MSI 6800 "Vanilla" User Impressions

jhtevans

Limp Gawd
Joined
Dec 14, 2002
Messages
374
Well, I bit the bullet and purchased an Nvidia based gfx card. My recent gaming performance was falling somewhat short of my expectations. My 9800Pro ran games like Far Cry and Doom 3, but it did not run them well enough IMO. However, I decided with this generation that I didn't want to buy myself into a corner, so to speak. My next platform upgrade will be in the spring and I'll be making the leap to PCIe at that time. Thus, I didn't want to invest a large amound of money into more AGP when I'll need to buy a top of the line PCIe card in about 6 months.

The solution? I initially preordered an eVGA 6800 Ultra. The buyers remorse was too much, however, and I cancelled my preorder a few days before it would've shipped. As painful as that was, I couldn't bear to spend $500 on a video card that I won't even be able to use on my system in another year. Based on the results and user opinions as well as my budget constraints, I purchased the MSI NX6800 which is their take on the "vanilla" 6800 card. I purchased the card from NewEgg for $290. Much easier on the wallet. The is the first "bang for the buck" card I've purchased since my original TNT and I've gotta say, it feels good. My upgrade progression looks a little something like this:

TNT--> GeForce DDR --> GeForce 3 --> GeForce Ti4600 --> 9800Pro --> 6800

The MSI card appealed to me because of the bundle, cooling solution, and overall image of the company. The game comes with an insane 17 games and a few applications. Granted, most of the games suck and people don't usually buy a video card just to get the games, but it's still nice to see a manufacturer put some value-added material with the card. At a $290 pricepoint which is still under the MSRP, why not buy a card that has a little extra?

As I mentioned, the cooling was also another selling point. The MSI card comes equipped with a solid copper heatsink with speed-controllable fan that reminds me of the 9800XT cooling. The fan has a speed slider that adjusts from 2800 - 4000 RPMS (30-39dBA respectively). At high speed, the card is audible over the CPU fan, but not so much with the case closed. Based on other posts, I'd say that the MSI 6800 runs slightly cooler than other branded cards using the reference cooling.

At stock speeds, the MSI NX6800 runs at around 51 C idle. This rises to a little of the 60 C mark after about 20 minutes of Doom 3. I have the fan speed set to the max speed.

After installing coolbits2, I detected my optimal frequencies as 381/811 (stock is 325/700). At these speeds, the cards runs at about 55 C idle and doesn't rise above 67 after another 20 minutes of Doom 3 goodness. I'd consider this a fairly nice overclock, especially without having to do anything to the card or modify the cooling in any way.

Real-World Gameplay.....

65.62 Beta ForceWare
High-Quality Image set via ForceWare, no optimizations enabled

Since everyone agrees that the 6800 series was made to play Doom 3, it comes as little surprise that the MSI NX6800 shines in this games. My 9800pro had trouble maintaining 30 fps (1280x1024, medium quality, 2xAA). The 6800 blazes ahead managing 60+ FPS the majority of the time (1280x1024, high quality, 2xAA) dipping into the 50s only on occasion. This is obviously cause for celebration because Doom 3 is all of a sudden a different game, and it's not on account of the visual splendor. The game is no longer as herky-jerky and monster movement is now much more fluid. I felt like the game looked and played like it was suppose to on my 9800 pro, but I was wrong. Moving to the 6800 definately brings things to a new level.

Haven't tried any other games yet, but as I do, I will continue to edit my post with new info.

Canned Benchmarks......

For those who like synthetic benchmarking, this section is for you. While I don't believe that 3DMarks tests represent 'real' gameplay, I believe there is value in using their benchmarks as a diagnostic tool. It's an especially quick and easy way to test for problems in a system. So, on with the numbers:

3DMark 2001 SE - 19065
3DMark 2003 - 10090

The 3DMark01 score surprised me. My 9800 Pro scored a 19184. 3DMark01 is an older test and doesn't show off the newer capabilities of the 6800 series and a reformat may alleviate the discrepency. For now, I'll just assume that the 9800 Pro's specs give it a slight advantage in the older benchmark. As for the 03 test, I'm pretty sure this falls right in line where it should. This score represents the card running at the overlocked frequencies.

After a day of using the card, I can honestly say that it was money well spent. If you've got an older video card in your rig and you're not looking to do any major upgrading until PCIe becomes more mainstream, then the 6800 "Vanilla" is for you. I'd say that it keeps pace nicely with most systems and doesn't cause everything else to be the bottleneck. Bottom line is that it's more than enough power 'right now' and it's not too much power either. A year down the line, the 6800 'vanilla' is likely to be severely handicapped by it's 12 pipes and 128 MB of DDR. The caveat to that is the eventual upgrade. If you plan on updgrading sometime next year, then the 6800 is a nice "low(er) cost" solution that should get you from here to there with relative ease without wasting too much money in the process.

Thoughts? Questions? Comments? Please post them.
 
Ugly_Jim said:
nice post. almost convinced me to buy a MSI 6800 :)

heh - Well, I figured there would be a number of people in a similar situation just wondering how much better a 6800 is than a 9800 Pro. Not having experienced anything in the middle, it seems like a minor leap in terms of performance. Image quality is quite good as well, dare I say slightly better than the 9800 Pro. Just providing the info in hopes that it'll help someone, somewhere, somehow.
 
No, it helped. The only thing preventing me from replacing my card is the fact that I have no job/money at the moment. :(
 
What are your system specs?

Never mind, I clicked on you name and got them.

I have an Athlon XP 3000+, 1 GB Corsair XMS PC3200, MSI RX9800PRO-TD128, Dell 2001FP, and have been wondering if I would notice a significant improvement with a vanilla 6800 myself.


Kirk
 
I will send your impressions to a few friends looking into a "plain" 6800. I got the GT so I could not really tell them to get a 6800 over an x800 or a 9800xt without an unbiased user opinion. Thanks for your thoughts.:)
 
The only one that I know of that has 256 is the ASUS card, which really isn't a viable option for most people because at this point, because it's more expensive than the GTs....
 
ugh...cant decide...this or GT...ugh.

If i found a NU with 256 ram for ~300 id buy it right there.
 
grats, 3dmark01 aint too useful its a cpu benchmark mainly now (cpu limited), so its ok not to see any difference in scores.
 
I decided on the BFG 6800GT since it was $100 more and it was the first one I found on the shelf at BB. :D I don't plan on upgrading for another 2 years. I may upgrade sooner if I get my hands on good used stuff. I figure there will be a plethora of inexpensive used i875/agp parts available in a year or so.

If I could have got my hands on a MSI 6800GT I would have. My last 3 cards were MSI's, and I liked them. If I want to do any serious overclocking, this BFG card will need a better HSF (I got the "custom fan" edition). It runs HOT

I've upgraded to the fastest cpu available for my motherboard (3.06Ghz P4) and also upgraded to 1Gb ram recently. This upgrade will need to last me anther 2 years. The only thing I would still like to get for my system is a new Seagate 15K.3 scsi drive to replace my 10K.3 scsi boot drive.

I'm also going to put a 120mm blowhole on the side of my case to blow right on my cpu & graphics card. I recently completely broke my computer down and cleaned it. I still haven't put the 4 80mm case fans back in, I have my case panels off right now so it's work in progress. I really like the quietness for now.
 
Great post jhtevans. You just helped me make my decision. I was trying to decide between the msi Radeon 9800 pro and this card. It looks like the msi 6800 is the winner. Just out of curiosity, are you running nvidia's reference drivers or msi's drivers? I'm not a big fan of manufacturer's "repackaged" drivers and was wondering if the msi 6800 works w/ the stock nvidia driver package (the ones you get from nvidia's site). This question may seem silly to some, but a friend of mine has a geforce 5700 by leadtek that won't work w/ the drivers from nvidia's web site. The card only works w/ the driver updates from leadtek's site. Weird huh? I just want to make sure that the msi 6800 doesn't have this problem.
 
I'm using the Nvidia 65.62 Betas - Not official, but they're not "re-packaged" by any manufacturer ---
 
Here are some numbers from Doom 3 Timedemo demo1, each consisting of the second run for each configuration.

High Quality, No AA
1024x768 - 70.3 fps
1280x1024 - 59.9 fps
1600x1200 - 44.6 fps
Extremely playable - the action moves so fast at some points that you can't even make out what's going on.

High Quality, 2xAA
1024x768 - 63.8 fps
1280x1024 - 51.7 fps
1600x1200 - 40.7 fps
Also extremely playable with some really fast action taking place.

High Quality, 4xAA
1024x768 - 53.1 fps
1280x1024 - 37.8 fps
1600x1200 - 3.5 fps (first run)
At 4xAA, the 6800 really shows it's lack of extra memory. 1024x768 is probably the only resolution under this mode that is playable. IMO, 1280x1024 reminded me of the performance of my 9800 Pro @ medium quality, 2xAA. Definately still an improvement, though. Notice that score at 1600x1200? That's right - 3.5!!!! The second run was a bit of an anomoly, scoring in at 40.3 fps - I wasn't able to duplicate this, so I assume it was some sort of fluke. If the 6800 vanilla had 256mb of DDR, I'd say this mode would be attainable.
 
Cali3350 said:
ugh...cant decide...this or GT...ugh.

If i found a NU with 256 ram for ~300 id buy it right there.

Excuse my lack of knowledge on the abbreviations, but what is an NU?

Thanks,
Kirk
 
how does doom 3 run on high quality settings ? does it stutter occasionalily like when opening doors or encountering a lot of demons at once because of the 128 mb of memory ?
 
one thing everyone must keep in mind is that the msi comes with 2.8ns ram whereas the leadtek and chaintech use 2.5 or 2.2. the gainwards also come with 2.5 or 2.2ns ram.

however when it comes to the bundle, msi's teh best.

and btw itll take quite some time for the 68gt's to come to the $300 price range.
 
gepetto said:
how does doom 3 run on high quality settings ? does it stutter occasionalily like when opening doors or encountering a lot of demons at once because of the 128 mb of memory ?

I only get stuttering on the first run through as the demo caches textures into memory - second time though, which is more indicative of in game play, doesn't skip at all. I included those time demo numbers for all of you, but from my experience with the game using a Ti4600, 9800Pro, and the 6800 - The timedemo numbers do not accurately reflect what goes on in the game.....

That said, the games runs great on the 6800, high-quality 1280x1024 2xAA. I could probably very easily play at 4xAA as well. On the 9800pro, medium quality 1280x1024 2xAA was as high as I could go. Anything more and the game was too herky-jerky.
 
Anyone have some numbers they can throw up for comparison? I'd be curious to see some other vanilla overclocking figures as well as some benchies and overall impressions....
 
High-Quality Image set via ForceWare, no optimizations enabled

My opinion - but I wouldn't bother with this. ALL cards (ATI included - yes, your old 9800) use various optimizations. In fact, ATI's cards use pretty much the exact same optimizations you just disabled there - only thing is that, so far, nVidia is the only manufacturer that gives you the option to disable them.

At this point - the GeForce 6-series - this just isn't worth doing. There is NO visual quality difference between having them on or off (at least, not discernable without doing a bit-wise comparison of two screen captures), and they can amount to some substantial performance increases having them on.

Ditto with the "High Quality" vs the "Quality". HQ pretty much just disables some of the optimizations, not much else. I'd bet if you turned the opts back on and used "Quality" rather than "High Quality" and took some before and after screen shots, you'd not be able to tell the difference anywhere other than the much higher FPS in-game.
 
I haven't used ForceWare in a few years so let me sum this up: There are basically 3 optimizations that can be enabled/disabled through Nvidia's drivers.

Trilinear Optimization
Anisotropic mip filter optimization
Anisotropic sample optimization

You're saying that on a 6800 series card such as the 'vanilla', the combination of all three of these optimizations will not noticeably change the visual quality
 
Correct. They are simple the same optimizations everybody uses - nVidia just gives you the option of turning them off. And, hey, if turning one back on DOES do something you don't like, just turn if back off again. That's why it's an option!

As an aside, there IS a registry hack to disable the trilinear optimizations, at least, on ATI cards, but not the other opts. Still, it's a pretty big performance hit - 15% - which makes ATI's position on it interesting:

“Disabling the trilinear optimization will not enable users to choose between performance and quality. Everyone on the web has been absolutely clear: there is no image quality degradation with this optimization. Therefore, a control panel option would not change image quality, only performance. We do not anticipate adding a checkbox to disable the optimization to future control panels. What would we label it? “Reduce performance”? “Lower frame rates”? Because that is all that switching off the trilinear optimization would do,” commented Chris Evenden for ATI Technologies.

Isn't that nice of them?
 
Back
Top