BFG 8800 GTX OC vs. Diamond HD 2900 XT 1 GB mini-review

[n7]

Vital System Info
E6600 @ 3375 MHz (9x375)
4 GB DDR2-750 4-4-4-12
Raptor 150 GB
Vista Ultimate 64-bit

Drivers Used
nVidia Forceware 158.24 (current official driver)
AMD Catalyst 7.6 (current official driver)

Driver Settings
nV 3D settings: top bottom
AMD 3D settings: top middle bottom

8800 GTX OC 768 MB clocks at defaults: 594/1404/900 (slightly higher than the reference 8800 GTX's 575/1350/900)
HD 2900 XT 1024 MB clocks: 825/1050

For all benchs, i ran tests at least twice, usually 3 times actually, & if there was more than a slight difference in scores, then 4 or more.
Results posted are averaged from the multiple runs.



Benchmarks!


3Dmark01 default
8800 GTX - 46773
2900 XT - 44147.67


3DMark05 default
8800 GTX - 17841.67
2900 XT - 19681
3DMark05 @ 2560x1600 4xAA/16xAF
8800 GTX - 8163
2900 XT - 7186
3DMark05 @ 2560x1600 0xAA/16xAF
8800 GTX - 10985
2900 XT - 10392
3DMark05 @ 1600x1200 4xAA/16xAF
8800 GTX - 11834
2900 XT - 11471


3DMark06 default
8800 GTX - 10636.5
2900 XT - 11501.5
3DMark06 @ 2560x1600 4xAA/16xAF
8800 GTX - 4579.5
2900 XT - 4392.5
3DMark06 @ 2560x1600 0xAA/16xAF
8800 GTX - 6645
2900 XT - 6079
3DMark06 @ 1600x1200 4xAA/16xAF
8800 GTX - 7219
2900 XT - 6641.5


Company of Heroes built in bench @ 2560x1600 maxed
8800 GTX
min - 20.8 fps
ave - 48.23 fps
max - 121.07 fps
2900 XT
min - 18.17 fps
ave - 44.27 fps
max - 85.93 fps
Company of Heroes built in bench @ 2560x1600 maxed no AA
8800 GTX
min - 21.47
ave - 53.13
max - 129.97
2900 XT
min - 19.67 fps
ave - 53.6 fps
max - 107.1 fps
Company of Heroes built in bench @ 1600x1200 maxed
8800 GTX
min - 47.17 fps
ave - 88.03 fps
max - 217.23 fps
2900 XT
min - 47.3 fps
ave - 103.17 fps
max - 209.67 fps


F.E.A.R. built in bench @ 2560x1600 everything maxed except soft shadows (4xAA/16xAF)
8800 GTX
min - 20 fps
ave - 36.33 fps
max - 86.67 fps
2900 XT
min - 12.33 fps
ave - 36.33 fps
max - 72.67 fps
F.E.A.R. built in bench @ 2560x1600 everything maxed incl. soft shadows (0xAA/16xAF)
8800 GTX
min - 22.5 fps
ave - 40 fps
max - 74.5 fps
2900 XT
min - 25 fps
ave - 39 fps
max - 67 fps
F.E.A.R. built in bench @ 1600x1200 everything maxed except soft shadows (4xAA/16xAF)
8800 GTX
min - 40.25 fps
ave - 71.75 fps
max - 137.25 fps
2900 XT
min - 23.25 fps
ave - 66.5 fps
max - 133 fps


CS: Source built in bench @ 2560x1600 everything maxed (4xMSAA/16xAF)
8800 GTX - 162.68 fps average
2900 XT - 102.03 fps average
CS: Source built in bench @ 2560x1600 everything maxed (0xMSAA/16xAF)
8800 GTX - 240.26 fps average
2900 XT - 178.21 fps average
CS: Source built in bench @ 1600x1200 everything maxed (4xMSAA/16xAF)
8800 GTX - 224.17 fps average
2900 XT -158.92 fps average


Painkiller built in timedemo 1 @ 2560x1600 maxed (4x MSAA)
8800 GTX
min - 5.37 fps
ave - 216.49 fps
max - 500.27 fps
2900 XT
min - 26.36 fps
ave - 147.54 fps
max - 501.55 fps
Painkiller built in timedemo 1 @ 2560x1600 maxed (0x MSAA)
8800 GTX
min - 22.38 fps
ave - 246.75 fps
max - 512 fps
2900 XT
min - 31.94 fps
ave - 246.23 fps
max - 502.84 fps
Painkiller built in timedemo 1 @ 1600x1200 maxed (4x MSAA)
8800 GTX
min - 10.7 fps
ave - 253.1 fps
max - 502.2 fps
2900 XT
min - 37.04 fps
ave - 237.73 fps
max - 512 fps

Painkiller built in timedemo 2 @ 2560x1600 maxed (4x MSAA)
8800 GTX
min - 1.8 fps
ave - 167.74 fps
max - 502.2 fps
2900 XT
min - 11.44 fps
ave - 162.26 fps
max - 512 fps
Painkiller built in timedemo 2 @ 2560x1600 maxed (0x MSAA)
8800 GTX
min - 8.67 fps
ave - 216.90 fps
max - 768 fps
2900 XT
min - 12.54 fps
ave - 211.81 fps
max - 853.33 fps
Painkiller built in timedemo 2 @ 1600x1200 maxed (4x MSAA)
8800 GTX
min - 8.67 fps
ave - 215.12 fps
max - 768 fps
2900 XT
min - 12.24 fps
ave - 216.92 fps
max - 1024 fps


Prey HOC bench 1 @ 2560x1600 maxed w/ boost enabled (4xAA/16xAF)
8800 GTX - 48.95 fps average
2900 XT - 45.3 fps average
Prey HOC bench 1 @ 2560x1600 maxed w/ boost enabled (0xAA/16xAF)
8800 GTX - 76.5 fps average
2900 XT - 76.27 fps average
Prey HOC bench 1 @ 1600x1200 maxed w/ boost enabled (4xAA/16xAF)
8800 GTX - 81.07 fps average
2900 XT - 79.4 fps average

Prey HOC bench 2 @ 2560x1600 maxed w/ boost enabled (4xAA/16xAF)
8800 GTX - 42.85 fps average
2900 XT - 42.15 fps average
Prey HOC bench 2 @ 2560x1600 maxed w/ boost enabled (0xAA/16xAF)
8800 GTX - 66.03 fps average
2900 XT - 62.73 fps average
Prey HOC bench 2 @ 1600x1200 maxed w/ boost enabled (4xAA/16xAF)
8800 GTX - 70.37 fps average
2900 XT - 70.05 fps average


Doom 3 built in timedemo @ 2560x1600 everything maxed (4xAA)
8800 GTX - 54.75 fps average
2900 XT - 66.45 fps average
Doom 3 built in timedemo @ 2560x1600 everything maxed (0xAA)
8800 GTX - 112.4 fps average
2900 XT - 104.4 fps average
Doom 3 built in timedemo @ 1600x1200 everything maxed (4xAA)
8800 GTX - 102 fps average
2900 XT - 124.5 fps average


S.T.A.L.K.E.R. "buildings_timedemo" flyby (Fraps) @ 2560x1600 everything maxed
8800 GTX
min - 32.67 fps
ave - 63.506 fps
max - 217.67 fps
2900 XT
min - 21.33 fps
ave - 47.04 fps
max - 193.33 fps
S.T.A.L.K.E.R. "buildings_timedemo" flyby (Fraps) @ 2560x1600 everything maxed no AA
8800 GTX
min - 32.67 fps
ave - 63.69 fps
max - 219.67 fps
2900 XT
min - 21.33 fps
ave - 46.62 fps
max - 196 fps
S.T.A.L.K.E.R. "buildings_timedemo" flyby (Fraps) @ 1600x1200 everything maxed
8800 GTX
min - 53.33 fps
ave - 89.65 fps
max - 233.67 fps
2900 XT
min - 38.33
ave - 78.05
max - 316.67

Quake 4 HOC bench @ 2560x1600 maxed w/ Ultra settings (4xAA/16xAF)
8800 GTX - 107.33 fps average
2900 XT - 97 fps average
Quake 4 HOC bench @ 2560x1600 maxed w/ Ultra settings (0xAA/16xAF)
8800 GTX - 113 fps average
2900 XT - 103.33 fps average
Quake 4 HOC bench @ 1600x1200 maxed w/ Ultra settings (4xAA/16xAF)
8800 GTX - 108.67 fps average
2900 XT - 107.5 fps average


UT2k4 Umark ONS-AlenjaForest-SE @ 1600x1200 maxed (8xAA/16xAF)
8800 GTX
min - 18.69 fps
ave - 74.71 fps
max - 176.29 fps
2900 XT
min - 16.82 fps
ave - 74.01 fps
max - 166.54 fps


Oblivion (Fraps indoor run) @ 2560x1600 everything maxed (4xAA + HDR)
8800 GTX
min - 33 fps
ave - 50.54 fps
max - 86 fps
2900 XT
min - 30.33 fps
ave - 43.2 fps
max - 77.67 fps
Oblivion (Fraps indoor run) @ 2560x1600 everything maxed (4xAA - no HDR/Bloom)
8800 GTX
min - 36 fps
ave - 54.78 fps
max - 105.67 fps
2900 XT
min - 33.33 fps
ave - 48.71 fps
max - 87 fps
Oblivion (Fraps indoor run) @ 2560x1600 everything maxed (0xAA + HDR)
8800 GTX
min - 35.33 fps
ave - 61.76 fps
max - 112.67 fps
2900 XT
min - 33.33 fps
ave - 53.21 fps
max - 99.33 fps
Oblivion (Fraps indoor run) @ 1600x1200 everything maxed (4xAA + HDR)
8800 GTX
min - 33.67
ave - 68.18
max - 134.67
2900 XT
min - 34
ave - 60.41
max - 125


Infernal (Fraps outdoor run) @ 2560x1600 everything maxed
8800 GTX
min - 26.67
ave - 36.81
max - 50.33
2900 XT
min - 15.67 fps
ave - 20.15 fps
max - 27 fps
Infernal (Fraps outdoor run) @ 2560x1600 everything maxed no AA
8800 GTX
min - 81.33 fps
ave - 111.07 fps
max - 150 fps
2900 XT
min - 44 fps
ave - 53.98 fps
max - 68.5 fps
Infernal (Fraps outdoor run) @ 1600x1200 everything maxed
8800 GTX
min - 59 fps
ave - 86.09 fps
max - 128.33 fps
2900 XT
min - 30 fps
ave - 37.01 fps
max - 43.5 fps

Call of Juarez DX10 demo doesn't run w/o BSODs on the 8800, & it artifacts on the 2900 @ 825/1050, so i have to lower the core speed to around 780.
Oddly enough, no other games have any problems with the HD 2900 XT @ 825/1050.

 

[n7]

7/30/07 update:
For those asking which card i decided to keep, it was the 8800 GTX.

There's not much question that right now for the majority of games, the 8800 GTX is a faster card overall.

The leaked then pulled 163.11 beta drivers fixed the main issues that were bothering me with the 8800 GTX.

Haven't had the "driver stopped" errors anymore, & UT2k4 doesn't have mini lockups anymore, so my main driver issues on the 8800 GTX were solved, making the decision of which card to keep easier.


7/7/07 update:
Benchs @ 2560x1600 w/ no AA have been added.
Benchs @ 1600x1200 4xAA/16xAF have beed added as per requests of benches at a lower resolution.

Pics of artifacting the HD 2900 XT 1 GB displays somewhat frequently randomly when hovering over text:
hmm
hmm2

Originally posted by: nitromullet
A couple of follow-ups on this thread.

That was a couple of days ago now that you posted that, and those are pretty harsh words about two of the three fastest video cards out at the moment. My question is this: where do you stand now with the cards? Which do you like better? What do you like/dislike about both of them?

I am still not happy with either.

8800 GTX pros:
- quiet cooler up to around 85% fan speed
- great performance

8800 GTX cons:
- with default driver settings, idles extremely hot
- random crashing during games due to driver stopped errors
- mini lockups in UT2k4
- BSODs when running Call of Juarez demo
- overall sh!tty Vista 64 drivers

HD 2900 XT 1 GB pros:
- cooler is very quiet idle (not audible above my other case fans, PSU fan, etc.)
- very stable Vista 64 drivers
- actually idles cooler than the 8800 GTX

HD 2900 XT 1 GB cons:
- cooler gets louder than ideal during games
- poor performance considering i paid about the same as for the 8800 GTX
- annoying weird black square artifact that occurs randomly when hoving over text (pics above)
- didn't come overclocked, as advertised (this has nothing to do with running 6-pin or 8-pin)

Notes:
- HD 2900 XT cooler isn't as bad for noise as i expected; i can tolerate the noise under load since i wear headphones when gaming anway.
- HD 2900 XT fan is insanely loud @ 100%. I believe Rivatuner reads it at 10000 RPM :Q
It truely is a jet engine at max
- Overall, i'd say temps are actually better with the HD 2900 XT; idle is slightly lower; load is slightly higher.
Cooler capability is drastically better with the 2900 though, as with fan speed higher, it can run far cooler than the 8800 (if you can tolerate the noise...i can't).
- Drivers in Vista are far better on the 2900.

I'd definitely not recommend an HD 2900 XT 1 GB to anyone running XP, unless you can get it for much less than the 8800 GTX.

As i've made quite clear, i don't like either card TBH...both are far overpriced considering the issues they have.

Which card will i keep?
I will likely be keeping the HD 2900 XT 1 GB, simply due the fact that i don't have any desire to downgrade to XP, & the 8800 GTX is just not reliable in Vista.
But the poor performance it has compared to the 8800 GTX is making the decision extremely difficult.
That's a simple answer to a question that's actually far more complicated for me.


Posted 6/29/07:
Diamond HD 2900 XT pics!

What came in the OEM packaging...
Soo sleek
It's real!

8800 GTX & 2900 XT w/ their backsides up :Q
Purty...
Inside teh case.


Since there's always the noise comments, i will say that the HD 2900 XT 1 GB is extremely quiet when idle (can't hear it over my case fans, etc.).
Under load, it's certainly louder than the 8800 GTX @ 100% even, but i wouldn't say it's unbearable at all.

I found the X1800/X1900 XT coolers much louder & annoying than i do the HD 2900's, & while it's certainly not as quiet as it could be, it's only during gaming it gets louder, which isn't too bad.

The cooler itself is very heavy, even heavier than the larger 8800 GTX's!
Looks good & feels solid though.


Performance from the card is very disappointing quite frankly, extremely disappointing.

Only postives @ 2560x1600 vs the 8800 GTX were:
Doom 3, oddly enough, where it clearly wins.
Painkiller it was also alot smoother too, since the 8800 GTX was stuttering a bunch for some reason, as the very low fps indicate.
UT2k4 = not much difference between either card.

It seems that overall, the 2900 XT made the stongest showing in OpenGL titles...certainly things have changed with this generation of video cards compared to last...


The most frustrating thing of all is actually the fact that card didn't come as advertised though, & since i don't feel like reiterating, i'll quote myself from a previous post below:

Originally posted by: n7
Not too happy right now...

I have the HD 2900 XT installed (i sneaked pics into the second post)

But it doesn't come clocked at the advertised 825/1050 :|

The really frustrating thing is that i basically expected it to be wrong clock speeds, since Diamond has 743/1100 listed, reviews of said card have 743/1000 listed, & NCIX was listing 825/1050.

Against better judgement, i decided to trust their numbers, & heck, even the damn sticker on the card reads 825/2100 :frown:
But the actual 3D clocks speed?

743/1000 :|

So i left a pretty unhappy post on NCIX's forums about the card.

I discovered AMD GPU Clock Tool does nicely work for manual OCing, so i have "OCed" the card from its 743/1000 to the advertised 825/1050, but i don't know of any tools that will work for setting that speed upon bootup, like Rivatuner did so well for the 8800.
ATi Tool doesn't work at all (doesn't recognize card at all), Rivatuner sees card but doesn't let me do anything with it.

Only other app i know of is ATI Tray Tools, which i didn't like last time i used.

I guess the CCC overclocking is unlocked if i get the 8-pin PCI-e adaptor?

Still though, it's stupid i have to OC to get the rated speed :frown:

Many of you might be wondering why in the world i bought an 8800 GTX, & now an HD 2900 XT 1 GB.

It's mainly because i have not been happy with nVidia's drivers in Vista, & i really do not wish to revert back to XP.
I've had "driver stopped" errors infrequently, & random semi-lockups in my favorite game, UT2k4.

I also really thought the faster clocked 1 GB version of the 2900 XT could give the 8800 GTX a run for it's money, but it's pretty obvious that's simply not the case.

So now i have to actually play some games on this HD 2900 XT to see if it exihibits any unbecoming behavior in order to choose which card to keep.

TBH, i was pretty sure i'd be happy enough with the 2900 it'd be replacing my 8800 GTX, but the incorrect default clocks coupled with poorer performance than i even expected is going to make my decision ridiculously hard.

I basically have to choose between nagging issues due to bad drivers & poor performance & extra noise, amongst a bunch of other things

I hope those of you who read thru the benchs & this gleaned something of help

 

[tuteja1986]

lol your an mature benchmarker : )

apoppin and keysplayr2003 are still trying to get all their 1st gpu benchmarking done.

Hey you should also do IQ testing with HL 2
ATI 8x AA vs Nvidia 8x AA vs Nvidia 8x QAA vs ATI CFAA 6x vs Nvidia 16x QAA , Nvidia 16xAA vs AT

 

[Cookie Monster]

Now this is something. Cant wait for your results!

So no TRAA, are you going to use some CSAA modes?

 

[dug777]

Looking forward to it

 

[lavaheadache]

I think these Benchmarking threads are great, but................ I would hold off on starting them till you are ready to post all the data due to people (like myself) filling the thread up with crap before the data is posted. Dare I call this thread a paper launch??? :laugh:

 

[manowar821]

I love you guys for doing these benchmarks.

Just out of curiosity, do you have a copy of Supreme Commander to test...?

 

[n7]

I get kinda confused by all the different kinds of AA settings these days.
Since i don't personally notice a huge difference between 4x AA or anything higher, & my main concern is playability @ 2560x1600, i don't know if i want to do anything higher.

But it'd help me if you guys could clarify what the matching terms on nV & AMD are though, & i can try comparisons?
TRAA on nV = Adaptive on AMD, right?
What about the other modes you guys are asking about, what is the comparable mode on AMD?

Originally posted by: manowar821
Just out of curiosity, do you have a copy of Supreme Commander to test...?

I don't, sorry I'll ask a buddy if he has that one...maybe i can "borrow" it just for benching for you.
You'll notice alot of the games i'm benching are older, sorry.
I'm kinda slow that way...don't tend to get to the newest games for a little while, & also, i haven't really found many newer games i like either...

 

[apoppin]

Originally posted by: n7
2900's results will be added when i receive the 2900, & then get time to bench it.

I also want to add Call of Juarex DX10 results, but thus far, i've only received BSODs toward the end of the bench :frown:

Oh, & if anyone has ideas for benching (with consistent results) Oblivion, Flat Out 2, Tomb Raider: Legend/Anniversary, please let me know.

Thanx.

OMG, you are the benchmarking master!

did you set up on a clean install of Vista

and you gave me GREAT ideas ... i will also show Painkiller benchs ... and MAYBE buy Call of Juarez IF it is a good game

i love you guys but i am not tossing $50 bucks in the trash for something i will never play


and tuteja1986, cut me some slack ! :|
this is my first "published" benchmarking ... i have little experience with 'professional' HW benching [except audio] and am taking my sweet time to make SURE i get everything right

and you have a GTX, right ... you are getting a HD2900xt, right?

why not join us ? you seem to know so much about IQ ... you DO these demanding tests
:roll:

or be quiet please
... new build, new HW, defective HW, brand new Vista and XP User ... and Gosh-Darn Dial-up ... i should have results AFTER i get my GTS ... i don't want to post "partial" bits and pieces of my 'review' ...

PLUS, unlike this one, i would have "reserved" the first 4 posts for me ... like Keys did

i have a 'feeling' that these are also 'preliminary' by n7 ... we will see it all on the same website about the same time

 

[n7]

apoppon, no clean install.

I believe i can clean up the driver in between video cards adequately to avoid issues, & honestly, i simply don't have the time.
Obviously, if there are problems w/ the 2900 XT, i'll have to

Can anyone who has the Supreme Commander demo test this benchmark method?
http://www.extremetech.com/art.../0,1697,2107341,00.asp

I am not downloading the 1 GB demo just to find out it only works on the retail game.


My HD 2900 XT is already on teh truck for delivery, but i won't be home, as i'm off to work, so i may not be able to get the card till tomorrow night.

 

[apoppin]

Originally posted by: n7
apoppon, no clean install.

I believe i can clean up the driver in between video cards adequately to avoid issues, & honestly, i simply don't have the time.
Obviously, if there are problems w/ the 2900 XT, i'll have to

Can anyone who has the Supreme Commander demo test this benchmark method?
http://www.extremetech.com/art.../0,1697,2107341,00.asp

I am not downloading the 1 GB demo just to find out it only works on the retail game.


My HD 2900 XT is already on teh truck for delivery, but i won't be home, as i'm off to work, so i may not be able to get the card till tomorrow night.

1GB demo ... that'd take me a week

if you do a clean install for HD2900xt, why not just create another couple of partitions? ... easy to transfer between them

and in my testing, i will have NO antivirus, no FF or any other proggies running ... just XP SP2 and/or Vista with the latest drivers, games/demos and benchmarking tools

My GTS is still scheduled for delivery on Friday ... that is when i get down to serious benchmarking .. since i will forgo lost Coast for now, i will set up Vista and XP tonight on their respective partitions in preparation for the weekend 'face off'
--my rig only has room for one of these cards
too bad i can't sli 'em

 

[Cookie Monster]

n7, here we go

ATi has 2xAA, 4xAA, 8xAa (box filter/your traditional run of the mill MSAA).
nVIDIA has 2xAA, 4xAA and 8xQ (which is 8xMSAA).

So you can compare those, but comparing e.g 16xCSAA to 16xCFAA is orange vs apple kind of comparison. I mean for instance, would you prefer 8xQ to 16xCSAA? the latter can clean more jaggies with less performance hit, while the former is more consistent with cleaning the jaggies with more performance hit.

Yes TRAA = AAA since they are for reducing jaggies on alpha textures, except that TR MSAA is abit inferior to Performance AAA, while TR SSAA is abit better than Quality AAA unless its been changed this round. (Someone could do some investigation here if they got time.)

 

[Sylvanas]

Love these informative threads

If you are comparing AA modes between the two, doesn't AMD have Gamma correct AA turned OFF in the drivers by default, and NV have it turned on? ATs' G80 review does a comparison of Gamma correct AA if I remember correctly and it does make a difference...for the worse. Could you Include any AA tests with it enabled and not enabled?

If you guys have the time of course keep it up!

 

[aCynic2]

<div class="FTQUOTE"><begin quote>Originally posted by:
F.E.A.R. built in bench @ 2560x1600 everything maxed except soft shadows (4xAA/16xAF)
8800 GTX
min - 20 fps
ave - 36.33 fps
max - 86.67 fps
2900 XT
min -
ave -
max -
</end quote></div>

Hmmm...is that an optimized resolution? I could use 1680x1050 on my machine (I have a BFG 8800GTX), but it's not optimal. I can't recall what I'm using right now, but the built-in benchmarkerhad 100% > 40fps. Avg was something like 128fps. I'll post my parameters when I got home tomrrow morning.

 

[BFG10K]

If you guys are doing benchmarks it's really important to do them with AA because this is where the 2900 seems to tank in performance compared to the G80.

doesn't AMD have Gamma correct AA turned OFF in the drivers by default,

Nope, ATi's had it on by default since the R3xx series.

 

[n7]

Originally posted by: aCynic2
<div class="FTQUOTE"><begin quote>Originally posted by:
F.E.A.R. built in bench @ 2560x1600 everything maxed except soft shadows (4xAA/16xAF)
8800 GTX
min - 20 fps
ave - 36.33 fps
max - 86.67 fps
2900 XT
min -
ave -
max -
</end quote></div>

Hmmm...is that an optimized resolution? I could use 1680x1050 on my machine (I have a BFG 8800GTX), but it's not optimal. I can't recall what I'm using right now, but the built-in benchmarkerhad 100% > 40fps. Avg was something like 128fps. I'll post my parameters when I got home tomrrow morning.

Not sure what you mean, sorry.

I have every setting maxed except soft shadows (they're turned off), w/ 4x AA & 16x AF.

Are you trying to compare your 1680x1050 to my 2560x1600? ?
2560x1600 puts a pretty big hurt on games compared to 1680x1050...

Originally posted by: BFG10K
If you guys are doing benchmarks it's really important to do them with AA because this is where the 2900 seems to tank in performance compared to the G80.

<div class="FTQUOTE"><begin quote>doesn't AMD have Gamma correct AA turned OFF in the drivers by default,</end quote></div>
Nope, ATi's had it on by default since the R3xx series.

Good.
I don't believe i touched anything in the nV driver except for turning quality to high quality (since i effing hate that they don't use that as default).

You'll notice every single bench i did is using at least 4x AA.
I don't mention what level of AA CoH is using, since i don't know; i just set AA to "Enabled"
For STALKER, i once again don't know what level it's using, i just dragged the slider to the max

No AA = unhappy n7.

I don't really think i'm going to have time for doing extensive testing with all the fancier higher versions of AA, but we'll see.

I'm just annoyed i didn't get the card tonight...stupid Purolator doesn't know where it is, & can't do anything for me till i phone then again tomorrow :frown:

 

[aCynic2]

<div class="FTQUOTE"><begin quote>Originally posted by:
F.E.A.R. built in bench @ 2560x1600 everything maxed except soft shadows (4xAA/16xAF)
8800 GTX
min - 20 fps
ave - 36.33 fps
max - 86.67 fps
2900 XT
min -
ave -
max -
</end quote></div>

Hmmm...is that an optimized resolution? I could use 1680x1050 on my machine (I have a BFG 8800GTX), but it's not optimal. I can't recall what I'm using right now, but the built-in benchmarkerhad 100% > 40fps. Avg was something like 128fps. I'll post my parameters when I got home tomrrow morning.


</end quote></div>
</end quote></div>

Not sure what you mean, sorry.

I have every setting maxed except soft shadows (they're turned off), w/ 4x AA & 16x AF.

Are you trying to compare your 1680x1050 to my 2560x1600? ?
2560x1600 puts a pretty big hurt on games compared to 1680x1050...

</end quote></div>

Well, my monitor won't go above 1680x1050, but FEAR doesn't run at 1680x1050. I think 1600x1050 was the highest rez, but that was not an optimized rez.

But, I think in fairness then, you should use the highest resolution you can get, but also the highest optimized resolution.

There are some resolutions for which FEAR is not optimized to use. It can still run at that, but it's not optimal. They usually have an asterisk next to them. Since choosing an optimized resolution provides a better performance, choosing the highest optimized should give the best performace/picture factor. I agree the test should be done with both cards at the same rez.

You can do as you please, but not only should the parameters of the benchmark should be stated, but whether or not those parameters are the best picture/performance ratio.

That old saying comes to mind: There are lies, damned lies and then there are benchmarks/surveys.

 

[tuteja1986]

n7 i want you to do a clean install otherwise your review has very little to stand on

 

[n7]

Ah, i see what you're saying.

There's nothing wrong with benching at 2560x1600; it's a standard resolution that games can run at, whether it be via .ini tweaks, console haxery, or default game options.

Now if i was benching at 2048x1280 (not a standard resolution), then yes, you'd be right.

But you might wanna look at some more benches
2560x1600 is likely more standard than 1680x1050 (which is a standard one also btw)

 

[n7]

<div class="FTQUOTE"><begin quote>Originally posted by: tuteja1986
n7 i want you to do a clean install otherwise your review has very little to stand on </end quote></div>

I'm okay with that
It's very rare than performing clean installs will magically improve performance.
You're more than welcome to read official reviews instead, but no matter what review you read, a level of trust is required, since really, how do you know they did a clean install?

I'm doing this little review for myself primarily, & secondly to give those of you who want to see two high end cards benched from a simple user's point of view, rather than an "official reviewer's."
Not to mention that i've yet to see any reviews other than one with this card, & that one is running slower clocks than the one i'm receiving is supposed to have.

 

[BFG10K]

n7 i want you to do a clean install otherwise your review has very little to stand on

I never do clean installs since I've never seen it make any difference.

 

[dug777]

Originally posted by: BFG10K
<div class="FTQUOTE"><begin quote>n7 i want you to do a clean install otherwise your review has very little to stand on</end quote></div>
I never do clean installs since I've never seen it make any difference.

2nd.

 

[apoppin]

Originally posted by: BFG10K

n7 i want you to do a clean install otherwise your review has very little to stand on

I never do clean installs since I've never seen it make any difference.

i can't say one way or another ... if the rig is stable, it is stable ... however, i would definitely suggest defragging the HDs and limiting background apps.

For me .. i must do a clean install ... it's just me ... perhaps i don't really care about the extra time it takes ... as long as i get the bulk of it done this weekend ... with the conclusions and retesting by/on the 4th [which i also have off ... not so sure yet about Thurs and Friday. ... and i am pretty sure i have the 'hang' of running benches pretty well - at least i am getting very consistent results.

i am nearly done, my rig is "right on" ... but NewEgg won't be sending my Crucial Ballistix PC8500 [i can 'make do' with the PC6400 XMS ]
--ANYWAY ... FC is set up and fully patched, STALKER, PREY and FEAR also [i guess i could also compare the Extraction Point bench]

i got PainKiller fully updated .. thankfully i had all but 115MB of it already in an earlier patch on DVD ... and i have most of Q4 already - patched to 1.3, already saved. ... and it even appears that Lost Coast may have enough files to be close to launch ... i am getting the "preparing for launch" message ... in my experience that is only overnight ... i will have Q4 and LC probably by tomorrow

Now What About Call of Juarez? is it a *good* FPS? Worth Paying $50 for? ... of course it would take till the 4 for me to patch it , so LMK right away so i can get it TODAY!!!!!!!!!

 

[tuteja1986]

Originally posted by: BFG10K
<div class="FTQUOTE"><begin quote>n7 i want you to do a clean install otherwise your review has very little to stand on</end quote></div>
I never do clean installs since I've never seen it make any difference.

But didn't you do just replace with another nvidia card or have done with ATI card ?

 

[BFG10K]

But didn't you do just replace with another nvidia card or have done with ATI card ?

I used to frequently jump backwards and forwards between my X800 XL and 7800 GT.