by Alexey Stepin , Yaroslav Lyssenko
07/23/2009 | 11:42 AM
As we know, ATI Radeon HD 4870 X2 didn’t have any competitors for a while: any Nvidia solutions with similar price point couldn’t compete against it in performance. At the same time, the idea to design a dual-processor accelerator on G200 chips that would be able to challenge the flagship “red” solution and possibly defeat it kept failing because of enormous size and extremely high heat dissipation of 65 nm GPUs each featuring 1.4 million transistors.
Luckily for Nvidia, Advanced Micro Devices graphics division got too carried away by their undefeated victory, which gave Nvidia some time to regroup and catch up. As a result, G200 was transferred to more progressive 55 nm manufacturing process, which lowered its heat dissipation and power consumption levels quite seriously. It finally became possible although still pretty hard to design an ideological successor to GeForce 9800 GX2. Nvidia simply had no choice and at last they dared do it. This is how GeForce GTX 295 came around that we first discussed in our article called “The King Is Dead, Long Live the King: EVGA GeForce GTX 295+ Graphics Card Review”.
The newcomer featuring two G200b processors tied together using SLI technology managed to defeat the 3D king – Radeon HD 4870 X2. However, it was a true Pyrrhic victory: GeForce GTX 295 turned out extremely complex in design and expensive to manufacture. Remember that it used two PCBs, one per GPU, that were tied together into a high-tech “sandwich” with a cooler inside. This cooler is also result of a compromise. Yes, formally, Nvidia took over the leadership, but in reality GeForce GTX 295 didn’t become very popular neither among the manufacturers, nor among consumers, and even turned out unprofitable for Nvidia. Its high production cost was one of the factors that determined such outcome, because a graphics accelerator with complex design like that couldn’t be cheap. As a result, the new solution was pretty hard to find in retail, and even if you did find it, then its price could be way higher than the recommended one. Moreover, the investment was not worth the performance advantage in games over Radeon HD 4870 X2. Another indication of the crisis is the fact that EVGA GeForce GTX 295+ remained the only such graphics solution that we managed to get into our lab.
So, Nvidia once again became the victim of their own strategy that was targeting the development and production of the most complex and most powerful graphics processors. But at that point it was impossible to give in, especially after AMD stroke a pretty tangible blow with their new Radeon HD 4890. It was impossible to make G200 any simpler; besides, Nvidia didn’t have any adequate replacement for it that is why they had to simplify everything else other than the GPU (see this news story for details).
The company decided to do the impossible: they attempted to design a simpler and cheaper version of a GeForce GTX 295 and they succeeded! Moreover, it could also be sort of a trial run, preparation for production of a new generation of dual-processor Nvidia graphics cards, since the initial GeForce GTX 295 design couldn’t be taken as a basis for them because of its extreme complexity and high production cost.
In our today’s review we will try to find out how Nvidia managed to turn defeat into victory that is why we will discuss the new GeForce GTX 295 revision in detail. Looks like the new design did indeed turn out a success: we managed to get three different graphics cards o0n the flagship “green” solution at the same time. The first one we got was Inno3D GeForce GTX 295 Platinum, so it will do the honor of representing the new solution in X-bit Laboratories.
Inno3D Company has long been known for their inexpensive solutions and, frankly speaking, this past got reflected in the design of the GeForce GTX 295 Platinum packaging: the amount of golden and silver colors in the box design exceeds all possible and impossible limits, in our opinion. The picture is also quite common for the graphics market: it depicts another female warrior with a humongous sword:
However, the widely spread mistake with the wrong graphics memory type is right there: as you can see in the lower right corner of the box, the card has 1792 MB of DDR3 with 896-bit access onboard. Of course, in reality the card is equipped with two GDDR3 memory banks, each 896 MB big and connected to its graphics core via a 448-bit bus. The specifics of contemporary multi-GPU technologies implementation imply that contemporary 3D applications can only use half of the indicated memory capacity, i.e. 896 MB, and this is where GeForce GTX 295 yields to Radeon HD 4870 X2 that is equipped with two 1024 MB memory banks.
As for other information that may be helpful to the potential buyer, I should mention the stickers about Far Cry 2 and Warmonger: Operation Downtown Destruction games bundled with the card. I am sure everyone is familiar with the first one, while the second one is a free multi-player shooter based on Unreal Engine 3 that broadly utilizes PhysX acceleration feature of Nvidia GeForce processors.
The box is fairly small and can fit into any bag or plastic bag, so you won’t have problems carrying the newly bought card home. Beneath a bright cover there is a solid box made of thick brown cardboard that holds two more boxes. One large box protected with extra polyurethane foam padding holds the card sealed into a antistatic bag, while the other contains all bundled cables and adapters. The list of bundled accessories includes the following items:
We think that an adapter for 8-pin PCIe 2.0 could be a nice addition to this bundle, because not all contemporary power supply units are equipped with this connector so far. The absence of user manual could also be considered a drawback, although there is something like that on the inside of the sleeve for the CD disk with drivers. The disk contains not only the drivers and outdated instructions for them, but also a copy of the Badaboom media-converter that can take advantage of the Nvidia GPU computational capacities.
Overall, the accessories bundled with Inno3D GeForce GTX 295 Platinum made a twofaced impression. On the one hand, two high-quality games included with the product are a definite advantage, but on the other, it is frustrating to have an important adapter and user manual missing. The same is true for the packaging, which very well fulfills its protective functions but features pretty tacky design. We get the feeling that Inno3D was in a real hurry to release their GeForce GTX 295 and didn’t have enough time to think all the details carefully through.
Of course, package and accessories are important, especially for a high-end solution, but they do not determine its performance and ability to win or lose. So let’s now take a closer look at the card itself.
Since all GeForce GTX 295 cards share the same reference PCB design, everything we say about Inno3D card refers to other manufacturers’ versions, too. The updated GTX 295 resembles the GeForce 7900 as it has a large fan placed in the center. The PCB dimensions are the only thing the updated version has in common with the older one.
GeForce GTX 295: new revision
GeForce GTX 295: old revision
The most exciting things are hidden inside, of course. In order to see them, we had to take the card apart and the new GeForce GTX 295 proved to have nothing in common with the old version even from this aspect: it is far easier to disassemble. The external plastic casing embellished with Inno3D stickers is fastened with latches and can be easily removed, revealing the cooling system.
The heatsinks can also be removed by unfastening the screws that connect the cooler’s base with the metallic back-plates. Then you will see the following:
GeForce GTX 295 PCB: new revision
GeForce GTX 295 PCB: old revision
So, the two versions of GeForce GTX 295 are very different indeed. The newer one is much simpler and has one PCB although we had suspected that would be technically impossible. We can’t but applaud Nvidia’s engineers who have managed to fit a couple of huge G200b processors, two sets of GDDR3 chips with 448-bit memory buses, two NVIO2 chips, an nForce 200 PCI Express switch and voltage regulators into such a limited space. The component density of the GeForce GTX 295 is far higher than that of the Radeon HD 4870 X2. There are almost no unpopulated areas on it.
Each GPU is powered by an individual 3-phase regulator based on an ADP3192A controller from Analog Devices. Nvidia advertises an enhanced pulse-width modulation technology that can react to changes in load faster and more flexibly, but its real benefits are unclear. It is also unclear if the GPU voltage can be controlled by software.
The memory sets are powered by single-phase regulators that seem to be controlled by a tiny chip marked as “N12 VGF936”. The GPU and memory voltage regulators are based on Renesas R2J20651NP integrated packs that combine power MOSFETs and drivers, making the power circuit compact.
The additional voltage regulators based on uP6161 and APW712 chips seem to power the other components of the card such as the PCI Express switch and the NVIO2 chips.
External power is provided to the card via two connectors, a 6-pin and an 8-pin one. As before, the 8-pin power cable must be plugged in. Otherwise the card won’t start, reporting a power problem with the LED on its mounting bracket. This protection can still be easily bypassed by closing the two outermost pins in the connector, so the GeForce GTX 295 can be powered by two 6-pin PCIe 1.0 connectors if these PSU outputs have a sufficient load capacity.
For some reason the card is still equipped with two NVIO2 chips. In the older, dual-PCB version of GeForce GTX 295, the second chip was necessary for a dedicated HDMI port but the new version has two DVI ports that can be supported by one NVIO2 chip. We suppose that the two chips may be needed to support multi-monitor configurations in multi-GPU mode. Nvidia used to have problems with that function and has only solved them recently.
The nForce 200 chip is an intelligent PCI Express 2.0 switch that supports 48 PCIe lanes and allows the connected devices to communicate directly in peer-to-peer mode. In fact, the nForce 200 moves the SLI logic from the mainboard to the graphics card so that the GeForce GTX 295 could work in SLI mode even on SLI-incompatible mainboards.
The PCB is equipped with a couple of revision B3 G200b processors. One of them was manufactured on the 23rd week of 2009 and is marked as “G200-401-B3”. The other is 2 weeks younger and is marked as “G200-400-B3”.
GPU-Z 0.3.4 does not recognize the new version of GeForce GTX 295 but reports a number of its GPU parameters correctly. Despite the word “Platinum” in the name of the card, its basic specs are standard: the GPUs are clocked at 576MHz and 1242MHz (the main and shader domains, respectively) and have 240 ALUs, 80 texture processors, and 28 RBEs each. For all their efforts, Nvidia’s engineers were unable to place two 512-bit memory banks on the PCB, so the memory chips are still connected to the GPUs with 448-bit buses, each GPU having only 7 active memory controller sections out of 8.
Each memory bank consists of 14 GDDR3 chips from Hynix (H5RS5223CFR-N0C). Half the chips are located on the reverse side of the PCB. Each chip has a capacity of 512 Mb (16Mb x 32), a voltage of 2.05V, and a rated frequency of 1000 (2000) MHz. The card’s actual memory frequency is 1008 (2016) MHz, which leaves almost no room for overclocking.
Each memory bank is 896MB large, but it would be wrong to say that the card has 1792MB of memory. 3D applications can only use half the total amount due to the specifics of today’s multi-GPU technologies. This should not be a problem since 896MB is more than enough for most modern games. And even if there is any shortage in memory, the developer can increase its amount by installing 1Gb GDDR3 chips.
The new GeForce GTX 295 comes with two DVI-I ports. As opposed to the older version, it does not support HDMI or analog video outputs natively. The developer has a reason for that since HDMI connection can be easily established by means of an appropriate adapter while outdated analog interfaces can hardly be necessary for people who buy graphics cards like GeForce GTX 295 in the year of 2009. With fewer connectors, the “second storey” of the card’s mounting bracket is all vent holes, which improves cooling efficiency. Besides that, the card has onboard S/PDIF and MIO headers. Like the older GeForce GTX 295, the new version has only one MIO connector because the second MIO line is wired on the PCB and connects the two GPUs. Thus, the maximum amount of GPUs in a multi-GPU subsystem is 4.
The new GeForce GTX 295 is also different from the old one when it comes to the cooling system. The latter had a sophisticated cooler that looked like an aluminum sandwich, but the new card’s cooler is very elementary.
This cooling system consists of two heatsinks made from thin aluminum plates and connected to the copper bases with a couple of heat pipes. These heatsinks are responsible for cooling the hottest parts of the GeForce GTX 295, namely its GPUs. Considering the reasonable thermal and electrical parameters of the 55nm version of the G200 chip, the design and size of these heatsinks is quite acceptable. They should do their job well enough. Each heatsink is fastened to the cooler’s base with four spring-loaded screws. There is a layer of dark-gray thermal grease between the cooler’s copper sole and the GPU cap.
The other components that require cooling, such as memory chips, voltage regulators, nForce 200 and NVIO2, touch the aluminum frame the fan is installed on. Fiber pads soaked in white thermal grease are used as a thermal interface for the memory chips and power circuit elements whereas the open-die chips contact with the cooler’s base via gray elastic thermal pads. The memory chips on the reverse side of the PCB are cooled with two metallic plates and have the above-mentioned pads with thermal grease. The plates are fastened through the PCB to the cooler’s base.
The cooling system uses an ordinary 4.8W 80x20mm axial fan from Forceon (DFB802012M00T). We could not find its specs but the letters DFB in the marking may mean Dynamic Fluid Bearing, which is supposed to ensure a longer service life and a lower noise level. We don’t know the bearing type for sure, but we will measure the card’s noise shortly. The fan’s airflow goes to both heatsinks but, like with the GeForce 7900 GTX, only the left airflow leaves the system case through the slits in the card’s mounting bracket. The right airflow goes in the opposite direction, i.e. into the system case, which is not good since even one G200b chip produces over 100W of heat despite the 55nm tech process. Anyway, this cooler should be quieter than the composite “sandwich” installed on the old dual-PCB version of GeForce GTX 295. It may also prove to be more effective, but we will check this out shortly.
Summing it up, the cooler of the new GeForce GTX 295 follows the engineering principle of making things simpler whenever possible. We do not know the manufacturing cost of the new and old versions of GeForce GTX 295, but we are sure that the new version must be considerably cheaper just because it has one instead of two PCBs. Thus, the Radeon HD 4870 X2 has got a dangerous rival in terms of both pricing and performance. Perhaps this will make AMD and partners hurry up with the launch of the faster Radeon HD 4890 X2.
The new GeForce GTX 295 has a new power circuit so we retested its power consumption to see how it changed after the changes in design. We used a special testbed configured as follows:
Following our standard procedure, the 3D load was created by the first SM3.0/HDR test from 3DMark06 running in a loop at 1600x1200 with forced 4x FSAA and 16x AF. The 2D load was emulated by the 2D Transparent Windows test from PCMark05. The graphics card worked at its default GPU and memory frequencies.
We’ve got the following results:
Click to enlarge
The new GeForce GTX 295 is more economical than the older version: the difference is 11.3W under load. Perhaps this is the benefit of Nvidia’s EPWM technology implemented in the PWM regulators of GPU voltage. The load is distributed among the power lines in a somewhat different manner, but the 8-pin connector still has the heaviest load of all. The peak power draw of 203W is a good result for a solution with two GPUs each consisting of 1.5 billion transistors. We guess AMD should think about power efficiency of its solutions as the Radeon HD 4870 X2 is less economical and delivers lower performance. And we doubt that a hypothetical Radeon HD 4890 X2 can beat the new GeForce GTX 295.
The new card is also impressive in terms of temperature. Its GPUs were never hotter than 74-78°C despite the summer heat!
The idle mode numbers are very good, too. The older version of GeForce GTX 295 does not even come close to that, notwithstanding its sophisticated cooler. Interestingly, the temperature of the GPU located near the mounting bracket is always 2-4°C lower than that of the other GPU because the hot air from the latter stays within the system case. Anyway, the new GeForce GTX 295 is the best in its class in terms of temperature.
When it comes to noisiness, we got the following results measuring the card’s noise level (at an ambient noise of 43dBA):
The new GeForce GTX 295 is as noisy as the older one in ordinary mode but its noise spectrum is more agreeable to the ear. When the card has been working long under load, its noise grows up to the level of the Radeon HD 4870 X2 which is considered one of the loudest graphics cards. However, the noise spectrum is less annoying. It is not the rattling of the fan bearing together with resonating plastic as with the Radeon HD 4870 X2, but the hiss of the airflow passing through the heatsinks. We guess this noise is not irritating and can be put up with. This is the price you pay for high cooling efficiency and good overclockability.
As for overclockability, we managed to increase the card’s frequencies from its default 576/1242MHz to 675/1455MHz and it was perfectly stable though all of our tests. Although its memory chips have a rated frequency of 1000 (2000) MHz, we overclocked them to an impressive 1250 (2500) MHz. Of course, such a good frequency growth should produce a nice performance boost. Therefore we benchmarked our Inno3D GeForce GTX 295 Platinum at the overclocked frequencies, too.
We are going to investigate the performance of Inno3D GeForce GTX 295 Platinum graphics card using the following testbed:
The graphics card drivers were configured in the following way:
ATI Catalyst:
Nvidia GeForce:
The list of benchmarks includes the following gaming titles and synthetic tests:
First-Person 3D Shooters
Third-Person 3D Shooters
RPG
Simulators
Strategies
Semi-synthetic Benchmarks
We selected the highest possible level of detail in each game using standard tools provided by the game itself from the gaming menu. The games configuration files weren’t modified in any way, because the ordinary user doesn’t have to know how to do it. We made a few exceptions for selected games if that was necessary. We are going to specifically dwell on each exception like that later on in our article.
Besides Inno3D GeForce GTX 295 Platinum graphics card we have also included the following graphics accelerators to participate in our test session:
We ran our tests in the following resolutions: 1280x1024, 1680x1050, 1920x1200 and 2560x1600. Everywhere, where it was possible we added MSAA 4x antialiasing to the standard anisotropic filtering 16x. We enabled antialiasing from the game’s menu. If this was not possible, we forced them using the appropriate driver settings of ATI Catalyst and Nvidia GeForce drivers.
Performance was measured with the games’ own tools and the original demos were recorded if possible. We measured not only the average speed, but also the minimum speed of the cards where possible. Otherwise, the performance was measured manually with Fraps utility version 2.9.8. In the latter case we ran the test three times and took the average of the three for the performance charts.
Starting from version 1.3 we use the game’s integrated benchmarking options together with a custom demo record. Unfortunately, this method does not report the bottom frame rate.
The start is not impressive. At resolutions up to 2560x1600 pixels the GeForce GTX 295 is close to the Radeon HD 4870 X2 and inferior to the Radeon HD 4890 CrossFireX tandem. Even overclocking does not make the card competitive to the couple of RV790 chips working in multi-GPU mode. The situation is different at 2560x1600, the resolution that the GeForce GTX 295 and similar cards are bought for: the Radeon HD 4870 X2 falls behind while the Radeon HD 4890 CrossFireX is only a few fps ahead of the GTX 295. When overclocked, the GeForce GTX 295 is the fastest single-slot solution.
The game uses deferred rendering and does not support FSAA. Instead, it uses integrated antialiasing algorithms that make use of the GPU’s computing resources.
The ATI Catalyst driver we used for this test session does not support CrossFireX mode in the new Call of Juarez. Therefore the performance of the dual-processor solutions from AMD is far from their theoretical best and actually equals that of the single Radeon HD 4870 and 4890. The GeForce GTX 295 is brilliant but redundant. An ordinary GeForce GTX 285 is quite enough to play the game comfortably at 2560x1600.
Crysis is one of those games luxurious graphics cards are bought for. Indeed, the GeForce GTX 295 is one of the few cards capable of delivering a comfortable frame rate at 1280x1024 if you enable maximum graphics quality settings. AMD has nothing to respond with: the Radeon HD 4870 X2 is not fast enough while the Radeon HD 4890 CrossFireX tandem is not easy to use as it requires two PCI Express slots and a CrossFireX-compatible mainboard.
Overclocking is obvious rewarding here. The overclocked Inno3D has good performance at 1680x1050 and catches up with the SLI pair of GeForce GTX 285 cards.
We disabled the integrated frame rate limiter in the game console for the sake of comparing the cards. The game’s built-in benchmarking options do not provide information about the bottom speed, so there is no such info in the diagrams.
The GeForce GTX 295 is not much better than its opponents and is only ahead at 1280x1024 where its high frame rate is redundant. The new card is equal to the Radeon HD 4890 CrossFireX at 2560x1600. Overclocking makes the GTX 295 unrivalled at resolutions of 1680x1050 and higher, but that’s not necessary for playing Quake Wars. It is enough to have a bottom speed of 30fps.
The GeForce GTX 295 has no rivals at 1280x1024 among single-PCB solutions. It is competitive to the Radeon HD 4890 CrossFireX at higher resolutions and is considerably ahead of the Radeon HD 4870 X2. Its frame rate is comfortable even at 2560x1600.
AMD’s solutions are unrivalled at low resolutions but the GeForce GTX 295 overtakes the Radeon HD 4870 X2 at 1920x1200 and goes ahead at the overclocked frequencies. Nvidia’s products are superior at 2560x1600 where they ensure a higher bottom speed.
The game runs on the Source engine and has an integrated benchmark, but the latter does not report the bottom speed information.
The GeForce GTX 295 is lagging behind the Radeon HD 4870 X2 throughout the test and only catches up with the latter at 2560x1600. When overclocked, Nvidia’s solution takes first place, being short of notching 100fps and leaving the GeForce GTX 285 SLI tandem behind.
To achieve a playable speed in this game we disabled FSAA and such resource-consuming options as Sun rays, Wet surfaces and Volumetric Smoke. We use the Enhanced full dynamic lighting (DX10) mode for our test and additionally enable the DirectX 10.1 mode for the ATI cards.
The GeForce GTX 295 delivers a comfortable performance at 2560x1600 but AMD’s solutions can do the same and even surpass the GTX 295 in terms of bottom speed. Oddly enough, everything is completely different at low resolutions where the Radeon HD 4870 X2 and Radeon HD 4890 CrossFireX are inferior to Nvidia’s solutions. Anyway, the new GeForce GTX 295 is better than its opponents in usability, power efficiency and noisiness.
It is clear that the GeForce GTX 295 is more than enough even for playing at 2560x1600, yet it is occasionally slower than the Radeon HD 4890 CrossFireX, especially at 1680x1050. Overclocking helps close this gap and put this card in the lead at 1920x1200, but there is no practical reason to overload the device in such a way. Today’s top-performance multi-GPU solutions offer more than enough speed in this game.
The first thing we can see in the Prince of Persia results is that the GeForce GTX 295 gets a tremendous performance boost through overclocking. The card is 4-11% slower than the Radeon HD 4870 X2 at the default frequencies but gets from 13% (at 1280x1024) to 35% (at 2560x1600) faster when overclocked. It should be noted, however, that the Radeon HD 4890 CrossFireX tandem delivers a comfortable frame rate at 2560x1600 even without any overclocking.
It is hard to find the winner at low resolutions. At a resolution of 2560x1600 ATI’s products deliver a somewhat higher average performance whereas Nvidia’s graphics cards, including the GeForce GTX 295, offer a higher bottom speed. Here, your shopping choice should be based on personal loyalty to the particular brand and on the consumer properties that do not affect performance directly such as noise, power consumption, etc.
As you can see, the single-chip flagship products Radeon HD 4890 and GeForce GTX 285 both make this game playable at 2560x1600. There is no reason to pay extra for expensive and power-hungry dual-chip solutions, let alone systems with three or four GPUs.
It’s like in the previous test: the dual-chip solutions are overkill even at 2560x1600 with 4x FSAA. However, we can see that the GeForce GTX 295 is somewhat slower than the Radeon HD 4870 X2 throughout the test, the gap varying from 4 to 11% depending on the resolution. Overclocking helps close this gap and makes the GeForce GTX 295 as fast as the GeForce GTX 285 SLI tandem but clearly offers no practical reward.
We use the in-game benchmarking tools that do not allow to measure the bottom frame rate. We also enable DirectX 10.1 support for ATI’s solutions.
Nvidia’s solutions are unrivalled in H.A.W.X. The gap from the Radeon HD 4870 X2 and Radeon HD 4890 CrossFireX is tremendous at resolutions up to 1920x1200. The difference is smaller at 2560x1600, yet the non-overclocked GeForce GTX 295 is almost 20% faster than the best single-PCB dual-chip solution from ATI/AMD and overclocking adds 14-15% more to that. The frame rate is overall high enough for the virtual pilot to feel no discomfort.
The game has a frame rate limiter fixed at 30fps.
G200-based products are not good in Red Alert 3 but the lack of speed can be made up for by SLI technology. When FSAA is turned on, the GeForce GTX 295 is almost as fast as the Radeon HD 4870 X2 although its bottom speed is somewhat lower at 2560x1600.
It is not easy to beat the Radeon HD 4890 CrossFireX, yet the GeForce GTX 295 is quite competitive, even though it has a lower bottom speed at resolutions higher than 1680x1050. The bottom speed is below playable at 2560x1600, but overclocking helps solve this problem. Comparing the GTX 295 with the Radeon HD 4870 X2, ATI/AMD’s solution has a comparable bottom speed but is noisier and less economical.
The GeForce GTX 295 is still inferior to the dual-chip solutions from AMD but this benchmark does not support newest 3D graphics techniques and does not allow the new card to show its best. The resolution of 1280x1024 seems to be a limiting factor, too, because overclocking does not increase the card’s performance much.
We minimize the CPU’s influence by using the Extreme profile (1920x1200, 4x FSAA and anisotropic filtering). We also publish the results of the individual tests across all display resolutions to provide a full picture.
3DMark Vantage is quite a different story. Backed up by PhysX support, Nvidia’s solutions feel at their ease here. First place among single-PCB multi-GPU solutions goes to the GeForce GTX 295. When overclocked, it scores over 11,000 points and comes very close to the result of the discrete tandem GeForce GTX 285 SLI.
The GeForce GTX 295 is in the lead in the gaming tests, too. However, the Radeon HD 4890 CrossFireX tandem delivers similar performance at 1920x1200 and 2560x1600 and is even ahead of the Nvidia card at 2560x1600. The overclocked GeForce GTX 295 is only rivaled by the SLI pair of GeForce GTX 285 cards.
By launching the new revision of their GeForce GTX 295 Nvidia proved that their developers can do almost the impossible, namely, turn and extremely complex and expensive graphics accelerator barely fit for mass production into a technological solution with acceptable price tag. I have to confess that we had our reasons to believe that it would be impossible to design a solution like that on G200 cores, even manufactured with 55 nm process. However, Nvidia engineers not just kept the entire potential of the GeForce GTX 295 solution, but also improved some of its consumer qualities such as energy-efficiency and acoustics.
The performance also remained the same, because the main specifications of the GeForce GTX 295 solution, such as clock frequencies and amount of onboard video memory remained unchanged. Our tests confirmed that GeForce GTX 295 is still the king of the hill in the high-performance single-card segment, although not completely without a proviso. Although it faces a dangerous rival, Radeon HD 4890 CrossFireX tandem, it is not quite correct to compare these two solutions because the latter consists of two graphics cards and hence requires a mainboard with two PCI Express x16 slots and corresponding multi-GPU technology support. This comparison will be justified only when we see Radeon HD 4890 X2 hit the streets, and at this point the only opponent for the new GeForce GTX 295 is Radeon HD 4870 X2. This is what things look like between them:
In 1280x1024 average advantage of GeForce GTX 295 over Radeon HD 4870 X2 is about 14 %, but it loses to AMD/ATI solution only in five tests out of fifteen: Call of Duty: World at War, F.E.A.R. 2: Project Origin, Prince of Persia, Race Driver: GRID and Left 4 Dead. We only see it fall seriously behind – by over 15 % - in F.E.A.R. 2: Project Origin, however, it doesn’t affect the gaming comfort at all, because of the very high overall performance level in this game.
If we compare GeForce GTX 295 against a Radeon HD 4890 CrossFireX configuration, the average advantage of the former over the latter will lower to 4%. However, the number of lost rounds remains the same, because tiny 1% lag in Devil May Cry 4 can be in fact disregarded.
Overclocking does help GeForce GTX 295 improve the performance: by raising the GPU frequency to 675/1455 MHz and memory frequency to 1250 (2500) MHz you increase the average performance by 10% and in some cases the gain reaches 27%.
The situation here is almost the same: average advantage of the new GeForce GTX 295 over Radeon HD 4870 X2 makes the same 14 %, although the win over Radeon HD 4890 CrossFireX drops down to 2%. We see the newcomer fall behind in the same five games mentioned above: Call of Duty: World at War, F.E.A.R. 2: Project Origin, Prince of Persia, Race Driver: GRID and Left 4 Dead. When we compare the results against those of Radeon HD 4890 CrossFireX we can also add here Devil May Cry 4.
Overclocking improves the performance of GeForce GTX 295 by 10% average with 24% maximum gain in some applications.
1920x1200 resolution is the real fit for GeForce GTX 295 and Radeon HD 4870 X2 solutions. The advantage of our today’s hero over the competitor here increases to 15% and the number of lost rounds drops to four, with the maximum performance difference never exceeding 10%.
However, when it comes to Radeon HD 4890 CrossFireX tandem, competition becomes really cut-throat: in eight tests out of fifteen ATI solution manages to outperform GeForce GTX 295, and in two more tests they tie. RV790 is a very successful graphics processor and when Radeon HD 4890 X2 finally comes out, GeForce GTX 295 will have really hard times competing against it, no doubt. However, the new solution has very good overclocking potential: we could easily increase the average performance of our particular sample by 13% and in some cases reach as high as 20-35%.
This is a showdown and GeForce GTX 295 again didn’t disappoint us having outpaced Radeon HD 4870 X2 by 10% on average. It lost in only four tests: Crysis Warhead, Prince of Persia, Fallout 3 and Race Driver: GRID. If we take a closer look at each of these four applications separately, we will see that the loss in Fallout 3 and Race Driver: GRID won’t have any consequences for the gaming comfort, and in Crysis Warhead the performance of both graphics accelerators is so low that any gameplay is out of the question. The situation in Prince of Persia stands out because GeForce GTX 295’s defeat can be in fact considered a victory, because despite lower average performance it provides higher minimal speed than Radeon HD 4870 X2.
Unfortunately, the new GeForce GTX 295 lost the battle against Radeon HD 4890 CrossFireX configuration. It yielded about 5% on average to the ATI solution and in Crysis Warhead the gap between them increased to almost 30%. The final score is not very good for the newcomer: 10:4 in favor of AMD/ATI and one tie. The only thing that saves the day for GeForce GTX 295 is the assumption that it is unfair to compare it against two graphics cards tied into a single configuration. But how long will it stand? No doubt that sooner or later a single-PCB graphics card modification with two ATI RV790 GPUs will come out and GeForce GTX 295 will have to live through really hard times, at least if its specs remain what they are today. Increase in clock frequencies could help, and overclocking experiments in 2560x1600 resolution show that this may turn out a very effective measure: in our case the average performance improved by 15% and the maximum gain we got was 35%. I would like to specifically mention World in Conflict: Soviet Assault, where overclocking provides a significant increase in GeForce GTX 295’s minimal performance ensuring comfortable gameplay.
Overall, as we have already said, the new GeForce GTX 295 revision is just as good as the old one and is still the fastest graphics card that requires only one PCI Express x16 slot. Later on, when Radeon HD 4890 X2 comes out things may change, but at this time we can recommend GeForce GTX 295 to all dedicated gamers who look for the ultimate highest performance but do not want to deal with several discrete graphics cards.
Inno3D GeForce GTX 295 Platinum graphics card is an exact copy of the reference Nvidia solution and differs from the other manufacturers’ products by the stickers on the cooler casing. Despite the word “Platinum” in the model name the GPU and memory clock frequencies match the reference ones. However, Inno3D solution does have an advantage that is unique for this graphics solution: the accessories bundle that includes two full versions of games, one of which is a real hit and another one allows you to fully enjoy the advantages of hardware PhysX accelerator. We haven’t seen popular games bundled with graphics cards too often lately. Moreover, there is also a discount coupon for 10% off when you buy up to five games in nZone.com store, which is also a nice bonus.
To cut the long story short, Inno3D GeForce GTX 295 Platinum is a good choice if you are looking for the today’s fastest graphics accelerator and the price is not an issue for you.
Highs:
Lows: