ATI HD 3800 Review Thread

[Zytek_Fan]

Might be, here's what it looks like

That reminds me of the glitch on Morrowind, where if you had 100+ acrobatics and you jumped you would fall through the canton and be stuck

[donitsi]

Fuad reports that 125,000 3800 have sold in the 1st 3 weeks! Impressive if true

Yes really nice if true. The 3800 series should put Ati back to profitability.

[Brother Esau]

Heres mine @ Defaults

[Donnie27]

That reminds me of the glitch on Morrowind, where if you had 100+ acrobatics and you jumped you would fall through the canton and be stuck

The problem is that there are points on each MAP where there was some type of Glitch. I got stuck between some lockers, I dropped a Laptop the wrong way and it killed me and etc....... I like the Game but I was surprised by how Buggy it is.

I kept my 3870 instead of taking it back. CoD4 and the other games I play look pretty good with it.

[ThatGuy16]

I can't wait, my 3870 should be here wednesday. What do you think i'll score in 3DMark06 with sig-rig?

Can't wait for the 45nm release, i plan on E8400 / X38 and might go crossfire

[mascaras]

Powercolor HD3850 256Mb Crossfire Default - Crysis GPU Benchmark



Q6600
Asus P5E
1680x1050
Windows XP



1680x1050 Medium Settings :


TimeDemo Run 1 Finished.
Play Time: 29.71s
Average FPS: 67.31
Min FPS: 46.50 at frame 155
Max FPS: 82.03 at frame 121




1680x1050 High Setings :


TimeDemo Run 2 Finished.
Play Time: 67.17s,
Average FPS: 29.78
Min FPS: 9.31 at frame 141,
Max FPS: 38.75 at frame 987












regards

[Cooper]

There's definitely something wrong with that "crysis benchmarking"
AFAIK all latest runs perform same or bit better

[Unbornchild]

I think that is absolutely not bad, for a HD3850 256Mb Crossfire Default .

I'm glad to see someone posting a real game benchmark again. (instead of the annoyingly useless 3DM06 scores), and showing how relatively strong this 'light' card actually can be in dual setup. Thanks Mascaras.

[DangeR!]

I think I am being bottle necked by my CPU.
See sig for full system speck.

One Card
http://service.futuremark.com/compare?3dm06=4026364

xFire
http://service.futuremark.com/compare?3dm06=4114217

Or do you think there is something wrong with my setup?

[Unbornchild]

Of course you aren't "bottle necked by CPU", see the same CPU score in both benchmarks. EDIT: wait. I just saw you have merely a E2140. Yeah, could be the bottleneck...

Your CF isn't working correctly? Make sure you have default settings in CCC when you use CF.

[madcho]

Q6600 @3ghz => 17500 with gpu and ram @ stock

[Donnie27]

I think that is absolutely not bad, for a HD3850 256Mb Crossfire Default .

I'm glad to see someone posting a real game benchmark again. (instead of the annoyingly useless 3DM06 scores), and showing how relatively strong this 'light' card actually can be in dual setup. Thanks Mascaras.

The problem is not the software, it is the silly miss use of it that's kind crazy.
Most folks with Common sense know the software has a good use. It for checking like set-ups to make sure they're properly set up or tweaked. IMHO, it was never meant as software used for contests LOL! Geeks change this view and FM went along with them instead of re-stating what the h@ll it was meant for Too many it became 3D Bung-Holi-marks because of this *misuse.

Of course the software has no correlation to any Game

[mascaras]

There's definitely something wrong with that "crysis benchmarking"
AFAIK all latest runs perform same or bit better

nothing wrong , not same settings on the Runs

1st run with Medium settings

2nd run with High settings

[DangeR!]

Of course you aren't "bottle necked by CPU", see the same CPU score in both benchmarks. EDIT: wait. I just saw you have merely a E2140. Yeah, could be the bottleneck...

Your CF isn't working correctly? Make sure you have default settings in CCC when you use CF.

I pressed "default" for everything in CCC before I did those runs.

I'm thinking of changing my E2140 in Jan for a Q9450.
If it was a setting, I could get my xFire to work well then I wouldn't have to upgrade the CPU.

[STaRGaZeR]

3870 CF is heavy CPU limited, even a 3,6Ghz quad is not enough.

[Lightman]

nothing wrong , not same settings on the Runs

1st run with Medium settings

2nd run with High settings

Good work Mascaras!

I did Crysis x32/x64 tests @1680x1050 HIGH DX9 (under Win XP x64)

Phenom@2470MHz
Hd3870CF@800/1200

Crysis x32 - Average from 2,3,4 loop 38,9FPS
Crysis x64 - Average from 2,3,4 loop 42,01FPS


Bear in ming that I'm running only 2GB of RAM ATM. With 3GB+ Crysis x64 can load everything into memory, so higher averages and no shuttering during gameplay .

[DangeR!]

3870 CF is heavy CPU limited, even a 3,6Ghz quad is not enough.

So what speed are we looking at, not to bottle the cards?

If I know what speed I can try and save for the CPU .

[STaRGaZeR]

I don´t know man, but when I tried 3870 CF a few days ago there was a huge difference in scores (I mean, in SM2.0 and SM3.0) going from 3,3Ghz to 3,6Ghz. With more Mhz those things would fly!

[mascaras]

I don´t know man, but when I tried 3870 CF a few days ago there was a huge difference in scores (I mean, in SM2.0 and SM3.0) going from 3,3Ghz to 3,6Ghz. With more Mhz those things would fly!

yep , see in my signature , only with +400mhz CPU (3200mhz to 3600mhz) >>> 1000+ marks

Q6600@ 3200mhz + HD3850 256mb Crossfire Default >>>> 16650 Marks

Q6600 @ 3600Mhz + HD3850 256MB Crossfire Default >> 17780 Marks


regards

[DangeR!]

My CPU seems a good clocker so I will aim for 3.6 now.

4GHz may be available lol (wish full thinking I know ) Bring on the water cooling.

[ThatGuy16]

not bad for a single 3870 and X2? all stock but cpu @ 3.33ghz

[DangeR!]

what are your scores?

[ThatGuy16]

what are your scores?

3DMark Score 10307 3DMarks

SM 2.0 Score 4411 Marks

SM 3.0 Score 5051 Marks

CPU Score 2385 Marks


http://service.futuremark.com/orb/re...0&UID=12663270

not bad, right?

[DangeR!]

Thanks alot.

you score is fine mate, here's mine for comparison.

One Card
http://service.futuremark.com/compare?3dm06=4026364
3DMark Score 10703 3DMarks
SM 2.0 Score 4470 Marks
SM 3.0 Score 5032 Marks
CPU Score 2744 Marks


xFire
http://service.futuremark.com/compare?3dm06=4114217
3DMark Score 13581 3DMarks
SM 2.0 Score 5753 Marks
SM 3.0 Score 7386 Marks
CPU Score 2742 Marks

With xFire my CPU is limiting it but I run 1680:1050 and crank up the eye candy to make up fir it .

Bring out the Yorkfields already!!!

[ThatGuy16]

oh yeah i hear ya, im running 1680x1050 too. Single card is almost exact same score!

I plan on picking up a x38 and a E8400 in late jan / early feb