Did you guys happen to try this card out in Crysis? I bought this card and every game except Crysis runs at high settings. Crysis "optimized" by the game asks for High settings but runs @ 10 FPS. On medium settings it does 12 fps, on low settings it does 10 FPS. I'm running a phenom 2 965be @ 3.8ghz, m4a88td-v evo from asus, 12gb gskill 1600 ram, sandforce ssd, etc - no other bottleneck in my system. Any potential help would be nice, I'm wondering if it's the drivers are the problem.

Link is broken in the first post.

Good article. Nice to finally see a decently quiet video card with everything at stock !

Excellent review, the only FAQ that you didn't answer in the article is how it would fit in a Solo (while I'm at it I'll mention that 9.5 inches is normally 9.5" not 9.5')

In addition to my Crysis question, has anyone tried to install a Scythe musashi on one of these? i am thinking of replacing my S1 + noctua fan combo for a musashi.

IMO it would fit in a Solo with plenty of spare room. My 4850 (in Solo) has the same length and I'm eyeing this 6850 to replace it.
If you'll put an S1 on the 6850, non-reference as they are, the GPU is moved to the right by about 1.5 cm, so the S1's right side will just overlap the card's right side.

Fits fine in a Solo -- with a little less than 1" to spare lengthwise.

After looking again at the pictures of the naked cards I noticed the casing of the DVI connector is very likely to prohibit the installation of an S1, at least without maiming.
MikeC, please confirm if this the case or not? This may be an issue with some other aftermarket coolers too.

Probably -- except the ones already modified for these cards. With the Asus, there's no need -- the thing is quiet enough for 99% of users interested in a card of this level of 3D performance.

I do not understand.

In TechPowerUp, that video card was noisy (noisier than here in SPCR): http://www.techpowerup.com/reviews/ASUS ... Direct_Cu/

There he says that does not run its fan under 1500 rpm.

I trust SPCR. Just I would know why that difference.

Thank for review.

I wish when people reviewed these cards they would include which version of the HDMI standard is implemented. Also, it would be great if someone tried using it on a their TV or A/V system.

I think many who read SPCR are interested in HTPC applications so it would be a bonus to have it on this site. The vendors and card manufactureres usually don't even include what version the HDMI port is.

I know the 6000 series cards CAN have HDMI 1.4 implemented which supports full digital sound and 3D, but does each vendor include it? Does the card have the protected audio path (PAP) so that TrueDTS-HD and DolbyHD? I think those are important questions.

I based my system of Anandtech's Build guide in conjunction with SPCR standbys: http://www.anandtech.com/show/4025/holi ... rs-guide/4

I bought the asus 6850 because in the guide it states:



It's definitely the quietest gaming card I've used, probably ever. My friend is an experienced modder and he modified the S1 slightly for my setup... but as I stated before, Mike - do you think it will fit a musashi? In the SFF gaming system article, SPCR recommended the 6850 because it was compatible with more AM heatsinks.

Nice review - thanks!

Typo in a page 4 table:
System Measurements: VGA Test System with
***Rerference*** AMD Radeon HD 6850

It's only an issue on the reference card.

Well, it's hard to talk about what they heard and measured... but our subject assessment is consistent with our measurements. And we saw the same 1500rpm min speed. Maybe they had a sample with a bum fan. Who knows?

arjunr, Now that I have seen. They can be wrong or have a bad sample.

Yes, Mike. I believe this.

Even that Anand and Tech Report found the same as you: That video card is very quiet.

Thanks for reply.

Glad I ordered my Asus 6850 on Saturday...

Once you guys do your reviews, the Newegg prices tend to go up!



Must be a driver problem. Hell, I get better than that on my current 8800GT!

I would advise caution for these cards, a lot of people are having issues playing games, myself included. Crysis was unplayable - super choppy, etc. Crysis warhead plays fine for about 3 minutes and then crashes the machine because the display goes completely static, etc.

Starcraft 2 is the only game that "just works."

However, these problems have been reported by AMD Win 64 users... which I am. It might be a driver issue as well. I kinda wish I bought a nvidia right now... but supposedly ATI is going to be releasing drivers soon.

IIRC these cards were released before the new drivers were ready. The current drivers kind of work but weren't designed for the 6850/6870. New drivers should be out soon.

I agree that Crysis is unplayable because it sends the whole computer into freeze and crash after a few minutes, but all other current generation game I have tried work out of the box with high settings. It is curious that all 6850 reviews upon release talked about being able to play Crysis on high. No-one seemed to have a problem. I wonder what the difference can be between their systems and those of so many users.

Woohoo, just ordered the Asus 6850 also Can't wait to undervolt this little friend too. Thanks for the timely review Mike!

Interesting read, do i cant see the point of using FurMark to do load test as it will give no usable data other then what the card will sound in theory if there would be a program out there that would load the GPU to its maximum load, ore if people start using F@H, ore what would be possible use is (re)encode video, other then that i see no real world use at all for FurMark to be the only load test.

what i would be interested in is power/sound load during
Idle (dose the system make sound when it doing noting)
MPEG2 and VC1(standard DVD and playback not really interesting as its silent on almost all cards)
h264@40mbps (Avatar?)
x264 (high bitrate re-encoded Avatar?)
Gaming WoW ore QuakeIII and Metro 2033 (low/hi-end DX9/11 gaming)

Ware sound levels during video playback would have the main focus of my interest, also a part of my interest would be HD-HQV and HQV test numbers as next to silent video quality is also part of my interest of buying HTPC parts

Not saying FurMark has to be kicked out of the test setup, just saying there are more interesting test that have more real world impact then a theoretical benchmark as FurMark what main focus is to test if a OCed video card is stable.

My 2 cents

player-x --

It sounds like you haven't read our video card reviews carefully enough.

1. idle noise is ALWAYS measured, described.
2. There is no difference in noise between idle and the most demand HD video playback. The GPU load is always quite modest -- and even most integrated GPUs have no trouble w/ HD video these days.
3. Furmark allows us to test for power draw, efficiency and maximum noise (after a long stretch of hard 3D gaming, for example) in a predictable, consistent way. This is the nature of all PC benchmark testing -- push the stuff to the max.
4. We generally don't run gaming benchmarks because it's done ad nauseam by everyone. Why bother? We focus on what we do best.

While we consider it a worst case scenario test, I have to point out that on our SFF gaming build, Furmark heated up the GPU only 4C higher than peak temperatures while playing Crysis Warhead: http://www.silentpcreview.com/Silent_Gaming_System_Build_Guide/Results

If you consider FurMark a theoretical maximum, the amount of headroom isn't big compared to playing Crysis. If we ran Crysis as a load test and the ambient temperature was a few degrees higher, we would get similar thermal results, so presumably a cooler with automated fan control would also generate similar noise levels.

Yeah finally an interesting card to replace my aging Gainward 4850 GS : higher performance than the 4850, lower idle power consumption, nearly the same power consumption at load, silent and not too expensive (bought my 4850@140€).

And the most important : fits into my Antec Solo

Just a followup, I finally got my system put together. Same mobo, but an x3 740BE (unlocked @ 3.2Ghz), DDR3-1600, with a 150GB velociraptor, with Crysis maxed out and it's smooth as butter using just the default 10.10 Catalyst drivers.

On a general note, the fan speed stays constant at 1600-1650rpm (in my system during heavy gaming) and is not audible from 1m. You can use MSI Afterburner to slow down the fan speed further if you need to, though due to the bearing noise and small size of the fan, it's unlikely it will help much at the expense of temps.

Hello,

Just bought this Asus DirectCu after long reflexion : trade my Antec Solo for a new case in order to support longer,more powerful, and more expensive cards, or be reasonable and buy only this 6850

My pc :
Core 2 QuadQ9550, Gigabyte P45 UD4, Enermax Modu 82+ 520, Noctua NH-U12P, Ocz vertex2, Samsung F3 1Tb, Samsung F2 ecogreen 1.5Tb, 3 year old Antec solo with stock fan + 2 noiseblocker Se2 92mm, 22" samsung 1680x1050 lcd.

My PC is not as silent as I would : clicking of my enermax fan, samsung f3 "woosh", antec tricool not slow enough, the noctua fan not enough undervolted(too lazy to look for my fan mate). However this 6850 does not seem to add noise to my configuration.

On the performance side, very happy since I come from a 4850.

Only drawbacks : price (i bought my 4850 at 140€), blocks one hard disk drive slot.

How would you undervolt/underclock this card further than in the review? Say, what would be good engine/memory settings at 1.05v or 1.00v (below reference)?

Mike,how would you say it compares to the hawk msi and the vapour x sapphire noise wise(as far as I know two of the most quiet factory coolers available).The reason why I'm asking is because I have a 5850 vapour x,and I found out here on the forum that the hawk(twin frozen) msi cooler is actually more quiet both on idle and load.I am not looking for an aftermarket cooler,just a very quiet stock one.

Got this card today and its very loud even at idle. Running 1750 RPM at its lowest (doesn't seem to go lower below 25% fan speed at all), even when reduced with MSI afterburner controls.
Do I have a faulty one or is this really how its supposed to be :/

The noise is somewhat high pitched making it even more annoying to the ear.

Temps are 32 C at idle, so its certainly not because its running hot either.