NV40 vs. R420. R420 better for HL2?

[HardWarrior]

I read a preview of HL2 in PCgamer yesterday morning (98%, BTW). The reviewer mentioned that, due to a plumbing differences between the two cores, the R420 would supply a 20% performance boost over the NV40 core. I haven't been into the tech side of video cards for a while (having a new card makes me want to learn) and I didn't understand his explanation of why he thought this was true.

What I'd like to know, in video-layman terms, is whether this is true and if so, why. I'm not all freaked up about it, just very curious. Thanks in advance.

 

[Keysplayr]

Originally posted by: HardWarrior
I read a preview of HL2 in PCgamer yesterday morning (98%, BTW). The reviewer mentioned that, due to a plumbing differences between the two cores, the R420 would supply a 20% performance boost over the NV40 core. I haven't been into the tech side of video cards for a while (having a new card makes me want to learn) and I didn't understand his explanation of why he thought this was true.

What I'd like to know, in video-layman terms, is whether this is true and if so, why. I'm not all freaked up about it, just very curious. Thanks in advance.

We won't know til the middle of November. And this time, we really mean it!!! LOL

However, if you look up all the HL2 beta tests and all the SOURCE engine stress test stuff, NV seems to have an edge. But if your not a believer in beta tests, then it's November for you.

 

[HardWarrior]

Originally posted by: keysplayr2003

Originally posted by: HardWarrior
I read a preview of HL2 in PCgamer yesterday morning (98%, BTW). The reviewer mentioned that, due to a plumbing differences between the two cores, the R420 would supply a 20% performance boost over the NV40 core. I haven't been into the tech side of video cards for a while (having a new card makes me want to learn) and I didn't understand his explanation of why he thought this was true.

What I'd like to know, in video-layman terms, is whether this is true and if so, why. I'm not all freaked up about it, just very curious. Thanks in advance.

We won't know til the middle of November. And this time, we really mean it!!! LOL

However, if you look up all the HL2 beta tests and all the SOURCE engine stress test stuff, NV seems to have an edge. But if your not a believer in beta tests, then it's November for you.

November 16th! (drool) The reviewer claimed to be talking about the retail version. I hope you're right about the nV advantage.

I bought the Steam version, so I expect to wake up on the morning of the 16th with that baby ready to go. I'll have a fresh bag of cheetos and an icy 2-liter ready.

 

[tfinch2]

I'm sure it will play just fine on both. The difference will probably be like 3-7 fps at the same resolution, if there is even a difference. When you're at 50-60 fps at high resolution that's nothing to get your panties in a bunch about...

 

[klah]

Here are some benchmarks from the leaked HL2: http://www.xbitlabs.com/articl...ics-cards-2004_26.html

 

[HardWarrior]

That's what I was thinking. I scored 109 on the CSS stress. I'd be really surprised if that turned into less than 60 for HL2.

 

[Marsumane]

Both cards will be fine for the game. It doesnt seem to be all that demanding compared to a game like doom 3. It will be more like 60 vs 53 fps with most, if not all settings on. Thats my speculation based from what ive seen.

 

[gururu]

yea this game seems to be so incredibly optimized to run on both platforms. valve hasn't forgotten anyone.

 

[HardWarrior]

Originally posted by: gururu
yea this game seems to be so incredibly optimized to run on both platforms. valve hasn't forgotten anyone.

In that they were damned smart. I was expecting to hear a bunch of crying from the CS crowd about the more demanding CSS, but even that hasn't appeared.

 

[lordtyranus]

Originally posted by: keysplayr2003

Originally posted by: HardWarrior
I read a preview of HL2 in PCgamer yesterday morning (98%, BTW). The reviewer mentioned that, due to a plumbing differences between the two cores, the R420 would supply a 20% performance boost over the NV40 core. I haven't been into the tech side of video cards for a while (having a new card makes me want to learn) and I didn't understand his explanation of why he thought this was true.

What I'd like to know, in video-layman terms, is whether this is true and if so, why. I'm not all freaked up about it, just very curious. Thanks in advance.

We won't know til the middle of November. And this time, we really mean it!!! LOL

However, if you look up all the HL2 beta tests and all the SOURCE engine stress test stuff, NV seems to have an edge. But if your not a believer in beta tests, then it's November for you.

I've seen the R420 dominate the built in stress test.

http://www.anandtech.com/video...oc.aspx?i=2182&p=5

 

[darkswordsman17]

Thats an old test. There have been newer ones that show the 6800 and X800 cards pretty close in the tests.

Of course Valve has said that the game performs 20% better on ATi's cards. For the R3xx vs NV3x this seems possible, but not with the new ones.

 

[dguy6789]

It dont really matter, either of the x800s and 6800s will run the game perfectly. The game was desogned around the R3XX architecture, and a radeon 9800 based card will run it perfectly at max settings, so obviously faster cards will as well.

 

[Keysplayr]

Originally posted by: gururu
yea this game seems to be so incredibly optimized to run on both platforms. valve hasn't forgotten anyone.

Lets hope valve still honors the HL2 coupons that were printed on papyrus.

 

[Pete]

Originally posted by: darkswordsman17
Thats an old test. There have been newer ones that show the 6800 and X800 cards pretty close in the tests.

Of course Valve has said that the game performs 20% better on ATi's cards. For the R3xx vs NV3x this seems possible, but not with the new ones.

I don't recall any newer HL2 tests, as (1) HL2 isn't out, and (2) Valve hasn't released any more benchmarks.

If you mean CS:S tests, is it known how comparable HL2 is to CS:S in terms of shaders and such? If you mean xbit's benches, they're of the year-old leaked alpha/beta/blahblah.

 

[RussianSensation]

It seems Nvidia does a little better, but it all depends on the website which posts the review.

HL2 Beta Test - digit life.com

Either way as long as you have the latest generation of videocards from either company, you should be fine.

 

[OinkBoink]

how the hell is nV performing better this time???the game is supposed to be optimised for radeons isnt it???

 

[quikah]

FYI: The part of PC Gamer which says that ATi runs 20% faster was from the Q&A with Valve. They are the ones saying it runs 20% faster, not PC Gamer.

 

[carage]

The German version of PCGamer did a review and benchmark.
NV40 only loses about 2 to 3 fps using the Source: Video Stress test.
Perhaps Valve is just sucking up to ATi, or nVidia pulled off another miracle in their drivers?

http://forum.pcdvd.com.tw/showthread.php?t=402788

 

[HardWarrior]

Originally posted by: quikah
FYI: The part of PC Gamer which says that ATi runs 20% faster was from the Q&A with Valve. They are the ones saying it runs 20% faster, not PC Gamer.

Well now, that changes things a bit.

 

[BFG10K]

Based on the benchmarks I've seen ATi has the edge though nVidia is close. I haven't seen anything that resembles a performance rift like Doom3.