Nvidia’s GeForce 8800 GTS 512 graphics card

Okay boys and girls, it’s time once again for the introduction of a new video card, a happy ritual performed multiple times each year during the lead-up to Christm… err, Holiday. The video card making its debut today is the GeForce 8800 GTS 512, a brand-new big brother to the GeForce 8800 GT, which is selling like a Nintendo Wii infused with the essence of Cabbage Patch Kids, or something like that.

The 8800 GTS 512 looks to supplant older versions of the GeForce 8800 GTS by offering a more attractive mix of price, performance, capabilities, and power consumption. So, you know, it’s the same old boring story in the world of GPUs. The 8800 GTS 512 is interesting in that it has, more or less, no direct competition from the Radeon camp, and its most potent competitor may be its little brother, the 8800 GT. Is there any reason to spend the extra to reach up to the new GTS? Let’s take a look.

Meet the new GTS, different from the old GTS

Despite the name, the GeForce 8800 GTS 512 is actually quite a different animal than previous versions of the GeForce 8800 GTS. If Nvidia were facing more competition in the high-end of the video card market, this product would probably be called the GeForce 8950 GTS or something like that. The new card is based on the same G92 graphics processor that lies under the slim cooler of the GeForce 8800 GT. The G92 traces its design lineage to the G80 GPU that powered the original 8800 GTS cards, but it’s substantially revised, with a new 65nm chip fabrication process, a PCI Express 2.0 interface, more of some things, and less of others.

The most obvious outward indication that the 8800 GTS 512 is a new product, of course, is the amount of memory it carries. Nvidia has decided to key on that fact and let the “512MB” label after the name denote the newer product. Cleverly subtle, I suppose, but to a fault. The 8800 GTS 512 is a better product in some notable ways.

At its default speeds, the 8800 GTS 512 packs quite a wallop. The G92 GPU in the 8800 GTS 512 has four ROP partitions capable of outputting four pixels each, for a total of 16 pixels per clock. Each ROP partition has a 64-bit path to RAM attached, yielding a total memory pathway that’s 256 bits wide. And the GTS 512 has a total of 128 stream processors clocked at 1625MHz.

So what do those numbers mean? Well, in several cases like pixel output capacity and memory bandwidth, the new GTS 512 trails the old GTS in per-clock power. However, the GTS 512 tends to make up any deficits by running at higher clock speeds. Here’s how some of the key stats look once you multiply the per-clock capacity by the clock frequency.

	Peak pixel fill rate (Gpixels/s)	Peak bilinear texel filtering rate (Gtexels/s)	Peak bilinear FP16 texel filtering rate (Gtexels/s)	Peak memory bandwidth (GB/s)	Peak shader arithmetic (GFLOPS)
GeForce 8800 GT	9.6	33.6	16.8	57.6	504
GeForce 8800 GTS	10.0	12.0	12.0	64.0	346
GeForce 8800 GTS 512	10.4	41.6	20.8	62.1	624
GeForce 8800 GTX	13.8	18.4	18.4	86.4	518
GeForce 8800 Ultra	14.7	19.6	19.6	103.7	576
Radeon HD 2900 XT	11.9	11.9	11.9	105.6	475
Radeon HD 3850	10.7	10.7	10.7	53.1	429
Radeon HD 3870	12.4	12.4	12.4	72.0	496

The 8800 GTS 512 beats the original 8800 GTS in all respects but peak memory bandwidth, where the two are very close. And the GTS 512 even surpasses the GeForce 8800 GTX in texture filtering capacity and peak shader arithmetic. That bodes very well for its performance overall.

The one caveat, if there is one, is that the GTS 512 doesn’t offer much more pixel fill rate or memory bandwidth than the GeForce 8800 GT. Thanks to an extra SP cluster, the GTS 512 does have more texture filtering and shader capacity than the GT, but those advantages may not always bring better performance. The G92’s ratio of shader and texture processing capacity to pixel output and memory bandwidth is much different than the G80’s. As a result, pixel fill rate and memory bandwidth are more likely to be the primary bottlenecks, which could make it hard for the GTS 512 to separate itself from the 8800 GT.

If that egghead math stuff doesn’t float your boat, you’d probably still want to pick the 8800 GTS 512 over the original GTS in order to see the pretty pictures from HD movies. Unlike G80-based cards, the 8800 GTS 512 supports HDCP over both links of a dual-link DVI connection—necessary for HD movie playback on very high resolution displays—and it includes Nvidia’s VP2 video processing unit that provides hardware assistance with H.264 decoding. We found that the VP2-equipped GeForce 8600 GTS consumes quite a bit less CPU time during H.264 video playback than the G80-based GeForce 8800 GTX.

The cards

Physically, GeForce 8800 GTS 512 cards look very similar to their predecessors.

EVGA’s take on the 8800 GTS 512

A single SLI connector means two-way SLI is probably the limit

The board is 9″ long, with a dual-slot cooler that reaches that full length. The GTS 512’s funky cooler places the blower at an angle from the card, presumably to allow room for air intake when the card is nestled up against another one. In front are a pair of dual-link DVI ports and an HDTV-out port. Around back is a single six-pin PCIe auxiliary power connector. Nvidia rates the card’s power use at 150W, so it’s just able to make do with a single aux plug. Like the GTS before it, the GTS 512 has just one SLI connector per card, not two like on the GeForce 8800 GTX. That means exotic three- and four-way SLI configurations probably won’t be possible.

The EVGA card pictured above is selling for $359 at Newegg with a bundled copy of Crysis. Like many GeForce cards, this one runs at higher clock speeds than Nvidia’s recommended baseline. The core clock is 670MHz, the shader clock is 1674MHz, and the memory clock is stock at 970MHz.

And here’s XFX’s take on the 8800 GTS 512. XFX plans several versions of the 8800 GTS, including a stock-clocked model for $349 and the one pictured above, the XXX version, which has a 678MHz core, 1728MHz shaders, and 986MHz memory, for $379. Our XXX card came with a bundled copy of Lost Planet: Extreme Condition, which means EVGA wins, in my book. Crysis is a much better game. XFX seems to have upped the ante by offering a stock-clocked version of the GTS 512 packaged with Company of Heroes for $369 at Newegg, however.

Overall, these prices are just a little lower than the going rate for a GeForce 8800 GTS 640MB. As we’ve mentioned, the GTS 512 doesn’t really have any direct competition from AMD, unless you count running a pair of Radeon HD 3850 256MB cards in CrossFire. A couple of those cards would cost about the same as an 8800 GTS 512. However, their effective memory size would be only half that of the GTS 512. We’ve tested this config to see how it compares. Although it’s a more expensive configuration, we’ve also tested a pair of Radeon HD 3870 512MB cards in CrossFire.

Since the EVGA card has the more conservative clock speed of the two GTS 512 cards we had on hand, we decided to use that one in single-card testing. We then paired it up with the XFX for SLI.

Test notes

You may notice that I didn’t engage in a lot of GPU geekery this time around. I decided instead to focus on testing these new video cards across a range of the amazing new games and game engines coming out. These GPUs are basically “refresh” parts based on existing technology, and their most compelling attributes, in my view, are their incredibly strong price-performance ratios.

In order to give you a better sense of perspective on the price-performance front, I’ve included a couple of older video cards, in addition to a whole range of new cards. Roughly a year ago, the Radeon X1950 Pro faced off against the GeForce 7900 GS at $199. This year’s crop of similarly priced GPUs have some substantial advantages in terms of specifications and theoretical throughput, but as you’ll see, the gains they offer in real-world performance are even larger—and they do it while delivering image quality that’s sometimes quite noticeably superior to last year’s models, as well.

You’ll find results for both the X1950 Pro and the 7900 GS in several of our gaming tests and in our power and noise measurements. I’ve had to limit their participation to scripted benchmarks because these cards were generally too slow to handle the settings at which we tested manually with FRAPS.

That leads me to another issue. As I said, the older cards couldn’t handle some of the settings we used because, well, they’re quite intensive, with very high resolutions, quality levels, or both. We tested at these settings because we wanted to push the cards to their limits in order to show meaningful performance differences between them. That’s hard to do without hitting a CPU or system-level bottleneck, especially with cards this fast running in multi-GPU configurations. We did test at multiple quality levels with a couple of games in order to give you a sense of performance scaling, which should help.

Also, please note that many of the GeForce cards in the tables below are clocked at higher-than-stock speeds. Nvidia’s board vendors have made a practice of selling their products at multiple clock speeds, and some of our examples are these hot-clocked variants. For instance, the 8800 GTS cards are all clocked at 575MHz (or in the case of the one XFX 320MB card, 580MHz) core clocks and correspondingly higher shader clocks. Obviously, that’s going to change the performance picture. We think it makes sense to include these cards because they’re typically fairly plentiful and available for not much of a premium over stock-clocked versions. They’re what we might buy for ourselves.

The one exception to that rule, at least right now, may be the GeForce 8800 GT. The first wave of these cards looks to have sold out at many online vendors, and all variants are going for something of a premium right now—especially the higher clocked ones. We have included one “overclocked” version of the 8800 GT (from MSI) in our tests in order to show you its performance. This card is very fast, but be aware that it is not currently a $199 or even a $249 option.

Our testing methods

As ever, we did our best to deliver clean benchmark numbers. Tests were run at least three times, and the results were averaged.

Our test systems were configured like so:

Thanks to Corsair for providing us with memory for our testing. Their quality, service, and support are easily superior to no-name DIMMs.

Our test systems were powered by PC Power & Cooling Silencer 750W power supply units. The Silencer 750W was a runaway Editor’s Choice winner in our epic 11-way power supply roundup, so it seemed like a fitting choice for our test rigs. Thanks to OCZ for providing these units for our use in testing.

Unless otherwise specified, image quality settings for the graphics cards were left at the control panel defaults. Vertical refresh sync (vsync) was disabled for all tests.

We used the following versions of our test applications:

• Unreal Tournament 3 demo
• Call of Duty 4: Modern Warfare
• Enemy Territory: Quake Wars 1.2
• BioShock 1.0 with DirectX 10
• TimeShift 1.2
• FRAPS 2.9.2

The tests and methods we employ are generally publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.

Enemy Territory: Quake Wars

We’ll start with Quake Wars since this game’s simple “nettimedemo” allows us to record a gaming session and play it back with precise repeatability on a range of cards at a range of resolutions. Which is what we did. A lot.

We tested this game with 4X antialiasing and 16X anisotropic filtering enabled, along with “high” settings for all of the game’s quality options except “Shader level” which was set to “Ultra.” We left the diffuse, bump, and specular texture quality settings at their default levels, though, to be somewhat merciful to the 256MB and 320MB cards. Shadows, soft particles, and smooth foliage were enabled where possible, although the Radeon X1950 Pro wasn’t capable of handling soft particles.

The GTS 512 does well enough at lower resolutions, but the pack really starts to separate as the display resolution climbs. At 1600×1200, the GTS 512 is the second-fastest single card in the bunch, just behind the GeForce 8800 GTX and just ahead of the Radeon HD 3870 in CrossFire. The most price-competitive config from AMD, the two Radeon HD 3850 cards in CrossFire, just isn’t in the same league as the GTS 512. Also, the GTS 512 easily outdoes the GeForce 8800 GTS 640MB, in spite of the fact that our GTS 640MB representative is clocked at speeds substantially higher than stock.

The most formidable foe to the GTS 512 is the “factory overclocked” GeForce 8800 GT from MSI. This card is selling for under $300 online and shadows the GTS 512 very closely.

Unreal Tournament 3 demo

We tested the UT3 demo by playing a deathmatch against some bots and recording frame rates during 60-second gameplay sessions using FRAPS. This method has the advantage of duplicating real gameplay, but it comes at the expense of precise repeatability. We believe five sample sessions are sufficient to get reasonably consistent and trustworthy results. In addition to average frame rates, we’ve included the low frames rates, because those tend to reflect the user experience in performance-critical situations. In order to diminish the effect of outliers, we’ve reported the median of the five low frame rates we encountered.

Because the Unreal engine doesn’t support multisampled antialiasing, we tested without AA. Instead, we just cranked up the resolution to 2560×1600 and turned up the demo’s quality sliders to the max. I also disabled the demo’s frame rate cap before testing.

The picture looks much the same in UT3. The 8800 GTS 512 slots in just between the GeForce 8800 GTX and the overclocked 8800 GT from MSI. The big difference here is the performance of the Radeon HD 3850 CrossFire config, which turns out higher average frame rates than the GTS 512. Unfortunately, though, the HD 3850 turns in a much lower median low FPS number. In fact, CrossFire doesn’t appear to help median low frame rates at all on the 3850, which means it’s not a huge help to playability in worst-case scenarios. The same is true of SLI on the GTS 512.

Call of Duty 4

This game is about as sweet as they come, and we also tested it manually using FRAPS. We played through a portion of the “Blackout” mission at 1600×1200 with 4X antialiasing and 16X aniso.

Not to sound like a broken record, but I think we’re getting a sense that the 8800 GTS 512 is faster than the GTS 640MB and nearly as quick as the big daddy, the 8800 GTX. The MSI 8800 GT OC remains the fly in the ointment, however.

Frustratingly, AMD doesn’t appear to have CrossFire working well with CoD4.

TimeShift

This game may be a bizarrely derivative remix of Half-Life 2 and F.E.A.R., but’s it’s a guilty-pleasure delight for FPS enthusiasts that has a very “action-arcade” kind of feel to it. Like most of the other games, we played this one manually and recorded frame rates with FRAPS. We had all of the in-game quality settings maxed out here, save for “Projected Shadows,” since that feature only works on Nvidia cards.

This time, the GTS 512 even outdoes the 8800 GTX in single-card performance, though not in SLI. The GTS 640MB just isn’t as fast, which is one reason I take issue with the naming of the GTS 512. As ever, though, the 8800 GT OC card shadows the GTS 512 closely.

The Radeon CrossFire scores look pretty good here, but I should mention that, although we were able to test, we encountered annoying visual artifacts (the screen was flashing) while playing on the CrossFire rigs, even with AMD’s latest drivers.

BioShock

We tested this game with FRAPS, just like we did the UT3 demo. BioShock’s default settings in DirectX 10 are already very high quality, so we didn’t tinker with them much. We just set the display res to 2560×1600 and went to town. In this case, I was trying to take down a Big Daddy, a generally unsuccessful effort.

The trends we’ve seen before continue unabated here, mercilessly offering me very little to write about. SLI seems to be no help at all in this game for the GTS 512. None of the GeForce cards gain much from SLI in BioShock, but most of them do gain some. We’re using slightly newer drivers on the GTS 512, which may explain the scaling difference.

Power consumption

We measured total system power consumption at the wall socket using an Extech power analyzer model 380803. The monitor was plugged into a separate outlet, so its power draw was not part of our measurement. The cards were plugged into a motherboard on an open test bench.

The idle measurements were taken at the Windows Vista desktop with the Aero theme enabled. The cards were tested under load running BioShock in DirectX 10 at 2560×1600 resolution, using the same settings we did for performance testing.

Although the GTS 512 performs nearly as well as the GeForce 8800 GTX, its power draw is much lower. This is solid improvement on the power-performance front—and nothing less than we expected from a card based on the G92 GPU.

Noise levels

We measured noise levels on our test systems, sitting on an open test bench, using an Extech model 407727 digital sound level meter. The meter was mounted on a tripod approximately 14″ from the test system at a height even with the top of the video card. We used the OSHA-standard weighting and speed for these measurements.

You can think of these noise level measurements much like our system power consumption tests, because the entire systems’ noise levels were measured, including the stock Intel cooler we used to cool the CPU. Of course, noise levels will vary greatly in the real world along with the acoustic properties of the PC enclosure used, whether the enclosure provides adequate cooling to avoid a card’s highest fan speeds, placement of the enclosure in the room, and a whole range of other variables. These results should give a reasonably good picture of comparative fan noise, though.

Few cards stand out from the pack here, and that’s generally a good thing, since they’re almost all nice and quiet. The big exceptions are the GeForce 7900 GS and the Radeon HD 2900 XT, which become noisy when running a game. The GTS 512’s cooler doesn’t suffer that fate.

Then again, neither does the GeForce 8800 GT. One thing’s worth noting, though, before we move on: we tested noise levels on an open test bench. I suspect that, when placed into a hot PC case, the GTS 512’s dual-slot cooler and rear exhaust will lend itself to lower noise levels quite a bit better than the single-slot cooler on the GeForce 8800 GT. As a rule, if we don’t mind sacrificing the extra expansion slot, we tend to prefer dual-slot coolers for this reason.