PDA

View Full Version : Rage3D GeForce 7800GT Review


Pages : [1] 2

Ratchet
08-26-05, 01:00 AM
Hey guys, you might remember that we posted a review of a reference GeForce 7800GTX (http://www.rage3d.com/reviews/video/nvidia7800gtx/) last month. NVIDIA was cool enough to also send out a 7800GT to us as well and I've just published the review here (http://www.rage3d.com/reviews/video/nvidia7800gt/) on Rage3D.

As before I would really appreciate some comments on this review from the guys here.

Toss3
08-26-05, 02:44 AM
Well I have to say I really liked the review! It was plain, simple, shor enough and you benchmarked most of the top games today! This was actually one of the best reviews I've read yet, but you should have read the overclocking thread here: http://www.nvnews.net/vbulletin/showthread.php?t=55571 first...

But very nice job anyways! :)

Muppet
08-26-05, 03:02 AM
Very nice it's good to see an unbiased review from an opposition web site. :)

F_L_C
08-26-05, 03:18 AM
Very professional, totally objective. But this market imbalance is causing weird and unnatural vibes. Competition absolutely needs to be restored. Sure, we, on the NV side of the aisle, privately (or publicly) gloat over reviews such as these (especially considering its origin). But as consumers, we lose if ATI does not get the ball rolling very soon.

Toss3
08-26-05, 11:16 AM
Very professional, totally objective. But this market imbalance is causing weird and unnatural vibes. Competition absolutely needs to be restored. Sure, we, on the NV side of the aisle, privately (or publicly) gloat over reviews such as these (especially considering its origin). But as consumers, we lose if ATI does not get the ball rolling very soon.
How so? I mean how would it effect us in any way? The prices on the 7800gtx's are the same as they have always been for the top cards. And I'm sure ATI's new card will be equal or faster than the 7800gtx, which is why(IMO) nvidia has been keeping back a 32pixelpipeline model of the 7800! :rolleyes: So "this" won't actually effect us, imho at least... Those who have gotten their 7800Gtx's have the fastest cards there is and the prices are likely to drop when ATI comes out with their new card, which is why this will only effect ATI, not us!

Marcos
08-26-05, 01:37 PM
get hamixada to review nvidia cards. then he will draw strong points to an argument on how his graphing calculator rapes your mother every night.

anzak
08-26-05, 03:40 PM
Great review ratchet. Your reviews are the most well thought out and organized I've read. The IQ comparison has to be my favorite part though.

SH64
08-26-05, 08:56 PM
I like your reviews & all Ratchet .. but hows the X850XT-PE is faster in FEAR SP bench ?? (i'm talking about the 4xAA/16xAF numbers)
its a well-known fact the GTX is faster in FEAR.
http://www.bit-tech.net/hardware/2005/08/09/gainward_7800gtx_gs/8.html (or see attachment)

even more so i did a quick bench on my GTX vs my X800XT-PE (520/570)
just to make sure & heres the results directly from FRAPs :
2005-08-27 04:33:54 - FEARSPDemo (X800XT-PE@1280x960,4xAA,8xAF)
Frames: 1424 - Time: 69770ms - Avg: 20.410 - Min: 0 - Max: 82

2005-08-27 04:35:47 - FEARSPDemo (7800GTX@1280x960,4xAA,8xAF)
Frames: 1843 - Time: 63169ms - Avg: 29.176 - Min: 12 - Max: 111

i picked places with no action just to test the graphics & not to get the CPU involved much in it.
you can disregard the min fps too because the XT-PE system had 1GB of RAM.
just take a look at the avarage framerate & add whatever you like for the increased core/mem clock for the X850XT-PE. you would still see the GTX is clearly the winner here .. heck even the GT might take the second place IMO.

maybe AA wasent running in the XT's case ?? notice that AA/Softshadows reset to defualts (0x/disabled) everytime you run the game.

Ratchet
08-27-05, 01:21 AM
I like your reviews & all Ratchet .. but hows the X850XT-PE is faster in FEAR SP bench ?? (i'm talking about the 4xAA/16xAF numbers)
its a well-known fact the GTX is faster in FEAR.
http://www.bit-tech.net/hardware/2005/08/09/gainward_7800gtx_gs/8.html (or see attachment)

even more so i did a quick bench on my GTX vs my X800XT-PE (520/570)
just to make sure & heres the results directly from FRAPs :

i picked places with no action just to test the graphics & not to get the CPU involved much in it.
you can disregard the min fps too because the XT-PE system had 1GB of RAM.
just take a look at the avarage framerate & add whatever you like for the increased core/mem clock for the X850XT-PE. you would still see the GTX is clearly the winner here .. heck even the GT might take the second place IMO.

maybe AA wasent running in the XT's case ?? notice that AA/Softshadows reset to defualts (0x/disabled) everytime you run the game.
How did you test? I did manual runs with Fraps through about half of the single player level (with all the enemies already dead), 3 times for each setting, then averaged the results (quite tedius). It might be that the sequence you ran, the sequence Tim ran, and the sequence I ran were all completely different. It's hard/impossible to generate like results when there are so many different variables.

I'm certain I benched the cards at the same settings for each card (as shown in the screenshot next to the benchmark chart). In fact, I distinctly remember seeing the "Detecting hardware" notice whenever I changed cards which reminded me to recheck the settings. I also had the screenshot of the settings (shown next to the charts in the review) showing on my work computer to match against to make sure that they were all the same (yes, the "detecting hardware" notice made me that paranoid about my settings).

However, to be absolutely sure, I reran the tests (only the 7800GT vs X850XT, I would have run the GTX if it became apparent the scores I originally generated were inaccurate) just to make double sure that the results I generated were accurate. The new results were well within an acceptable margin of error from what I've got in the review (+/- ~1 fps), so I stand by the original results.

edit: I noticed that Tim used Medium (http://www.bit-tech.net/content_images/gainward_7800gtx_gs/fear-settings2.jpg) for most of his Advanced Video Card Options whereas I used Maximum (http://www.rage3d.com/image.php?pic=/reviews/video/nvidia7800gt/shots/fearperf2.jpg&comment=Settings), do you think that might have something to do with the differences between our results?

SH64
08-27-05, 07:32 AM
I used absoulte maximum settings aswell (& tried the game with/without softshadows).. the general feel when i play the game on both PCs is undoutly showing the GTX superiority.
the scene i took was from game start up till the first checkpoint .. same path same everything.
if you see the results above you can conclude a difference of about 30% right?
well heres 2 standstill screenshots showing appox. the same percentage.

http://img377.imageshack.us/img377/157/fearspdemo20050827043016073jd.jpg

http://img395.imageshack.us/img395/6654/fearspdemo20050827043353293pk.jpg

i can take more screenshots from throught the game or bench through different places or through the place you tested if thats of any interest for you.

of course different places will give different numbers but i assure that i played the game on both PCs one next to the other & without counting the hitching/stuttering with 1Gb on the X800 PC i can clearly see the higher framerate is always on the GTX side.

EDIT : hmm you have the texture filtering set to Trilinear ??! i thought it was set to 16x Anisotropic!
also when AA is set to off its no problem .. it starts once you enable it because it will be back at 0 everytime you start the game. the same happens with the softshadows.
i'm not sure how aware are you with this bug though. anything wrong will result in playing without AA or SShadows or both.
also i see you've set the volumetric light density to medium so its not all max afterall ?

MUYA
08-27-05, 07:37 AM
Hey guys, you might remember that we posted a review of a reference GeForce 7800GTX (http://www.rage3d.com/reviews/video/nvidia7800gtx/) last month. NVIDIA was cool enough to also send out a 7800GT to us as well and I've just published the review here (http://www.rage3d.com/reviews/video/nvidia7800gt/) on Rage3D.

As before I would really appreciate some comments on this review from the guys here.
I presume you have permission to to post the link from MikeC? If not, please do so in advance before you post links to your own website. We do not allow web site self pimping in these forums.
Since this is the first time I have seen you doing so, I will just remind u of our policy. Next time please obtain prior permission.

Ratchet
08-27-05, 08:19 AM
I used absoulte maximum settings aswell (& tried the game with/without softshadows).. the general feel when i play the game on both PCs is undoutly showing the GTX superiority.
the scene i took was from game start up till the first checkpoint .. same path same everything.
if you see the results above you can conclude a difference of about 30% right?
well heres 2 standstill screenshots showing appox. the same percentage.

<snip>

i can take more screenshots from throught the game or bench through different places or through the place you tested if thats of any interest for you.

of course different places will give different numbers but i assure that i played the game on both PCs one next to the other & without counting the hitching/stuttering with 1Gb on the X800 PC i can clearly see the higher framerate is always on the GTX side.

EDIT : hmm you have the texture filtering set to Trilinear ??! i thought it was set to 16x Anisotropic!
also when AA is set to off its no problem .. it starts once you enable it because it will be back at 0 everytime you start the game. the same happens with the softshadows.
i'm not sure how aware are you with this bug though. anything wrong will result in playing without AA or SShadows or both.
also i see you've set the volumetric light density to medium so its not all max afterall ?
I set the game to "high" then modified only the AA/AF settings from there, so I guess "high" sets volumetric lighting to Medium. The area I took benchmarks from is near that place where the chopper comes out of the sky (there's a screenshot of the area on the right hand side of that review page). I actually start in the corridor just after that area and go toward the start of the level. As for the trilinear thing, that changes as I benchmark (that was the default setting shown, No AA/No AF). Like I said, I checked each setting before I started benchmarking, so I'm certain that is not the issue.
I'm not disputing your findings, I'm just saying my findings were different.

Ratchet
08-27-05, 08:21 AM
I presume you have permission to to post the link from MikeC? If not, please do so in advance before you post links to your own website. We do not allow web site self pimping in these forums.
Since this is the first time I have seen you doing so, I will just remind u of our policy. Next time please obtain prior permission.
He never said anything the last time when I posted my GTX review for feedback. In fact, he posted it on the main page as news.

MUYA
08-27-05, 08:22 AM
I am asking you, that you do seek permission from now on. It was out of irony that he linked your review on the front page maybe. DO not assume and DO RESPECT OUR RULES. Thanks

Ratchet
08-27-05, 08:26 AM
I am asking you, that you do seek permission from now on. It was out of irony that he linked your review on the front page maybe. DO not assume and DO RESPECT OUR RULES. Thanks
Wow, reputation deserved.

You won't have any problems from me any longer, I can guarantee that.

SH64
08-27-05, 08:31 AM
I set the game to "high" then modified only the AA/AF settings from there, so I guess "high" sets volumetric lighting to Medium. The area I took benchmarks from is near that place where the chopper comes out of the sky (there's a screenshot of the area on the right hand side of that review page). I actually start in the corridor just after that area and go toward the start of the level. As for the trilinear thing, that changes as I benchmark (that was the default setting shown, No AA/No AF). Like I said, I checked each setting before I started benchmarking, so I'm certain that is not the issue.
I'm not disputing your findings, I'm just saying my findings were different.

Of course its no problem as long as you are 100% sure about your results , but i just cant accept that esp when i'm testing the same thing infront of my eyes & i can clearly see that the GTX is faster.
it might be the difference in settings or other stuff .. i'm curious what it could be.

OWA
08-27-05, 10:17 AM
In fact, I distinctly remember seeing the "Detecting hardware" notice whenever I changed cards which reminded me to recheck the settings.
...
Like I said, I checked each setting before I started benchmarking, so I'm certain that is not the issue.
The FEAR settings issue is that you have to reselect the settings each time, not just recheck to see if they're showing what you think they're showing. The options menu will show the correct settings but it won't actually be enabled unless you set it again. You have to do that each time you start the game (for AA, soft-shadows, that is).

So, are you sure that you reselected them each time or did you just check the options menu to see if it was still showing what you thought it should be showing?

SH64
08-27-05, 10:49 AM
Thats what i'm suspecting too OWA.
FEAR settings can get pretty confusing if ones didnt pay a good attention to them. even if the options in the menu are shwoing AA/SS to be enabled it dosent really mean they are working in the game itself.

Superfly
08-27-05, 11:06 AM
The FEAR settings issue is that you have to reselect the settings each time, not just recheck to see if they're showing what you think they're showing. The options menu will show the correct settings but it won't actually be enabled unless you set it again. You have to do that each time you start the game (for AA, soft-shadows, that is).

So, are you sure that you reselected them each time or did you just check the options menu to see if it was still showing what you thought it should be showing?

EXACTLY - you MUST re-select the options not just look at them, annoying bug that I hope is gone in the retail version.

Ratchet
08-27-05, 11:50 AM
ok, I did everything again (7800 GT vs X850XT-PE) and made sure to reselect all the same options, and then reselect them again after each time I changed resolutions/settings. I still got the same exact results.

I've benchmarked this at least 4 times now. Sorry but I can't put more time into it, I have other reviews to get started on.

FraGTastiK
08-27-05, 02:19 PM
but i just cant accept that esp when i'm testing the same thing infront of my eyes & i can clearly see that the GTX is faster

I was wondering the same thing...

why the results are a bit off at some points?

btw F.E.A.R Demo is a bit tricky to bench.

reading the F.E.A.R Demo Tweak Guide (http://www.tweakguides.com/FEAR_1.html) should help.

OWA
08-27-05, 02:27 PM
ok, I did everything again (7800 GT vs X850XT-PE) and made sure to reselect all the same options, and then reselect them again after each time I changed resolutions/settings. I still got the same exact results.

I've benchmarked this at least 4 times now. Sorry but I can't put more time into it, I have other reviews to get started on.

Okay, the results still seems strange to me but thanks for the time spent re-checking the results.

F_L_C
08-27-05, 05:50 PM
@SH64: I doubt it would make any difference, but I noticed the reviewer is benching with CAT 5.7. Are you using CAT 5.8?

sniggle
08-27-05, 07:00 PM
LOVE your image comparison page. really well done there. great review, keep it up!

5150 Joker
08-27-05, 07:25 PM
I'm sure it's a good review but I refuse to read anything nVidia related on R3D. :)