by Alexey Stepin
07/05/2006 | 11:26 AM
As you know from our earlier reviews called PowerColor Radeon X1800 GTO Graphics Card: the Best in Its Class?, ATI Technologies’ reply to Nvidia’s GeForce 7600 GT was the Radeon X1800 GTO graphics card that utilized the company’s large stock of R520 chips. A viable attempt at leadership in the $199-249 price segment, this graphics card indeed became one of the best mainstream products, even though not an all-around leader.
<%BANNER[article]%>The $249-299 segment being occupied by the Radeon X1800 XT 256MB and the Radeon X1900 XT costing as much as $449, there was a great gap between them which couldn’t be filled by the Radeon X1800 XT 512MB due to a number of reasons, the main of which was its out-dated design. An R580-based solution was necessary to offer something really competitive in this price sector because the potential of the R520 GPU with its only 16 pixel processors had been exhausted while the new graphics card was supposed to have better technical characteristics than the Radeon X1800 XT’s, but inferior to the Radeon X1900 XT’s. That’s the background behind the arrival of the ATI Radeon X1900 GT.
It’s a traditional trick of ATI Technologies’ to employ one and the same graphics processor on several graphics card models positioned in difference performance categories. Considering the flexible modular architecture of the Radeon X1000 series on the whole and of the R580 GPU in particular, it wasn’t a big challenge for the company to create a new performance-mainstream product. We know the R580 chip (or Radeon X1000) originally incorporates 48 pixel shader processors, 16 texture-mapping units and 16 raster operators. This 3:1:1 ratio of functional subunits embodies ATI’s current approach to designing graphics processors which is focused on achieving maximum performance at executing pixel shaders with complex math1ematical calculations.
Although the number of functional subunits may vary in each particular Radeon X1000 series GPU, ATI decided to stick to the 3-to-1 ratio in its Radeon X1900 GT. This GPU has 36 pixel processors, 12 TMUs and 12 ROPs or three fourths of what the Radeon X1900 XT has. This is a rather arguable solution considering the performance-related aspects of games we play today. Nvidia’s cards from the GeForce 7900 family have 24 TMUs and 16 ROPs and it will be hard for the newly arrived Radeon X1900 GT to compete with them, especially when full-screen antialiasing is enabled and in scenes with a lot of high-resolution textures or using shaders with multiple texture lookups. On the other hand, ATI claims its new product has a good reserve of performance for upcoming games. The claim sounds logical in view of the tendency towards putting ever more math1ematical operations into pixel shaders.
We put the specifications of the Radeon X1900 GT, Radeon X1800 XT and GeForce 7900 GT into one table for better comparison:
As you see, it’s hard to exactly pinpoint the Radeon X1900 GT’s position relative to other products. On one hand, the 36 pixel processors and the considerably higher GPU frequency should give the new card an edge over the GeForce 7600 GT in newest games. On the other hand, the Radeon X1900 GT has two times less of TMUs, which may affect its performance negatively in applications requiring high texturing performance. The use of only 12 TMUs is somewhat justifiable because the X1900 GT has a lower memory bandwidth than the GeForce 7900 GT, even though its efficient rig-bus memory controller may make up for this difference to some extent.
The new card also has obvious advantages over its market opponent in the image quality improvement area. We mean its support of FSAA and HDR simultaneously and its use of a higher-quality anisotropic filtering algorithm.
The Radeon X1900 GT can work in a CrossFire tandem, and you don’t need a Master card with a Compositing Engine for that. All you have to do is to install two Radeon X1900 GT cards into your CrossFire-compatible system (based on a CrossFire Xpress 3200 or Intel 975X chipset, for example). Then you just enable the appropriate option in Catalyst’s Control Center. The two cards will communicate via the PCI Express bus.
We will be examining PowerColor’s version of Radeon X1900 GT in this review. Let’s have a closer look at this card now.
The PowerColor X1900 GT graphics card belongs to PowerColor’s family of products based around Radeon X1900 GPUs, and its box is designed in the appropriate manner.
Pink is now added to the previous three colors (blue for the PowerColor X1900 XT, red for the PowerColor X1900 XTX and yellow for the PowerColor X1900 XTX CrossFire Edition). This kind of color coding allows telling different graphics card models from PowerColor apart without reading the text on their boxes.
And again, the use of the metallic effect doesn’t look garish because of the overall discretion of design. Besides some text, there is a sketchy image of the graphics card on the box – this makes it look like the package of the PowerColor X1800 GTO (for details see our article called PowerColor Radeon X1800 GTO Graphics Card: the Best in Its Class? ).
The small box will easily fit into any plastic bag for you to carry it home. As usual, there’s a main cardboard box under the colorful envelope. The graphics card lies in it on a cardboard tray. The accessories to the card are placed under this tray. Unlike in packages from ASUS, the card is simply wrapped into an antistatic bag and is not firmly fixed in place, but the contents won’t survive anyway if the box is severely damaged during transportation.
We opened the box up to find the following, besides a sample of the PowerColor X1900 GT graphics card:
This is a typical set of accessories you can expect to get hold of when you buy a PowerColor product. Exceptions are rare. It’s only with some top-end models that the company may include a couple of games. This lowers the price of the end product, making it more appealing to people who are not interested in games they can get along with a graphics card. Still, we think the PowerColor X1900 GT is a rather advanced product and might have come with at least one game and with a paper user manual, notwithstanding the fact that you receive a full electronic version of the manual on the enclosed Drivers CD.
Just like it had been the case with Radeon X1800 GTO, ATI Technologies didn’t develop a special PCB for its new card, but took the existing design employed in Radeon X1900 XT/XTX. All samples of Radeon X1900 GT are manufactured at the facilities of Sapphire Technology, the main partner of ATI’s. It means that everything in the subsequent description applies not only to PowerColor’s but also to other companies’ versions of Radeon X1900 GT. The cooling system may differ, however. Radeon X1900 GT cards can be equipped with one of two standard coolers from ATI or, in some cases, with a third-party cooler (e.g. from Arctic Cooling).
The PowerColor X1900 GT’s having fewer pixel processors, TMUs and ROPs active, its power circuitry has been greatly simplified. You can spot this in the right part of the PCB. The power circuit of Radeon X1900 XT/XTX cards includes seven Pulse PA0511.101 inductance coils, but there are only 6 such coils here. The bottommost coil and its accompanying elements are not installed. We didn’t remove the heatsink on the MOSFETs, but we suspect there are fewer of them there, too, due to the low power consumption of the Radeon X1900 GT (you’ll see how the new card compares with the Radeon X1900 XT in terms of power consumption in the next section of the review).
An ordinary R580 graphics processor is installed here and is clocked at 575MHz. Some of its functional subunits are disabled by melting the fuses inside the die package, which makes it impossible to transform the PowerColor X1900 GT into a Radeon X1900 XT. This would be a very appealing option since the R580 chip yield is high and they must mostly be using fully operable chips to make the Radeon X1900 GT. Alas, this overclocking dream won’t come true unless an exclusive batch of cards is released in which the “redundant” subunits are disabled on the BIOS level.
The memory frequency is greatly reduced in ATI’s new solution in comparison with the Radeon X1900 XT and its total amount is cut down from 512MB to 256MB, so the PowerColor X1900 GT carries 8 chips of Samsung K4J55323QG-BC14 GDDR3 memory in 136-pin FBGA packaging. The chips have 256Mbit capacity and work at 1.8V voltage. The memory is rated to work at 700 (1400) MHz frequency, but it is clocked at 600 (1200) MHz here in strict compliance with the Radeon X1900 GT specification. Since the card’s PCB was originally designed for much higher frequencies, we can expect the PowerColor X1900 GT’s memory to reach at least its nominal 700 (1400) MHz clock rate at overclocking.
Like almost all cards on ATI’s GPUs, the PowerColor X1900 GT features a Rage Theater chip for capturing analog video signal. This feature isn’t in high demand nowadays, though. We think they might have not installed this chip at all to reduce the manufacturing cost of the Radeon X1900 GT even further.
We were a little surprised to find on the PowerColor X1900 GT a sample of the old cooler which traces its origin back to the Radeon X1800 XL (we described this cooler in detail in our All-in-Wonder X1800 XL review). This cooler has a lower-performance, but noisier fan in comparison with the new cooler we talked about in our PowerColor X1800 GTO Review. Incorporating heat pipes and a well-ribbed heatsink, this cooler can even cope with junior R520-based graphics card models, but it may find it difficult to handle an R580, even with some of the chip’s subunits disabled. We know that there are versions of Radeon X1900 GT, e.g. the Connect3D Radeon X1900 GT, which feature coolers with an updated design. The older design is probably used because the manufacturer’s got a large amount of such coolers in stock.
Following the usual testing procedure we apply to each graphics card that enters our labs, we measured the power consumption of the Radeon X1900 GT on a modified testbed configured as follows:
The measurements were performed with a digital multimeter Velleman DVM850BL (0.5% measurement accuracy). We loaded the GPU by launching the first SM3.0 graphics test from 3DMark06 and running it in a loop at 1600x1200 resolution and with enabled 16x anisotropic filtering. The Peak 2D load was created by means of the 2D Transparent Windows test from Futuremark’s PCMark05 benchmarking suite. The results follow below:
The Radeon X1900 GT’s using an R580 graphics processor, which was originally designed for top-end graphics cards, tells on its power consumption characteristics. The new card has a healthy appetite, as you can see. Its power draw is only lower than that of the GeForce 7900 GTX, a graphics card from a higher class. The GeForce 7900 GT, the direct market rival to the Radeon X1900 GT, consumes only two thirds of the amount.
Like all other members of the family, the Radeon X1900 GT puts a bigger load on the external +12V line than on the internal one, although the misbalance isn’t as big as with the Radeon X1900 XT and XTX. In 3D mode the external line provides 38 watts of power whereas the PCI Express slot yields 32 watts more. The remaining 5 watts come to the card via the +3.3V line (this is characteristic of ATI’s solutions, too; the GeForce 7900 GT consumes nearly nothing from this line, for example).
So, the Radeon X1900 GT has nothing to boast about when it comes to power consumption. In order to see better results in this test we have to wait for ATI’s upcoming solutions on the economical RV570 chip that are expected to deliver the same performance but have much better power-related characteristics.
Next we also tried to overclock our sample of the PowerColor X1900 GT. We were not very successful, to tell you the truth, but that might have been expected considering the modest capabilities of the cooler the card is equipped with and the high power consumption level of the GPU. However, we did manage to increase the GPU frequency to 620MHz and memory frequency to 700 (1400) MHz by only adding a 120mm fan to the card’s standard cooler – the additional fan was blowing air along the card’s PCB. The card was stable at the said frequencies, but a hang-up and a BSOD occurred when we tried to set the GPU clock rate at 625MHz. The same thing happened when we attempted to lift the memory frequency higher up to 720 (1440) MHz.
Yes, our overclocking attempts were seriously restrained by the cooler which was a little too week for a card that generates so much heat. Anyway, we think the GPU and memory frequency gain we achieved is going to positively affect the performance of the PowerColor X1900 GT. Enthusiasts who are into serious overclocking may achieve better results than ours but this would call for more effective cooling methods and perhaps GPU/memory volt-modding.
We measured the level of noise produced by the PowerColor X1900 GT with the help of a digital sound-level meter Velleman DVM1326. This instrument has a resolution of 0.1dB and allows measuring in a range up to 130dB with A or C weighting.
The background noise level was 36dBA in our lab when we were conducting the measurements. At a distance of 1 meter from the test platform in which a graphics card with passive cooling was installed the level of noise was 40dBA. These are the reference numbers we’ll base our judgments upon.
Here are the results:
Most of the time the PowerColor X1900 GT isn’t audible against the noise from other components of a working gaming system. You surely won’t hear the card in a closed system case as the numbers suggest: the loudness of its noise is 42.6dBA. Compare this with the near-silent GeForce 7900 GTX’s 44.6dBA.
Lacking any fan speed management system, the GeForce 7900 GT produces much more noise. The difference is 3.6dBA when the measurements are performed at a distance of 1 meter from the testbed. This is quite a lot, considering the logarithmic nature of the decibel. You can always hear the card when the system is working while the PowerColor X1900 GT only makes itself heard at those rare moments when the GPU temperature reaches a certain threshold value after long running 3D applications. It then produces noise with a sound intensity of 45.2dBA. This seems to be not very loud, but the spectrum of the noise from the small but high-speed fan is very annoying. Fortunately, this doesn’t happen too often and may not happen at all if your system case is well ventilated.
The noise characteristics of the PowerColor X1900 GT are overall acceptable except for the start-up moment when the cooler works at its full speed for a few seconds and for those moments when the GPU overheats. This graphics card is surely better from this point of view than the GeForce 7900 GT which has a rather poorly designed cooler and cannot reduce the speed of the fan (although the G71 chip itself doesn’t generate too much heat).
We have no complaints about the image quality the PowerColor X1900 GT provides in 2D applications. No wonder as all senior models in the Radeon family are manufactured at one and the same fab where each card undergoes a strict quality check. So, the card yielded a sharp picture without any shadowing or fuzziness in all display modes, including 1600x1200@85Hz and 1800x1440@75Hz. We’d like to note that the quality of a graphics card’s analog output isn’t as important nowadays as it used to be. LCD monitors with a DVI input are currently widespread and provide an immaculate 2D image quality due to the digital nature of this interface.
We tested PowerColor X1900 GT in the following testbed:
We set up the ATI and Nvidia drivers in the same way as always:
ATI Catalyst:
Nvidia ForceWare:
We selected the highest graphics quality settings in each game, identical for graphics cards from ATI and Nvidia, except for the Pacific Fighters flight simulator that requires vertex texturing for its Shader Model 3.0 rendering mode. Radeon X1000 doesn’t support this feature therefore we ran the game in Shader Model 2.0 in this case. We did not edit the configuration files of the games. We also didn’t use the driver profiles optimized for given games. To measure the performance we either used the integrated tools of the games we tested in, or if there were none available, resorted to FRAPS utility. If it was possible, we measured minimal performance as well.
To load the video subsystem to the full extent and to minimize the influence of the CPU speed on the performance results we didn’t test the systems in the “pure speed” mode. We only ran the tests in “eye candy” mode with full-screen anti-aliasing and anisotropic filtering. The graphics subsystem potential is used in a more optimal manner in this case. Besides, the image quality is considerably higher than in those cases when no full-screen anti-aliasing and/or anisotropic filtering is enabled.
We enabled FSAA and AF from the game if possible. Otherwise we forced the necessary mode from the ATI Catalyst and Nvidia ForceWare graphics card driver. We have also tested the following graphics cards besides the PowerColor X1900 GT:
These games and applications were used as benchmarks:
First-Person 3D Shooters
Third-Person 3D Shooters
Simulators
Strategies
Synthetic benchmarks
Even with fewer TMUs the PowerColor X1900 GT is no worse than the GeForce 7900 GT and leaves it behind in high resolutions (not that it’s very important here since modern GPUs all give you enough speed in Battlefield 2). Note that the new card has almost the same performance as the Radeon X1800 XL in 1024x768, but breaks away from it in the higher resolutions thanks to faster memory.
The overclocking ensures a 10% performance boost in resolutions up to 1280x1024, enabling the new card to compete with the Radeon X1800 XT 512MB.
The Radeon X1900 GT allows to play Battlefield 2 in every resolution with turned-on FSAA and anisotropic filtering; the average frame rate is never lower than 60fps.
The PowerColor X1900 GT is as fast as the Radeon X1800 XL in all the three standard resolutions and is always keeping rather far away from the GeForce 7900 GT.
Like in the previous test, the overclocking provides a 10% speed bonus. This makes the gap between the PowerColor and the Radeon X1800 XT 512MB smaller, but doesn’t eliminate it completely.
The PowerColor X1900 GT’s average performance is 50fps in 1280x1024, which is not enough for comfortable play as it lacks any speed reserve. It means you have to use 1024x768 resolution or lower the level of detail or disable full-screen antialiasing. You don’t have to face this choice if you’ve got a GeForce 7900 GT: thanks to the game’s using OpenGL and dynamic stencil shadows, this card delivers over 65fps in 1280x1024 and is almost as fast as the Radeon X1900 XT.
It’s only in 1600x1200 that the Radeon X1900 GT enjoys a considerable advantage over the Radeon X1800 XL, but this advantage is of no real use for the gamer because the average speeds of these two cards aren’t higher than 20fps then. 1024x768 is in fact the single resolution you can play the game with more or less comfort in even on the GeForce 7900 GT which is 25% ahead of the PowerColor card.
Even the mighty Radeon X1900 XT doesn’t suffice for normal play in the higher resolutions if full-screen antialiasing is turned on. So like in the previous test, the user has to choose between lowering the level of detail, which will make the game less attractive visually, and disabling FSAA. We’d prefer the latter option as it affects the game’s visuals less.
Like it could have been expected, the PowerColor X1900 GT isn’t any faster than the Radeon X1800 XL on the Pier level and is even slower in 1024x768. This comes from the specifics of the design of the test scene. The flight on a hang-glider means a vast field of vision and, accordingly, an increased textural load. In the higher resolutions the PowerColor card does better due to its high GPU/memory frequencies. It equals the GeForce 7900 GT here.
The overclocking adds more vigor to the card but it still cannot reach the level of the Radeon X1800 XT 512MB. This is not a problem for this rather undemanding game, however. Even at its default frequencies the Radeon X1900 GT allows playing Far Cry with comfort using 1600x1200 resolution, full-screen antialiasing and anisotropic filtering.
The new card is better than the Radeon X1800 XL. Like in the previous case, its performance is high enough for you to use FSAA and anisotropic filtering in 1600x1200 resolution. The rest of the participating cards deliver comfortable performance, too.
We’ve got interesting results in the game’s HDR mode. The Radeon X1900 GT isn’t impressive in 1024x768, being a mere 5fps ahead of the Radeon X1800 XL, but after that it suddenly soars up to the level of the Radeon X1800 XT 512MB. The triumph didn’t last long, however, as in 1600x1200 the PowerColor X1900 GT is again closer to the Radeon X1800 XL, probably because it has less of graphics memory than the Radeon X1800 XT 512MB. Well, none of the participating graphics cards provides even a minimum of comfort in this resolution.
The new card behaves in a similar way when we turn on HDR on the Research map. Despite the inefficient implementation of this mode for Radeon X1000 cards in the current version of Far Cry, the PowerColor allows playing in 1280x1024 while the Radeon X1800 XL doesn’t. Note that you can have a comfortable frame rate on the GeForce 7900 GT in 1280x1024 resolution in the HDR mode.
Although F.E.A.R. is a pixel shader rich application, the PowerColor X1900 GT has no advantage over the Radeon X1800 XL or the GeForce 7900 GT; the 12 TMUs must be the limiting factor. Otherwise this graphics card would have every chance to outpace the GeForce 7900 GT. Nvidia’s solution isn’t too much faster, though. It doesn’t allow playing the game comfortably in 1280x1024 with enabled FSAA whereas in 1024x768 an acceptable average frame rate is provided by the Radeon X1900 GT and by the Radeon X1800 XL as well.
The game uses the deferred rendering technique which makes the use of FSAA impossible. That’s why we tested it only with enabled anisotropic filtering. The game also has resolution and graphics quality limitations for cards with less than 512MB of memory on board.
The PowerColor X1900 GT feels at ease here. Despite having only 12 TMUs and ROPs, it is successfully competing with the Radeon X1800 XT 512MB and leaves it behind in 1280x1024. This game obviously appreciates the availability of many pixel shader processors.
This is a rather difficult test and acceptable speeds are only achieved in 1024x768. In higher resolutions even the Radeon X1900 XT with its 48 pixel processors cannot keep the average frame rate at 50-55fps.
With only 12 TMUs the PowerColor X1900 GT is confidently ahead of the Radeon X1800 XL from 1280x1024 resolution onward. When overclocked, it is just a little slower than the Radeon X1800 XT 512MB. This means the GPU frequency is the decisive performance-related factor in this test.
This sequel to Half-Life 2 picks up at the moment of the explosion of the Citadel’s main reactor. Episode One utilizes an improved version of the Source engine our readers should be familiar with by the tech demo Half-Life 2: Lost Coast. This engine employs an integer based HDR method (INT16) which is different from Far Cry’s HDR implementation. This method is less accurate, but it is compatible with full-screen antialiasing and works on any graphics card that supports Shader Model 2.0.
The new card is as fast as the GeForce 7900 GT in the lowest resolution, but the latter leaves it behind in the higher display modes. This allows the Nvidia card to show an acceptable average speed in 1600x1200 whereas the Radeon X1900 GT is limited to 1280x1024 resolution in which its performance is about 70fps.
Prey shares with Duke Nukem Forever and S.T.A.L.K.E.R.: Shadows of Chernobyl the dubious fame of one of the longest-under-development projects. This sci-fi shooter was announced back in 1997, but the developer completely forgot about the project afterwards. However, Prey survived eventually and is scheduled for a July 11, 2006 release.
In brief, this game is a sci-fi shooter utilizing a greatly modified Doom 3 engine with an Indian as the main character. Prey features high-quality, even though not outstanding, visual effects and surely deserves a place in our reviews.
The GeForce 7900 GT is a little slower than the Radeon X1900 XT and is far ahead of the Radeon X1900 GT, which in its turn is faster than the Radeon X1800 XL but slower than the Radeon X1800 XT 512MB. The PowerColor X1900 XT provides what seems to be close to a comfortable speed in 1024x768, but it should be noted that the GeForce 7900 GT cannot deliver a comfortable average and minimum frame rate in the higher resolutions, either. At least this is true for the full-screen antialiasing mode.
The demo version of Prey we used in this review doesn’t offer integrated benchmarking options, so we had to measure the frame rate with the Fraps utility. It means the results may be somewhat inaccurate since it’s hard to repeat exactly the same sequence of the player’s actions.
Quake 4 makes use of stencil shadows the GeForce 6/7 architectures have a special technology, called UltraShadow II, to accelerate the rendering of. So, it is no surprise that the GeForce 7900 GT feels much better here than its young opponent. You can see that the Radeon X1800 XL with 16 pixel processors, TMUs, Z-units and ROPs has nearly the same speed as the Radeon X1900 GT whose main advantage – 36 pixel processors – isn’t called for in this game.
Even the increased GPU and memory frequencies cannot help the new card. It cannot deliver a normal speed in 1600x1200 with enabled full-screen antialiasing. Its results are on the verge of being acceptable even in 1280x1024, unlike the Radeon X1800 XT’s.
Just like all ATI Radeon graphics cards with 256 megabytes of memory, the PowerColor X1900 GT isn’t brilliant in Serious Sam 2 when the game’s highest graphics quality settings are selected. The card yields an average frame rate of less than 30fps even in 1024x768. The GeForce 7900 GT has an average speed of 45fps, but its min speed is only 5fps which is far from comfortable. This means that graphics cards with 256MB of onboard memory are not suitable for Serious Sam 2, at least if you want to use the highest graphics quality settings and FSAA. Note also that the GeForce 7900 GT is again ahead of the new card from ATI thanks to its 24 TMUs.
We do not benchmark this game with FSAA because the GeForce 7 series doesn’t support full-screen antialiasing along with HDR. And we do not want to test without HDR because it makes the game much more beautiful.
The graphics card is expected to quickly process pixel shaders, shadows and lighting in Oblivion’s closed environments. The PowerColor X1900 GT leaves the Radeon X1800 XT 512MB far behind, but has a little lower average performance than the GeForce 7900 GT in resolutions above 1024x768. On the other hand, its minimum performance is always higher than that of the Nvidia solution because it has more pixel processors (36 against 24) which are also clocked at a higher frequency.
If you use the PowerColor X1900 GT (or any other version of the Radeon X1900 GT) in 1280x1024, you can be sure the frame rate will never sink below 30fps, which guarantees smooth game play. And if you lower the resolution to 1024x768, you can have an average performance of over 60fps which is comfortable for any 3D shooter.
The 24 TMUs of the GeForce 7900 GT come into play in the game’s open scenes, and this card has a higher average speed than the Radeon X1900 GT. However, the latter provides a considerably higher minimum speed. We guess having a min speed of 32fps and an average speed of 19fps is better for the gamer than 37/12fps, so this round seems to be won by ATI Technologies’ solution.
TES IV: Oblivion doesn’t have integrated benchmarking options, so we had to test it manually with Fraps. The results may be somewhat inaccurate as a consequence.
It’s strange, but the PowerColor X1900 GT isn’t superior to the Radeon X1800 XL and GeForce 7900 GT in SC: Chaos Theory which uses math1ematics-heavy shaders in its visual effects. The cards show similar speeds, although the PowerColor is just a little faster than the other two in 1600x1200. Judging by the results of the Radeon X1800 XT 512MB, the reason is that the game doesn’t need many shaders to be processed in parallel, but needs a high speed of execution of a rather small number of them, which depends directly on the GPU clock rate.
When overclocked, the PowerColor is quite successfully competing with the GeForce 7900 GT despite having two times less of texturing-related subunits. There’s nothing surprising in this. The game was ported from video consoles and doesn’t have complex high-resolution textures, focusing on shader-based special effects instead. The Radeon X1900 GT delivers a comfortable performance in 1600x1200 with enabled full-screen antialiasing. We want to remind you that the requirements to minimum and average speeds are less strict for a third-person shooter than for a first-person one in which your shooting accuracy depends on the smoothness of play.
The PowerColor X1900 GT can’t impress with its performance in this Nvidia cards-oriented game. However, it runs Pacific Fighters fast enough to provide you an enjoyable flight experience.
Unfortunately, the Radeon X1000 family do not support vertex texturing and render a lower-quality water surface than is available on the GeForce 7900 GT.
Unlike the flight sim Pacific Fighters, the space sim X3: Reunion evidently prefers the Radeon X1000 architecture and the GeForce 7900 GT has to compete with the Radeon X1800 XL only (but its speed isn’t disappointing, after all). The PowerColor X1900 GT is slower than the Radeon X1800 XT 512MB, but the overclocking helps to narrow the gap to 5%.
The PowerColor card allows playing with comfort in 1280x1024 resolution with enabled 4x FSAA and 16x anisotropic filtering. Well, you can also try to switch to 1600x1200 – the minimum speed is never lower than 30fps.
The PowerColor X1900 GT isn’t far better than the Radeon X1800 XL, both these cards lagging behind the GeForce 7900 GT. We know that this game can put all of the graphics card’s pixel processors to use as is indicated by the performance of the Radeon X1900 XT relative to the Radeon X1800 XT 512MB. The small number of TMUs and ROPs must be the reason for the humble performance of the Radeon X1900 GT.
The PowerColor allows playing this game comfortably in 1024x768 with enabled FSAA. Its performance is too low in the higher resolutions, so you have to choose between FSAA and high display modes.
The game using stencil shadows and high-resolution textures, the PowerColor X1900 GT is initially handicapped in Dawn of War, which is a DirectX 8 application. This graphics card delivers a rather high average performance, but its min sped is below 20fps in 1024x768. You’ll probably have to turn full-screen antialiasing off to ensure accurate control over your virtual army.
Not looking very strong in the gaming tests, the Radeon X1900 GT is similar to the Radeon X1800 XT 512MB in 3DMark05 and scores about 1,500 points more than the GeForce 7900 GT although has fewer TMUs and ROPs. It makes sense to analyze the benchmark’s results test by test as they differ a lot among themselves.
The large-scale first test is not a problem for the PowerColor X1900 GT which is on the same level with the Radeon X1800 XT 512MB and far faster than the GeForce 7900 GT. Considering that this test requires a high fill rate from the graphics card, this is an excellent result, although we are not quite sure where it comes from. It probably comes from the architectural peculiarities of the R580 GPU which has more general-purpose registers than the R520 or from some optimizations implemented in ATI’s Catalyst driver.
The second test brings about rather surprising results. The fill rate parameter isn’t at all important here, but the PowerColor X1900 GT somehow cannot compete with the Radeon X1800 XT 512MB as in the previous case, but is considerably slower at the default frequencies. The overclocking helps to nearly eliminate this gap, thus giving us a glimpse of the answer. The performance of the vertex processors seems to determine the overall result of a graphics card in the second test. The Radeon X1800 XT 512MB and the Radeon X1900 GT have the same number of vertex processors, and the GPU frequency growth on the latter card leads to an increase in its vertex processors performance, diminishing the gap between the two products.
The PowerColor X1900 GT (and any other version of the Radeon X1900 GT, too) could do better in the third test, which requires fast processing of complex shaders, if it had more TMUs and ROPs. Anyway, it easily left the Radeon X1800 XL behind and would surely beat the Radeon X1800 XT 512MB if it had 16 texture-mapping units and 16 raster operators. It’s only at the overclocked frequencies that the new card can overtake the latter.
The results of the separate tests agree with the total score, considering the use of full-screen antialiasing. The Radeon X1900 GT is indeed good in 3DMark05, except for the second test where it is slower than the Radeon X1800 XT 512MB.
It’s different in 3DMark06: the PowerColor X1900 GT can’t defeat the GeForce 7900 GT because the tests are more complex, requiring a fast execution of shaders as well as a high scene fill rate. Let’s view the results by the groups of tests.
As might have been expected, the PowerColor X1900 GT is a little slower than the GeForce 7900 GT and the Radeon X1800 XT 512MB in the SM2.0 graphics tests. The biggest gap can be observed in the first test of this group, the one which is the most sensitive to the fill rate parameter.
Contrary to the previous case, the PowerColor X1900 GT shows its best in the SM3.0/HDR tests where its 36 pixel processors have got a lot of work to do executing numerous pixel shaders. Its speed in this group of tests is even a little higher than that of the Radeon X1800 XT 512MB.
The supposition is true: the PowerColor X1900 GT is indeed not very good in the first test that features 26 light sources and high-resolution shadow maps. It is slower than the GeForce 7900 GT and the Radeon X1800 XT 512MB which have 24 and 16 TMUs, respectively.
Using its Fetch4 technique to accelerate the rendering of shadows with the CSM method, the Radeon X1900 GT isn’t slower than the GeForce 7900 GT in the second test. The relatively low GPU clock rate puts it behind the Radeon X1800 XT 512MB. The overclocking eliminates this gap.
Thus, the Radeon X1900 GT is slower than the GeForce 7900 GT in one test only. It is the first test in the SM3.0 group which is very sensitive to texturing and lighting-processing speed. In the second test of this group the situation changes into the opposite. Considering the use of FSAA in the individual tests, the overall scores look well-deserved.
So, the tests are over and we now have to analyze the results and see if ATI Technologies’ attempt to fill in the niche between the Radeon X1800 XT and the Radeon X1900 XT with the new graphics card called Radeon X1900 GT is a success. The company has once again come up with a new product that utilizes the existing resources, but does it stand any chance on the market?
Having the results of the gaming tests on our hands, we cannot answer this question with a definite Yes. Following their 3-to-1 concept, ATI’s developing team didn’t take it into account that 12 TMUs and ROPs is insufficient for a performance-mainstream graphics card, especially since the GeForce 7900 GT, the main opponent of the Radeon X1900 GT, has 24 TMUs and 16 ROPs. The competing product from among the Radeons, the Radeon X1800 XT, not only has 16 TMUs, but also high GPU and memory frequencies. So each time you launch an application that requires a lot of texture lookups, the performance of the Radeon X1900 GT goes down, often below the level of the Radeon X1800 XT.
The Radeon X1900 GT generally provides a high enough performance in today’s games. However, it is only in 4 tests out of 19 that it enjoyed a confident victory over its market opponent and in 4 tests more equals the performance of the GeForce 7900 GT. These 8 tests are Battlefield 2, Far Cry (except in the HDR mode), Half-Life 2, TES IV: Oblivion, Splinter Cell: Chaos Theory, X3: Reunion and both 3DMarks. As you see, Half-Life 2 is the only game in the list that doesn’t use math1ematics-heavy shaders. In other cases the new solution from ATI was hamstringed by its having too few texture-mapping units as we’ve repeatedly said throughout this review.
Being a compromise solution with a graphics processor originally meant for top-end graphics cards, the Radeon X1900 GT is also inferior to its main opponent in terms of size and power consumption, but produces less noise at work. It allows using FSAA and HDR (FP16) simultaneously, supports high-quality anisotropic filtering and features the Avivo video engine.
As far as we know, Radeon X1000 graphics processors permit to flexibly configure the number of the functional subunits, so it’s not quite clear why ATI Technologies didn’t enable, for example, 16 TMUs and 12 ROPs in the Radeon X1900 GT (especially considering that this graphics card is obviously manufactured out of fully operable R580 chips). We don’t know if it’s due to an unwavering adherence to the 3-to-1 concept or some technological issues do not permit to change the number of TMU and ROP subunits relative to the number of pixel processors at will. Anyway, the Radeon X1900 GT didn’t win the battle with the GeForce 7900 GT for the title of the best performance-mainstream graphics card of today, at least in today’s games. The situation may change in the future in view of the tendency towards ever more complex math1ematics in pixel shaders, but the Radeon X1900 GT will have probably been ousted from the market by then with newer solutions.
The market success of the ATI Radeon X1900 GT depends on the company’s ability to keep its price below $299 which is the official price of the GeForce 7900 GT. In its last quarter report ATI Technologies promised to increase its profitability from 30.1% to 31.0-31.5% in the current quarter. This may mean that the company isn’t going to reduce the prices or that it will release something new that can be sold at a high profit. But it may also find some means to lower the prices for its existing products in which case ATI may make its Radeon X1800 XT and Radeon X1900 GT cheaper.
As for PowerColor’s version of the Radeon X1900 GT which we’ve tested today, it is a typical product of this company, i.e. an ordinary “workhorse”. It ships without extra accessories, but may be very appealing in the eyes of users for whom the graphics card itself and its cost are more important than the design and size of its package and free games in the included software pack. If you are interested in the Radeon X1900 GT and want to buy a performance-mainstream graphics card with some reserve of speed for future games, but do not want to pay extra for any designing whims and games you won’t play, consider the PowerColor X1900 GT. It may be just what you need.
Highs:
Lows: