• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

[H] Cpu Scaling Review

A

agar

Limp Gawd
Joined
Oct 11, 2003
Messages
170
Found this review very interesting. Is it me or is the X800pro holding its own against the 6800U? I have to go over and look at it again, I just can't believe what I just saw. Yep, I guess those are the results they found. Another interesting detail was that the XT was put at a higher res and it would still either beat the 6800U or be withing 3-5fps. Is it me or are these results quite different from the recent reviews using the 61.34 drivers (ex. Guru3d)?

http://www.hardocp.com/article.html?art=NjMy
 
Do you think its the use of FRAPS instead of timedemos? I know they are custom, but this is strange.
 
Don't know what to tell ya, the results shown are from real gameplay in those games on the cards tested. No timedemos were used, only real gameplay evaluation. The performance was repeatable.
 
Brent_Justice said:
The performance was repeatable.
Click to expand...

By HOCP only. I have yet to see another site match these results, especially in the light of how close the X800Pro is to the 6800U Ultra. I'm not making accusations, simply pointing out this rather large discrepancy. I'm afraid I can't accept these results unless other sites back them up and they simply are not. And to top it off people with retail 6800Us are saying that Far Cry is fully playable at 1600x1200 4xAA16xAF with an average of 40 frames and the occasional dip towards the 20s. I'm afraid while I respect HOCP a great deal, I simply cannot take their word over people like Jackup who have the cards. Ah well, I'll find out myself in a few days as I will have a very similar system to the ones used in the article.
 
I trust your revies Brent, I'm not doubting your competency in writing fair and comprehensive reviews, I just find it very interesting that many other sites seem to have quite drastic differences in their results while using custom timedemos. Hell, after reading anand's NV45 preview I wonder if he even ueses custom timedemos. His results were also remarkably different when comparing Farcry at 1600x1200 4XAA/8XAF XT/PE ~ 54.2 to [H]'s using similar settings 1600x1200 4XAA/16XAF XT/PE ~ 35.4. I know its not the best comparison between FRAPS and timedemos, but I still find it strange.
 
so can anyone tell me, is my athlon xp barton 3200+ useless. i just spent 482 on an x800xt from cdw, i dont want to spend another 400 on cpu+mobo
 
Vagrant Zero said:
By HOCP only. I have yet to see another site match these results, especially in the light of how close the X800Pro is to the 6800U Ultra. I'm not making accusations, simply pointing out this rather large discrepancy. I'm afraid I can't accept these results unless other sites back them up and they simply are not. And to top it off people with retail 6800Us are saying that Far Cry is fully playable at 1600x1200 4xAA16xAF with an average of 40 frames and the occasional dip towards the 20s. I'm afraid while I respect HOCP a great deal, I simply cannot take their word over people like Jackup who have the cards. Ah well, I'll find out myself in a few days as I will have a very similar system to the ones used in the article.
Click to expand...

Doesn't Jakup have his 6800U at like 450+ or around there? That will definely show a difference.
 
FYI: Recorded timedemos in FarCry do NOT playback the AI, there is no shooting, no damage being taken, nothing, it is like a flyby.

Of course performance is going to be better in a recorded timedemo. It does not accuretly reflect gameplay.

Our real world gameplay evaluations are manual run-throughs in God mode taking on all the gun firring, all the AI reacting, and the visual damage that occurs to the player.

Basically our run-throughs show you real world game performance that the cards deliver.
 
KillDFurby said:
so can anyone tell me, is my athlon xp barton 3200+ useless. i just spent 482 on an x800xt from cdw, i dont want to spend another 400 on cpu+mobo
Click to expand...

It's obviously not going to give you what a FX-53 or 3.4EE can, however I don't see why you won't be able to play at high res, high quality AA/AF.
 
agar said:
Doesn't Jakup have his 6800U at like 450+ or around there? That will definely show a difference.
Click to expand...

Yes, but when he was reporting the Far Cry numbers he was running it as stock (by request).
He's also reported on the numbers using 61.34, 61.45, and 61.71 noting that there was no big performance leaps as from the 6072.s to 61.34s (ie, Average was 40s, indoors typicaly was 50s, and the game did occasionally slip into the 20s but not enough to warrant lowering the settings).
 
agar said:
It's obviously not going to give you what a FX-53 or 3.4EE can, however I don't see why you won't be able to play at high res, high quality AA/AF.
Click to expand...
good because i plan on playing at my usual 1600x1200 2xAA 4xAF
 
KillDFurby said:
so can anyone tell me, is my athlon xp barton 3200+ useless. i just spent 482 on an x800xt from cdw, i dont want to spend another 400 on cpu+mobo
Click to expand...
It's fine, just crank up the aa and af, and enjoy minimal performance hits.
 
Vagrant Zero said:
Yes, but when he was reporting the Far Cry numbers he was running it as stock (by request).
He's also reported on the numbers using 61.34, 61.45, and 61.71 nothing that there was no big performance leaps (ie, Average was 40s, indoors typicaly was 50s, and the game did occasionally slip into the 20s but not enough to warrante lowering the settings).
Click to expand...

Then what are you implying? The 6800U is faster than what Brent found? Have anyone with a 6800U clock it at 400mhz, same drivers, same settings, same game and use FRAPS and let us know what they get. I have a pretty decent feeling it won't be much different.
 
There was an old article on the 'net somewhere that had a theory on why some games would feel "laggy" even at 60fps. There is no way I'd be able to find the article now.

It was the ones that had on double or triple buffering, and Tom kind of hit on it when he was rightfully knocking the Rage Fury Maxx's AFR rendering technique (BTW, it should not affect Nvidia's new SLI much, as it does not need a triple buffer like AFR to achieve its effect.)

Most games always try to render at least one frame ahead of time, regardless of input from your mouse or keyboard. IE: If you switch from a left movement to a right movement, in that exact instance the 1/x hz fps buffer still thinks and is rendering your left movement. In a worst case scenario, with an active triple buffer (99 percent full on one frame, one frame pre-buffered and one being shown) if the FPS displayed 60fps, your mouse input could actually make it feel like 20.0001 FPS
 
agar said:
Then what are you implying? The 6800U is faster than what Brent found? Have anyone with a 6800U clock it at 400mhz, same drivers, same settings, same game and use FRAPS and let us know what they get. I have a pretty decent feeling it won't be much different.
Click to expand...

I'm not implying it. I'm stating it. Jackup's and others [who DID run it at those speeds, same drivers, same settings, same game] are reporting remarkedly different results. But instead of sharing your feelings with me, might I suggest you go straight to the horses mouth and ask the people over at nvnews what their experiance with the 6800 has been in Far Cry. Best way to resolve this way I see it.
 
It used to be commonly reffered to as "mouse lag" as long as pre-buffers exist and things are not instantly drawn to the screen, it will always exist.
 
ZenOps said:
It used to be commonly reffered to as "mouse lag" as long as pre-buffers exist and things are not instantly drawn to the screen, it will always exist.
Click to expand...

wrong thread? ;)
 
 
Vagrant Zero said:
I'm not implying it. I'm stating it. Jackup's and others [who DID run it at those speeds, same drivers, same settings, same game] are reporting remarkedly different results. But instead of sharing your feelings with me, might I suggest you go straight to the horses mouth and ask the people over at nvnews what their experiance with the 6800 has been in Far Cry. Best way to resolve this way I see it.
Click to expand...

Curious Jackup was using Fraps? You didn't say that or not.
 
Brent_Justice said:
FYI: Recorded timedemos in FarCry do NOT playback the AI, there is no shooting, no damage being taken, nothing, it is like a flyby.

Of course performance is going to be better in a recorded timedemo. It does not accuretly reflect gameplay.

Our real world gameplay evaluations are manual run-throughs in God mode taking on all the gun firring, all the AI reacting, and the visual damage that occurs to the player.

Basically our run-throughs show you real world game performance that the cards deliver.
Click to expand...

quoted again just incase anyone missed it at the bottom of the first page
 
oh man time to try to break 2.4ghz so i can get all out of my arriving x800pro(and my bartonf or that matter)

and i still find it interesting how the x800PRO holds it own against the 6800behemoth

hey Brent, any chance you will get a shootout of the 6800GT and x800pro anytime soon?
and possibly the 6800nu
 
retardedchicken said:
oh man time to try to break 2.4ghz so i can get all out of my arriving x800pro(and my bartonf or that matter)

and i still find it interesting how the x800PRO holds it own against the 6800behemoth

hey Brent, any chance you will get a shootout of the 6800GT and x800pro anytime soon?
and possibly the 6800nu
Click to expand...

6800GT, count on it
 
i've been waiting for a x800pro and 6800gt head-to-head review... still on the fence on which to get or keep.

hey Brent, if you had the choice between an x800pro and a 6800gt, which would YOU choose and why? thanks
 
bobmanfoo said:
i've been waiting for a x800pro and 6800gt head-to-head review... still on the fence on which to get or keep.

hey Brent, if you had the choice between an x800pro and a 6800gt, which would YOU choose and why? thanks
Click to expand...

i don't know yet, i' haven't played any games on a 6800GT yet

when i do the review it'll all be in there
 
Cryect said:
Curious Jackup was using Fraps? You didn't say that or not.
Click to expand...

I wasn't sure and didn't want to guess. It's a 37 page thread. I'm sure now since I just sat down and sifted through all 37 pages [yup he used FRAPS]. You guys better apppreciate this, I've got a headache now.

http://www.nvnews.net/vbulletin/showthread.php?t=29518&page=33&pp=15&highlight=Cry+Fraps

1024x768 8xAA/8xAF Ave 60s Min 50s Max 90s [FORCED R300 PATH using ???, I don't know if he was using 61.34s or 61.71s]

1600x1200 8xAA/16xAF Ave 30s Min 10s Max 40s [again, forced R300 path using ???]

Also all the optimizations and stuff were turned off [doesn't say on that page, says it on the next page].

BTW, he ran those at stock speeds [might not say on that page...and there's too many for me to remember by heart...but I know he stopped the OC since UT2K4 refuses to run at anything but stock speeds for him].

PS. Personally, I don't consider Nvidia's 8xAA a playable mode, you take a MASSIVE performance hit going from 4xAA to it. I'd bet good money that if he'd run FC at 4xAA we'd have seen on average a 10-15FPS increase. Even more if he'd left the optomizations on.
 
Thank you again Brent for a great article.

I do have a suggestion for your next comparison/review(not preview) of the 6800 vs. x800.

I feel if you are going to use the Platinum Editon for the x800, then the 6800 Extreme Edition should be used as well. If you are using Ati's best then Nvidia's best should be used at the same time. If I read an article comparing the 6800 Ultra EE against the x800,it would appear unbalanced by me. Even though I know it is not deliberate.




Please scroll down to the yellow chart.
http://www.firingsquad.com/hardware/geforce_6800_ultra_extreme/

Fresh off the introduction of GeForce 6800 Ultra a few weeks ago, NVIDIA is back again with a pair of new GeForce 6 products, and we’re not talking about the GeForce 6800 either. On the ultra high-end of the market, NVIDIA and its board partners will be producing an overclocked GeForce 6800 Ultra part, a final name has not been officially announced, but the running title is GeForce 6800 Ultra (Extreme). This board ships with a 450MHz NV40 graphics core, 50MHz higher than the default GeForce 6800 Ultra and effectively boosting fill-rate by nearly 10%.
Curiously enough, the memory clock frequency remains the same at 550MHz (1.1GHz effective) for a peak memory bandwidth figure of 35.2GB/sec.

These figures are certainly impressive, but do fall shy of ATI’s X800 XT Platinum Edition card, which features a 520MHz core clock and 560MHz memory clock (1.12GHz effective). NVIDIA and its board partners have not come up with a final price point on the Extreme 6800 Ultra board yet, but it will hit retail in June.

Sitting between the GeForce 6800 Ultra and the GeForce 6800 is NVIDIA’s GeForce 6800 GT. This is a 16-pipeline part, just like NVIDIA’s GeForce 6800 Ultra series and ships with 256MB of GDDR3 memory clocked at 500MHz (1.0GHz effective). Its 350MHz graphics core is boasts a peak texel fill rate of 5.6 Gigatexels/second.

This board is priced at the same $399 price point as ATI’s X800 PRO, which sports a 12-pixel pipeline architecture operating at 475MHz (for a peak fill rate of 5.7Gigatexels/second) and 256MB of GDDR3 memory operating at 450MHz (900MHz effective).

With these changes in mind, let’s take a look at NVIDIA’s entire GeForce 6800 lineup:
Notes
NVIDIA has also announced the final details on its GeForce 6800 part, 325MHz core/350MHz memory. We’ve heard from one of NVIDIA’s board partners that the GeForce 6800 should offer a nice performance improvement over GeForce FX



Maybe they are impossible for you to get right now. But they do exist.
http://www.evga.com/6800ultra.asp
 
Vagrant Zero said:
I wasn't sure and didn't want to guess. It's a 37 page thread. I'm sure now since I just sat down and sifted through all 37 pages [yup he used FRAPS]. You guys better apppreciate this, I've got a headache now.

http://www.nvnews.net/vbulletin/showthread.php?t=29518&page=33&pp=15&highlight=Cry+Fraps

1024x768 8xAA/8xAF Ave 60s Min 50s Max 90s [FORCED R300 PATH using 61.34s]

1600x1200 8xAA/16xAF Ave 30s Min 10s Max 40s [again, forced R300 path using 61.34s]

Also all the optimizations and stuff were turned off [doesn't say on that page, says it on the next page].

BTW, he ran those at stock speeds [might not say on that page...and there's too many for me to remember by heart...but I know he stopped the OC since UT2K4 refuses to run at anything but stock speeds for him].

PS. Personally, I don't consider Nvidia's 8xAA a playable mode, you take a MASSIVE performance hit going from 4xAA to it. I'd bet good money that if he'd run FC at 4xAA we'd have seen on average a 10-15FPS increase. Even more if he'd left the optomizations on.
Click to expand...


BTW with the 61.34's you have to set AA from inside the game, if you set it from the control panel it does nothing in the game. And the setting inside the game has no 8X equivelent setting. Low = 2X, Medium = 4X and High = 4X.
 
I've read alot of reviews about the new cards from ATi and nVidia but I can't decide what to get my hands on. Currently I have a 9800 Pro@440/780. My CPU is a P4 3.2@3.84Ghz. So my question is what should I get to couple with my CPU?
 
Bad_Boy said:
how much does the x800XT (not PE) cost?
Click to expand...

xxx + the price of a pci-e motherboard. Last I heard there will be no XTs for AGP. At least not during the initial launch.
 
Vagrant Zero said:
By HOCP only. I have yet to see another site match these results, especially in the light of how close the X800Pro is to the 6800U Ultra.
Click to expand...


I have yet to see another site use REAL GAMEPLAY to make their evaluations. I am waiting for some of you to see the forest......
 
davidj said:
Thank you again Brent for a great article.

I do have a suggestion for your next comparison/review(not preview) of the 6800 vs. x800.

I feel if you are going to use the Platinum Editon for the x800, then the 6800 Extreme Edition should be used as well. If you are using Ati's best then Nvidia's best should be used at the same time. If I read an article comparing the 6800 Ultra EE against the x800,it would appear unbalanced by me. Even though I know it is not deliberate.
Click to expand...

All reviews will deal with retail available cards.

You will see some cards from manufacturers, like BFG, that are "OC'd" by default, for example the BFG 6800U OC which has a higher clocked core speed.

Or the one you posted.

This is taken directly from NVIDIA's website:

Available models: GeForce 6800 Ultra, GeForce 6800 GT, and GeForce 6800
Click to expand...

No mention of an "Extreme Edition".
 
Also for a CPU scaling review the number of games you benched was more than adequate, but if you do conduct the inevitable x800pro/xt vs. 6800nu/gt/u shootout, would it be possible to test with more than 4 or 5 games - say 10-15 of the more popular ones. Also, since they are quite the rage these days, how about a subjective analysis of performance on one of the popular MMORPGs (i.e. lineage 2, city of heroes, anarchy online, star wars galaxies, etc)? Would be tought to get solid data but it would be nice to know how the cards perform in these games which can be extremely taxing at times with tons of characters and textures on screen at once.
 
tranCendenZ said:
Also, since they are quite the rage these days, how about a subjective analysis of performance on one of the popular MMORPGs (i.e. lineage 2, city of heroes, anarchy online, star wars galaxies, etc)? Would be tought to get solid data but it would be nice to know how the cards perform in these games which can be extremely taxing at times with tons of characters and textures on screen at once.
Click to expand...

We are DEFINITELY going to add that genre of games to our mix.

In your opinion which is the most taxing?

City of Heroes pretty good?
 
mine really isnt different, i like using god mode :D


i'd like to see warcraft 3, but it will get pwned and a half by the new cards


i can run it at 12x10 2xAA/16xAF with pretty good rates (only slows when there are about 300 guys on the screen....which doesnt happen often)
 
Vagrant Zero said:
Regardless, a real users experiance [and not just his though he's the most active of the lot] is vastly different from the ones being shown in the article. Granted I'd expect a little bit of parity, the unverise is like that...but this is more of a day/night difference than a parity. Care to comment or give an opinion on what we're seeing here? Surely you must have some idea I'd imagine.
Click to expand...

I'm not seeing the same thing you are. He didn't test the same way we did, how could you possibly compare?
 
You must log in or register to reply here.
Back
Top