Nvidia’s Kepler architecture has displayed excellent, sometimes record-breaking performance since its grand debut in March. The GeForce GTX 680 was the fastest single-GPU graphics card until AMD recaptured the mojo last week, and the GeForce GTX 690 is still the undisputed king of single-card solutions, thanks to its dual GPUs (and prodigious thousand-dollar price tag).
In spite of these successes, Kepler has seemed eerily reluctant to get its feet wet in the shallow waters of the low-end market. Until earlier this month, the cheapest Kepler card Nvidia could offer cash-strapped enthusiasts was the GeForce GTX 670—which, at $400, really isn’t very cheap at all. Anything more affordable was derived from Nvidia’s older Fermi architecture and manufactured using a 40-nm fab process. All the while, AMD acolytes have had a whole lineup of state-of-the-art, 28-nm Radeons to choose from, including the $109.99 Radeon HD 7750 that came out back in February.
It couldn’t last…
…and it didn’t. In early June, Kepler finally jumped in the kiddie pool aboard the GeForce GT 640, an offering situated firmly in budget territory. The card launched at $99 and currently retails in the $99.99-109.99 range, almost squarely opposite the Radeon HD 7750. Nvidia makes a couple other versions of the GT 640, too, but those are for PC vendors, and you won’t find them listed at e-tailers like Newegg and Amazon.
In any case, Nvidia has given buyers a very real option if they want Kepler goodness minus the lofty price tags. That, in turn, raises an important question. How well does the GeForce GT 640 fare against other budget cards vying for supremacy around the hundred-dollar mark? By that we mean not just the Radeon HD 7750, but also Nvidia’s own, previous-gen offering.
More simply put, is the GeForce GT 640 a new budget wonder, or are you better off spending your lone Benjamin on another card? Let’s find out.
Here’s the star of our show, sans heatsink. Note the tiny GPU. That’s the GK107, which packs 1.3 billion transistors into a scant 118 mm² footprint using TSMC’s 28-nm fab process. The chip is slightly smaller than the Cape Verde GPU that powers AMD’s Radeon HD 7700 series. Cape Verde measures 123 mm² and plays host to 1.5 billion transistors, though some of those transistors are disabled on the 7750, which has two fewer compute units than the 7770.
The GK107 is also much smaller than Nvidia’s 40-nm GF106 and GF116 graphics chips, which drive the GeForce GTS 450 and GeForce GTX 550 Ti, respectively. Both of those older parts host just under 1.2 billion transistors in a die area of about 240 mm². Chips with physically smaller dimensions can cost less to produce, so although the GTS 450 is priced in the same neighborhood as the GT 640 now, Nvidia may have more freedom to apply future price cuts to the latter. So far, however, the GT 640 seems to be staying put at around $99.
Peer inside the GK107 with an electron microscope, and you’ll see lots of transistors and gates arranged in all kinds of crazy patterns. The diagram above provides an easier-to-parse overview of the chip’s various bits and pieces.
The GK107 features a single graphics processing cluster containing dual SMX shader multiprocessors. For reference, the GK104 chip inside the GeForce GTX 680 has four GPCs and eight SMXs. On both chips, each GPC contains two SMX units and a raster engine capable of rasterizing one triangle per clock. Each SMX has 192 arithmetic logic units (ALUs) and texture units capable of filtering 16 texels per clock cycle. The GK107’s lone GPC is backed by a single ROP partition capable of producing 16 pixels per clock. The chip also has dual 64-bit memory controllers that give the GeForce GT 640 a 128-bit path to its 2GB of DDR3 memory.
Hold on a minute—DDR3?
Yes, believe it or not, Nvidia has equipped this card with DDR3 RAM instead of the speedier GDDR5. It’s pretty sluggish DDR3, too, with an effective transfer rate of only 1782 MT/s. A version of the GT 640 with GDDR5 RAM does exist, but it’s one of those pesky cards reserved for PC vendors and not available for sale to the general public.
It doesn’t take a profound understanding of GPU architectures to guess that DDR3 could needlessly hamstring the GeForce GT 640 compared to its GDDR5-toting rivals. The peak theoretical numbers below lend weight to that notion:
|GeForce GTS 450||13||25||25||0.7||783||98|
|GeForce GTX 550 Ti||22||29||29||0.7||900||98|
|GeForce GT 640||14||29||29||0.7||900||29|
|Radeon HD 7750||13||26||13||0.8||800||72|
While the GeForce GT 640 compares favorably to AMD’s Radeon HD 7750 in terms of peak shader, texturing, and rasterization throughput, it falls considerably short when it comes to memory bandwidth, with only 29GB/s to the Radeon’s 72GB/s. We’ll gauge the real-world performance implications of that shortcoming in a minute, but it certainly doesn’t bode well. Graphics cards today need plenty of memory bandwidth to juggle textures and frame data.
The GT 640 may have one minor trump card, and that’s its tight power envelope. Nvidia rates the card for peak power draw of only 65W, while AMD says the Radeon can draw up to 75W. The GeForce GTS 450 and GTX 550 Ti aren’t even in the same league, with respective TDPs of 106W and 116W.
Our GeForce GT 640 sample was kindly volunteered by Zotac. Here it is again, this time with its single-slot heatsink and fan still firmly attached:
Zotac is known for its amped-up, er, AMP! Edition cards, but this GT 640 couldn’t be more sober. Its GPU runs at the standard 900MHz speed defined by Nvidia, and its two gigabytes of DDR3 memory are clocked at 891MHz, per the official specs. Zotac’s cooler at least looks different from the one on Nvidia’s reference design. The heatsink seems to have a fair bit more metal, which likely helps keep temperatures lower. The fan looks to be about the same size, though. It’ll have to spin quickly to generate decent airflow, and the accompanying noise may not blend into the background whoosh of a quiet desktop PC. We’ll look at noise levels in a bit.
The Zotac GeForce GT 640 sells for $109.99 at Newegg right now, which is in line with the prices of other GT 640 variants. We’re going to compare it to a stock-clocked Radeon HD 7750 and MSI’s slightly souped-up version of the GeForce GTX 550 Ti, the GTX 550 Ti Cyclone. For reference, vanilla 7750 variants sell for as little as $99.99 (or $89.99 after a mail-in rebate) at Newegg, and MSI’s GTX 550 Ti Cyclone costs $124.99 (or $109.99 after MIR) at Amazon.
Our testing methods
As ever, we did our best to deliver clean benchmark numbers. Tests were run at least three times, and we reported the median results. Our test systems were configured like so:
|Processor||Intel Core i5-750|
|North bridge||Intel P55 Express|
|Memory size||4GB (2 DIMMs)|
|Memory type||Kingston HyperX KHX2133C9AD3X2K2/4GX
DDR3 SDRAM at 1333MHz
|Memory timings||9-9-9-24 1T|
|Chipset drivers||INF update 18.104.22.1680
Rapid Storage Technology 10.5.0.1026
|Audio||Integrated Via VT1828S
with 22.214.171.12400a drivers
|Hard drive||Crucial RealSSD C300 256GB|
|Power supply||Corsair HX750W 750W|
|OS||Windows 7 Ultimate x64 Edition
Service Pack 1
|Driver revision||GPU core
|MSI GeForce GTX 550 Ti Cyclone||GeForce 304.48 beta||950||1075||1024 (GDDR5)|
|Zotac GeForce GT 640||GeForce 304.48 beta||900||900||2048 (DDR3)|
|AMD Radeon HD 7750||Catalyst 12.6 beta||800||1125||2048 (GDDR5)|
Thanks to Asus, Corsair, Crucial, Kingston, and Intel for helping to outfit our test rigs with some of the finest hardware available. AMD, Nvidia, and the makers of the graphics cards used for testing, as well.
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 test applications:
- Batman: Arkham City
- Battlefield 3
- The Elder Scrolls V: Skyrim
- LuxMark 1.0 Win64
- ShaderToyMark 0.2.0
- TessMark 0.3.0
- Fraps 3.5.5
- GPU-Z 0.5.8
Some further notes on our methods:
- We used the Fraps utility to record frame rates while playing a 90-second sequence from the game. Although capturing frame rates while playing isn’t precisely repeatable, we tried to make each run as similar as possible to all of the others. We tested each Fraps sequence five times per video card in order to counteract any variability. We’ve included frame-by-frame results from Fraps for each game, and in those plots, you’re seeing the results from a single, representative pass through the test sequence.
We measured total system power consumption at the wall socket using a P3 Kill A Watt digital power meter. 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 desktop with the Aero theme enabled. The cards were tested under load running Skyrim at its High quality preset.
We measured noise levels on our test system, sitting on an open test bench, using a TES-52 digital sound level meter. The meter was held approximately 8″ from the test system at a height even with the top of the video card.
You can think of these noise level measurements much like our system power consumption tests, because the entire systems’ noise levels were measured. 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.
- We used GPU-Z to log GPU temperatures during our load testing.
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.
Before we get into game benchmarking, let’s run through some synthetic tests to get a better sense of the GT 640’s capabilities.
|GeForce GTS 450||25||25||98|
|GeForce GTX 550 Ti||29||29||98|
|GeForce GTX 550 Ti Cyclone||30||30||103|
|GeForce GT 640||29||29||29|
|Radeon HD 7750||26||13||72|
The GeForce GT 640’s low memory bandwidth is no handicap in these texture filtering tests. Kepler’s relatively strong texturing performance puts the GT 640 ahead of the Radeon HD 7750, especially in FP16 filtering, where the AMD card’s peak theoretical throughput is half its integer rate.
|GeForce GTS 450||783||98|
|GeForce GTX 550 Ti||900||98|
|GeForce GTX 550 Ti Cyclone||950||103|
|GeForce GT 640||900||29|
|Radeon HD 7750||800||72|
In this tessellation test, the GT 640 manages to outpace the 7750 and the older GeForce GTX 550 Ti Cyclone—a card that has both greater memory bandwidth and higher peak theoretical rasterization performance. This isn’t the first time we see a Kepler part zoom ahead of the competition in this test.
|GeForce GTS 450||0.7||98|
|GeForce GTX 550 Ti||0.7||98|
|GeForce GTX 550 Ti Cyclone||0.7||103|
|GeForce GT 640||0.7||29|
|Radeon HD 7750||0.8||72|
The GT 640 falls behind in these tests, which are chiefly designed to gauge its shader performance. Considering these three cards all have roughly comparable peak theoretical throughput, the GT 640’s low memory bandwidth is likely a bottleneck here.
Of course, the Radeon HD 7750 has a huge advantage in LuxMark regardless, even over the GeForce GTX 550 Ti. We’ve seen a similar pattern with high-end Radeon HD 7800- and 7900-series GPUs stacked up against the GTX 680 and older GeForces in this test. Nvidia has told us it doesn’t optimize its drivers for LuxMark, which likely explains the discrepancy.
Let’s now see how things shake out in some actual games.
The Elder Scrolls V: Skyrim
Our Skyrim test involved running around the town of Whiterun, starting from the city gates, all the way up to Dragonsreach, and then back down again.
We tested at a 1080p resolution using the “Medium” detail preset. (The “High” preset was playable on the GT 640, but it wasn’t completely smooth.)
Now, we should preface the results below with a little primer on our testing methodology. Along with measuring average frames per second, we delve inside the second to look at frame rendering times. Studying the time taken to render each frame gives us a better sense of playability, because it highlights issues like stuttering that can occur—and be felt by the player—within the span of one second. Charting frame times shows these issues clear as day, while charting average frames per second obscures them.
For example, imagine one hypothetical second of gameplay. Almost all frames in that second are rendered in 16.7 ms, but the game briefly hangs, taking a disproportionate 100 ms to produce one frame and then catching up by cranking out the next frame in 5 ms—not an uncommon scenario. You’re going to feel the game hitch, but the FPS counter will only report a dip from 60 to 56 FPS, which would suggest a negligible, imperceptible change. Looking inside the second helps us detect such skips, as well as other issues that conventional frame rate data measured in FPS tends to obscure.
We’re going to start by charting frame times over the totality of a representative run for each system—though we conducted five runs per system to sure our results are solid. These plots should give us an at-a-glance impression of overall playability, warts and all. (Note that, since we’re looking at frame latencies, plots sitting lower on the Y axis indicate quicker solutions.)
The GT 640 exhibits much higher and more frequent latency spikes than the Radeon HD 7750 and the GeForce GTX 550 Ti in our Skyrim test. It also produces less consistent frame times overall (as evidenced by the fatter-looking plot), which is exactly the opposite of what we want.
We can slice and dice our raw frame-time data in other ways to show different facets of the performance picture. Let’s start with something we’re all familiar with: average frames per second. Though this metric doesn’t account for irregularities in frame latencies, it does give us some sense of typical performance.
Next, we can demarcate the threshold below which 99% of frames are rendered. The lower the threshold, the more fluid the game. This metric offers a sense of overall frame latency, but it filters out fringe cases.
Of course, the 99th percentile result only shows a single point along the latency curve. We can show you that whole curve, as well. With single-GPU configs like these, the right hand-side of the graph—and especially the last 5% or so—is where you’ll want to look. That section tends to be where the best and worst solutions diverge.
Our raw frame-time plots show several big spikes above 80-ms for the GT 640, and those are reflected in the latency curve.
Finally, we can rank solutions based on how long they spent working on frames that took longer than 50 ms to render. The results should ideally be “0” across the board, because the illusion of motion becomes hard to maintain once frame latencies rise above 50-ms or so. (50 ms frame times are equivalent to a 20 FPS average.) Simply put, this metric is a measure of “badness.” It tells us about the scope of delays in frame delivery during the test scenario.
No matter how you slice it, the results are clear. In Skyrim, the GT 640 isn’t just slower than the Radeon HD 7750 overall; it’s also worse at keeping frame times consistent, which means it doesn’t do as good a job of maintaining the illusion of motion.
From a seat-of-the-pants perspective, Skyrim doesn’t feel completely fluid on the GT 640. While the game responds well to input, movement that should be smooth is punctuated by skips at short and regular intervals. Perhaps the GPU and its anemic memory configuration are to blame, or maybe it’s just a driver optimization issue. Either way, playing Skyrim on the GT 640 just isn’t a very good experience, even at this rather low detail preset.
Batman: Arkham City
We grappled and glided our way around Gotham, occasionally touching down to mingle with the inhabitants.
Arkham City was tested at 1080p using the “Medium” detail preset. Medium FXAA was enabled, as well, but DX11 effects were left disabled.
Frequent latency spikes are common in this particular stretch of Arkham City, likely because the game has to stream parts of the city on the fly—and this is a very rich and detailed environment.
Ignoring those spikes, we’re seeing basically the same pattern we saw in Skyrim. The GT 640 is clearly much slower than the Radeon HD 7750 and the GeForce GTX 550 Ti, as evidenced by its shorter plot (higher frame rates generate more frames in total) and frequent latency spikes. The Kepler-based GeForce’s frame times are much less consistent, too.
Yeah, the GT 640 isn’t really even in the same league as the other two cards.
Out of the three, the Radeon HD 7750 probably delivers the best overall experience. Its average frame rate may not be the highest, but it’s close enough, and the card’s frame times are more consistent, with fewer spikes than the GTX 550 Ti Cyclone.
We tested Battlefield 3 by playing through the start of the Kaffarov mission, right after the player lands. Our 90-second runs involved walking through the woods and getting into a firefight with a group of hostiles, who fired and lobbed grenades at us.
BF3 is a demanding game, and the GT 640 struggled with it until we bumped down the graphical preset to “Low.” We stuck with a 1080p resolution, however.
In BF3, the GT 640 doesn’t seem to have a problem maintaining relatively consistent frame times. However, those frame times are clearly higher overall than on the other cards.
Happily, none of the cards spent 50 ms or more on a single frame. Once we lower our threshold to 33.3 ms, we see the Radeon once again does a better job than its rivals of avoiding unusually long frame latencies. Considering its average FPS rate is close to the monitor’s refresh rate, I think we can give this victory to the Radeon.
The 7750 sips less power than the GeForce GT 640 at idle—especially when the display goes to sleep. AMD’s ZeroCore tech switches the GPU to an ultra-low-power state and disables its cooling fan in that scenario. Under load, however, the GT 640 manages the lowest power draw of the bunch. I suppose that’s not entirely unexpected, given that the GT 640 also happens to have the tightest thermal envelope of the three.
Noise levels and GPU temperatures
The numbers from our decibel meter are all pretty close, but to my ear, neither the GT 640 nor the 7750 sound particularly quiet. Their tiny fans both emit a sort of high-pitched hiss, while the bigger spinner on the GTX 550 Ti makes a fainter whooshing noise that’s much easier to tune out. That said, keep in mind our Radeon HD 7750 isn’t a retail offering; it’s a reference card from AMD, and most (if not all) the models in stores have different coolers.
Zotac’s little single-slot cooler keeps the GeForce GT 640 nice and chilly. The reference AMD heatsink and fan on the Radeon HD 7750 don’t do such a good job, on the other hand. Here’s hoping retail versions of the 7750 have beefier coolers that keep temperatures lower.
Let’s round things out with a couple of our famous scatter plots. We’re laying average performance (based on the results from the games we tested) along the Y axis and prices along the X axis. The sweet spot will be the card closest to the top left of the plot, while the worst will be closer to the bottom right. We fetched prices from Newegg and Amazon.
We can also compile a value scatter plot out of our 99th percentile frame time data. For consistency’s sake, we’ve converted the frame times to frame rates, so desirable offerings are still at the top left.
I think it’s pretty clear the retail GeForce GT 640 doesn’t belong anywhere near its introductory $99 price point. AMD’s Radeon HD 7750 delivers substantially higher performance for the same amount of money, and it does so using silicon that’s not much larger or hungrier for power. For folks shopping for a latest-generation, 28-nm GPU in this price range, the choice is clear.
The GeForce GT 640 feels like so much wasted potential. I can’t help but think it would have been much more compelling had Nvidia paired the retail card with GDDR5 RAM instead of DDR3. As we saw a few pages back, the GK107’s peak rates are all in the same league as the 7750’s, so a GDDR5-equipped version of the GT 640 could conceivably keep up with its Radeon rival. Heck, it might even be a little quicker here and there. With DDR3, however, the Nvidia card is hopelessly hamstrung.
In the end, it’s likely we’ll see the GT 640 quietly and unceremoniously work its way down the price ladder to a more reasonable level—maybe, say, $60 or $70. At the same time, I’d be surprised if the GK107 GPU on this card didn’t eventually pair up with GDDR5 RAM in a faster budget GeForce better equipped to compete at $99. The chip seems to be capable, and Nvidia should be able to loosen the power envelope and boost the clock speeds.