

I feel that is is increadible to see they still have not realsed any demo or something so that WE! can see this for our self.

I know!!! that hl2 will work on my Pentium 4 2.4ghz 512mb, nvidia mx 420 so valve why not just release a mineute of gameplay of half life 2 or cs source. ????

Or maybe it is so full of bugs that u dont dare.
 
Thanks for the update paulj.

It's always good to see benchmarks for video cards that the average gamer uses. Hopefully mine will be included. :D
 
Paul your benchmarks seem inaccurate. Just based on my system:

Athlon xp 3200
ATI 9800XT
1 Gig (2 512 cosair dual channel DDR400)
60 gig (2 - 30 GIG SATA Raid 0 setup)
Gigabyte nforce2 motherboard....

and i got AVG FPS of 50 both times I tried with

1280 x 1024
6X AA
8x AF
everything else on highest...

so why are you showing a 31 FPS for it..I can see if it was close but that is a whole 20 FPS
 
hmmm yeah that is odd..... that is really wierd, i'll re-benchmark it a few times and see if i get those scores.

what are your scores with noaa enabled?
 
maximus0402 said:
Paul your benchmarks seem inaccurate. Just based on my system:

Athlon xp 3200
ATI 9800XT
1 Gig (2 512 cosair dual channel DDR400)
60 gig (2 - 30 GIG SATA Raid 0 setup)
Gigabyte nforce2 motherboard....

and i got AVG FPS of 50 both times I tried with

1280 x 1024
6X AA
8x AF
everything else on highest...

so why are you showing a 31 FPS for it..I can see if it was close but that is a whole 20 FPS

Why would you say "average 50" if you ran the stress test? It tells you your exact average :p

I have a 9800 Pro AND AND...well, just see my sig for the score and system. I get 27.50 (4xAA/16xAF), and 33.50 (2xAA/16xAF)
 
I got like 49.13 or somthing like that I just rounded up to 50...close enough...it is still way off... I will try with 1280 x 960 with no AA or AF (by the way its 1280 x 960, because 1280 x 1024 says LCD on the pull down meu in the game so I did not pick that but it should not make the difference.) I be right back
 
Ok the results were:

with 4x AA 8 AF - 49.06 FPS

Without AA or AF - 73.42

I did notice that I the REFLECTIVE option, I put REFLECT ALL and run both of these tests but everytime I leave the CS beta and go back in it changes bac k to REFLECT WORLD. Weird but this may have nothing to do wiht it and it may do that to your machine too I assume its just not a true available option until the halflife 2 comes.
 
Any idea on what a 9600 256mb scores?

2.8 gig P4
1 gig DDR 400 pc3200
9600 256mb
60 gig drive

any ideas about FPS on full setting but without AA or AF at 1024X756 (i can live with jaggy edges)
 
Dougy said:
Any idea on what a 9600 256mb scores?

2.8 gig P4
1 gig DDR 400 pc3200
9600 256mb
60 gig drive

any ideas about FPS on full setting but without AA or AF at 1024X756 (i can live with jaggy edges)

I think the paul guy that did the articles will do a test for your card tomorrow on part 3. However seeing the trend for average FPS I think your going to run great with no AA or AF at 1024 x 768
 
maximus, put it on reflect world and see if that makes any difference.
 
Slick review.
Doom 3 and HL2 are flip flops between the high-end cards for sure.
 
paulj said:
maximus, put it on reflect world and see if that makes any difference.

I did since that is was it ends up gong back to anyway...but I just ran it with 4x aa and 8x af with reflect world and it gave me 51.60 FPS
 
I wonder what kind of machine could run Half-Life 2 with the resolution at 2048x1536 with all the detail textures turned to high and had all the advanced options turned to on and but on FSAA, Anit-Aliasing and Anistropic Filtering to the highest and still get 120 framerates per second?

A NASA Super-Computer? Perhaps.
Or maybe... The NASA Omega-Ultra Super Computer 10K.

Damn, I wish I could own one.
;(
 
Tredoslop said:
I wonder what kind of machine could run Half-Life 2 with the resolution at 2048x1536 with all the detail textures turned to high and had all the advanced options turned to on and but on FSAA, Anit-Aliasing and Anistropic Filtering to the highest and still get 120 framerates per second?

A NASA Super-Computer? Perhaps.
Or maybe... The NASA Omega-Ultra Super Computer 10K.

Damn, I wish I could own one.
;(

The next generation of CPU's from AMD should be able to help with that. Also, the new NVIDIA SLI systems that will be coming out later this year should be pretty good at HL2. :cool:
 
A-Train said:
That's interesting to see how the GT still outperforms the X800 Pro, glad I bought it!..Wohoo

They're using a GT OC, now what's the OC? Overclocked? in that case it's normal that it beats the X800P at stock speeds. Besides there's only one or two frames of difference. And yeah I'm an ATI fanboy haha :LOL:
 
CB | Para said:
They're using a GT OC, now what's the OC? Overclocked? in that case it's normal that it beats the X800P at stock speeds. Besides there's only one or two frames of difference. And yeah I'm an ATI fanboy haha :LOL:

I think they make that exact insertion in the article, but they were quick to point out that the Pro and the GT were both priced the same (400 bucks), and that for the dollar, the GT was coughing up better performance and that the decision to put 12 texture pipelines instead of 16 in the Pro was a bad move by ATI, :D suck it down monkeyboy!
 
A-Train said:
I think they make that exact insertion in the article, but they were quick to point out that the Pro and the GT were both priced the same (400 bucks), and that for the dollar, the GT was coughing up better performance and that the decision to put 12 texture pipelines instead of 16 in the Pro was a bad move by ATI, :D suck it down monkeyboy!

Ah you got me ;)
 
Dougy said:
Any idea on what a 9600 256mb scores?

2.8 gig P4
1 gig DDR 400 pc3200
9600 256mb
60 gig drive

any ideas about FPS on full setting but without AA or AF at 1024X756 (i can live with jaggy edges)

My brother has an almost similar system, although with a 9600 XT, so I pressume there shouldn't be much difference between pro and XT.

1024 x 768, no AA and AF = 58.9 FPS
1024 x 768 2 x AA, (not sure about the AF) = 42.6 FPS
1024 x 768 4 x AA, (not sure about the AF) = 30.4 FPS
 
WOW WOW WOW WOWWY!

My pc has consistantly been getting frame rates of 117 to 125 through all the variations of AA, AF and resolutions. I can't wait to takle the tough stuff!
 
Tackle*

And that stress test IS the tough stuff, that is why it is called a STRESS test.
 
Paul I think you should have released part 3 first and part 1 last. :| I got a Ti 4200 and wanna see how it fares.
 
I'd be more interested to know if:

a. The HL2 benchmark will make use of ShaderModel 3.0.

b. Whether this effects the outcome when DirectX 9.0c is installed allowing the NVidia cards to take advantage of SM3.0.

It's been shown previously that the 6800 series cards get a significant performance boost from SM3.0 ... a feature ATI doesn't currently have in their cards.
 
my result: 52.77 fps

res: 1024x768
details: all 'high'
reflection: world only
AA: none
AF: 8x

athlon 2100
kt266a
768MB
9600xt 128MB

:cheers:
 
I dont trust the Stress test, AT ALL.

I ran it with my system.

Pentium3 1 ghz
512 Megs RAM
Geforce FX 5200 Asylum Ultra

And I get, with EVERYTHING on, 48 FPS.

What?? 48 FPS?? I thought, NO WAY! I can run HL2 NO PROB!!! I then jump into CS:S and find myself running at 9 fps. Thats the high end of my FPS, not the average, I average like 3 fps. I start my own server and just play with myself in there, same thing, but at like 15 fps.

I think the Stress test is bunk guys. That is of course, we are all talking about the one in CS:S. Theres no way a P3 1ghz with a SUPER powerful FX 5200 is rocking 48fps.
 
Back
Top