On the example of
BFG GeForce 9800 GTX 512MB PCI-E,
MSI GeForce 9800 GTX 512MB PCI-E (N9800GTX-T2D512),
Zotac GeForce 9800 GTX 512MB PCI-E

Gaming Tests

We used the following benchmarks:

• Call Of Juarez (Techland/Ubisoft) - DirectX 9.0, shaders 3.0 (HDR), demo. Tests were run with maximum quality settings. The batch file is included.
  
  
• S.T.A.L.K.E.R. 1.003 (GSC Game World/THQ) - DirectX 9.0, demo. You should unpack the files into the game folder with savegames (Documents and settings). When the game starts up, load ixbt3, open the console, and start the demo: demo_play ixbt3. Test settings - maximum (DYNAMIC LIGHTS are ENABLED).
  
  
• 3DMark05 1.20 (FutureMark) - DirectX 9.0, multitexturing, trilinear filtering.
  
  
• CRYSIS 1.2 (Crytek/EA), DirectX 9.0c/10.0, shaders 3.0/4.0, (batch file to start the test), test settings in MS Windows Vista - Very High (DX10), Windows XP - High (DX9), RESCUE and HARBOR levels). You can e-mail us to get the timedemo.
  
  We express our thanks to CRYTEK for writing timedemo for iXBT.com.
  
  
• Company Of Heroes (Relic Entertainment/THQ) - DirectX 9.0, shaders 2.0, startup batch file. When you start the game, you should go to options, choose the graphics section, and press the test button. Tests were run with maximum quality.
  
  
• 3DMark06 1.10 (FutureMark) - DirectX 9.0c, multitexturing, trilinear test settings.

If you have a decent understanding of 3D graphics, you may draw your own conclusions for the diagrams below. However, if you are interested in our comments on test results, you may read them after each test. Anything that is important to beginners and those who are new to the world of graphics cards will be explained in detail in the comments.

First of all, you should look through our reference materials on modern graphics cards and their GPUs. Be sure to note the operating frequencies, support for modern technologies (shaders), as well as the pipeline architecture.

ATI RADEON X1300-1600-1800-1900 Reference
NVIDIA GeForce 7300-7600-7800-7900 Reference

If you have just begun realizing how large the selection to choose a video card is, don't worry, our 3D Graphics section offers articles about 3D basics (you will still have to understand them - when you run a game and open its options, you'll see such notions as textures, lighting, etc) as well as reviews of new products. There are presently two companies that manufacture popular GPUs: AMD (its ATI department is responsible for graphics cards) and NVIDIA. We can also mention Matrox, S3. However, their share of the discrete graphics market is less than 1%, so we can discard them. So most of the information is divided into these two sections. We also publish monthly i3DSpeed that sums up all comparisons of graphics cards for various price segments.

Thirdly, have a look at the test results of our today's cards. We are not going to analyze each test in this article, primarily because for us it makes sense to draw a bottom line in the end of the article. We will, however, make sure that we make our readers aware of any special circumstances or extraordinary results.

S.T.A.L.K.E.R.

CRYSIS, Rescue, DX9, High

CRYSIS, Harbor, DX9, High

CRYSIS, Rescue, DX10, Very High

CRYSIS, Harbor, DX10, Very High

Call Of Juarez

Company Of Heroes

3DMark05: MARKS

3DMark06: SHADER 2.0 MARKS

3DMark06: SHADER 3.0 MARKS

Conclusions

What can we cay about test results? - Everything is logical. The 9800GTX is expectably faster than 8800 GTS 512, its performance gain corresponds to the increase in operating frequencies. As the product has lower memory bandwidth and less memory than 8800 Ultra, the new card is outperformed by the old product (which is already leaving the market) in heavy modes with AA in high resolutions. However, the lag is not that big to worry about.

That is as in case of 8800 GTS 512, NVIDIA offers a solution for a moderate price (299-349USD plus a price markup for a new product), which can outperform a much more expensive 8800 GTX card and even compete with a very expensive 8800 Ultra. This is very good news.

However, it's a nasty surprise for owners of 8800 GTS 512, who have bought this card only recently - the new card is a tad faster and it comes with the same price tag.

I repeat that 9800 GTX is not a revolutionary product, it's not 50% as fast as the 8800 GTS 512. In fact, it's the same product operating at a tad higher frequencies. If the prices for 8800 GTS 512 and 9800 GTX differ much, it will be a hard choice. All 8800 GTS 512 cards can operate at the core clock rate of the 9800 GTX (what concerns memory, we have to test whether 1.0-ns chips will manage to operate at 1100 MHz).

Now let's say a few words about each card.

BFG GeForce 9800 GTX 512MB PCI-E is a reference card (manufactured by NVIDIA's orders, BFG just buys these cards). The company gives a 10-year warranty for its products (it replies to messages of private users and can replace a defective product - you can read the details on the vendor's web site). Unfortunately, it's not that easy to find BFG products in our stores so far (you can mostly come across old products). You should keep in mind that the bundle lacks the HDMI adapter and the audio cable.

MSI GeForce 9800 GTX 512MB PCI-E (N9800GTX-T2D512) is also a reference card bought from NVIDIA. It's a standard product with a standard bundle. It will all depend on its price.

Zotac GeForce 9800 GTX 512MB PCI-E is also a reference card (manufactured by NVIDIA's orders at third-party plants). I have nothing to add. Just compare the prices and choose a card. If the price is adequate, it will be a good choice. Besides, the bundle includes a modern game.

In fact, if we forgive NVIDIA such a mess with product names, it's a very good card for its price. Unfortunately, our comparison with the 3870 card was only nominal, we didn't focus on it, because this card is much cheaper, and AMD has nothing to offer in the 299 USD price range. The 3870 x2 is more expensive by a hundred dollars.

So we cannot discard HD 3870 products - even though they are outperformed by the 8800 GTS/9800GTX, they are cheaper.

As always, the final choice is up to the reader. We can only inform you about products and their performance, but can't make a buying decision. In our opinion, that should solely be in the hands of the reader, and possibly their budget. ;)

And another thing that we are not tired to repeat from article to article. Having decided to choose a graphics card by yourself, you have to realize you're to change one of the fundamental PC parts, which might require additional tuning to improve performance or enable some qualitative features. This is not a finished product, but a component part. So, you must understand that in order to get the most from a new graphics card, you will have to acquire some basic knowledge of 3D graphics and graphics in general. If you are not ready for this, you should not perform upgrades by yourself. In this case it would be better to purchase a ready chassis with preset software (along with vendor's technical support,) or a gaming console that doesn't require any adjustments.

Now a few words about our benchmarks:

In our article about FRAPS, we illustrated how crude and inaccurate tests with this utility are. Testers do not have other tools, except for benchmarks built into games.

The article explains that it's sometimes possible to test games with integrated and IDENTICALLY looped demos (although with lots of potential errors, because measurement accuracy depends on a given tester: will he start/stop FRAPS in time or do it later/earlier). But I have run across situations, when demo load changes abruptly in the very beginning or end. So a half-second delay in starting/stopping the utility changes the average FPS by 15-20%. That's not a measurement error, such a test is a total waste of time. One time you delay the test, the other time you start it too early (not intentionally, of course), and you end up with absolutely different performance ratios.

But even that's not the most important thing. The fact is, there are almost no games with built-in demos anymore. So testers are forced to use a totally unacceptable method (in our opinion), when they measure gaming performance by walking a straight line from the initial point in a scene to a selected finish spot (the nearest fence, tree, etc.)

We all understand that it's impossible to navigate to a finish spot in PRECISELY the same route with different cards and in different resolutions. Besides, such games always introduce random elements into a scene, objects may be placed slightly differently ON THE SAME SCENE.

Unfortunately, web sites that publish a huge number of tests do not reveal their procedures in each game (except for those with built-in benchmarks)

So we are of the opinion that it's better to offer a limited number of game tests, but each one should be crystal clear and accurate, and they should all show differences between graphics cards.

To find more information regarding the current graphics card market and the performance of various cards, feel free to read our monthly special i3DSpeed.

Our Excellent Package award goes to Zotac GeForce 9800 GTX 512MB.