• Nvidia`s Gtx 650 Ti Makes Too Many Inches
• Nvidia's Gtx 650 Ti Makes Too Many Red Blood Cells
• Nvidia Geforce Gtx 650 Ti 2gb
• Gtx 650 Ti 2gb

Can a GeForce GTX 650 Ti Run 3 monitors? Baldey Apr 11, 2013, 1:53 PM. NVIDIA GeForce GTX 650 Ti cannot make surround mode to make 3 monitors behave as 1 large monitor. Jan 20, 2019  Nvidia's small-sized 750 Ti runs only US$149 (or £114, AU$219), half the price of the GTX 760, and just $20 more than AMD's R7 260X, if you account for the 260X's recent price cut. Nvidia’s GTX 650 Ti makes too many compromises for a $170 games-oriented graphics card October 2, 2014 – 18:27 Specifications for the next graphics card in Nvidia’s Kepler-based lineup, the GTX 650 Ti, have leaked.

• Gigabyte GTX 650 Ti in Detail. Additionally, Nvidia's specification sees the memory capacity halved from 2048MB to 1024MB, though Gigabyte has still outfitted its GTX 650 Ti with 2048MB of memory, which is clocked at Nvidia's specified 1350MHz (5400MHz DDR) and therefore falls in line with the company's spec for memory bandwidth at 86.4GB/s.
• GeForce GTX 690, GeForce GTX 680, GeForce GTX 670, GeForce GTX 660 Ti, GeForce GTX 660, GeForce GTX 650 Ti BOOST, GeForce GTX 650 Ti, GeForce GTX 650, GeForce GTX 645, GeForce GT 645, GeForce GT.
• Because my video card, a double width 650 GTX Ti, is on the SECOND Slot of my motherboard, and my motherboard has 4 x x16 slot + (x16 lane EACH), it's Z8PE, $600 motherboard. The reason it's not the on 1st x16 slot is because the Noctua giant 14 in. CPU Cooler is blocking it.
• After years of development, Nvidia's new Turing graphics cards are finally here: the RTX 2080 and RTX 2080 Ti. While the RTX 2080 Ti offers a clear performance advantage over the best of the last-gen Pascal cards (plus a ton of new features) the RTX 2080's broadly similar performance to the GTX 1080 Ti makes it a less obvious choice.

OK, wow, this is awkward. You know, we really can’t keep meeting like this. Every few weeks, it seems like, we’re back in this same place, and I’m telling the same story again. You know the one, where Nvidia has taken its potent Kepler GPU architecture, shaved off a few bits, and raised the performance ceiling for a lower price range. By now, you know how it ends: with me explaining that this new graphics card delivers enough performance for most people and questioning why anyone would spend more. You probably expect me to say something about how the competition from AMD is pretty decent, too, although the Radeon’s power draw is higher. By now, the script is getting to be pretty stale. Heck, I can see your lips moving while I talk.

Well, listen up, buddy. I am nobody’s fool, and I’m not going to keep playing this same record over and over again, like Joe Biden at the DNC. I can do things, you know. I should be, I dunno, explaining write coalescing in the Xeon Phi or editing a multicast IP routing table somewhere, not helping you lot decide between a video card with 10 Xbox 360s worth of rendering power and another with 14. This second-rate website can get a new spokesmonkey.

I’m totally not going to tell you about the video card shown above, the GeForce GTX 660. You can see from the picture that it’s based on the same reference design as the GeForce GTX 660 Ti and GTX 670. And if you have half a working lobe in your skull, you know what’s coming next: the price is lower, along with the performance. Look, it’s as simple as a few key variables.

You really don’t need me for this. Versus the GTX 660 Ti, this ever-so “new” product is a tad slower in texture filtering, rasterization, and shader flops rates. And yes, that really is a drop from 14 Xboxes worth of filtering power to 10. The ROP rate and memory bandwidth haven’t even changed, and yet the price is down 70 bucks. This value proposition doesn’t involve difficult math.

Heck, you probably don’t even care that the card has a mixed-density memory config with three 64-bit interfaces driving 2GB of GDDR5 memory. Who needs to know about that when you’re Calling your Duties or prancing around in your fancy hats in TF2? All you’re likely to worry about are pedestrian concerns, like the fact that this card needs only 140W of power, so it requires just one six-pin power input. I could tell you about its high-end features—such as support for up to four displays across three different input types, PCI Express 3.0 transfer rates, or two-way SLI multi-GPU teaming—but you’ll probably forget about them two paragraphs from now. Why even bother?

A different chip

You know what’s rich? This apparently pedestrian branding exercise actually involves new GPU silicon. They’re calling this thing “GeForce GTX 660,” but it’s not based on the same chip as its purported sibling, the GeForce GTX 660 Ti. That’s right: the GTX 660 is based on the GK106 chip, not the GK104 part that we’ve been talking about for months.

Functional block diagram of the GK106 chip. Source: Nvidia.

This is a smaller, cut-down chip with fewer resources throughout, as depicted in the block diagram above. The unit counts in that diagram are correct for the GTX 660, right down to that third GPC, or graphics processing cluster, with only a single SMX engine inside of it. Is that really the GK106’s full complement of units? Nvidia claims, and I quote, that the GTX 660 “uses the full chip implementation of GK106 silicon.” But I remain skeptical. I mean, look at it. Really, a missing SMX? I know better than to trust Nvidia. I’ve talked to Charlie Demerjian, people.

With its five SMX cores, the GK106 has a total of 960 shader ALUs (calling those ALUs “CUDA cores” is crazy marketing talk, like saying a V8 engine has “eight motors”). Beyond that, look, the specs are in the table, people. The only thing missing is the L2 cache amount, which is 384KB. (Note to self: consider adding L2 cache sizes to table in future.) You’ve probably noticed that the GK106 is just two square millimeters larger than the Pitcairn chip that powers the Radeon HD 7800 series. Seriously, with this kind of parity, how am I supposed to conjure up drama for these reviews?

The GK104 (left) versus the GK106 (right)

I probably shouldn’t tell you this, but since I’ve decided not to do a proper write-up, I’ll let you in on a little secret: that quarter is frickin’ famous. Been using the same one for years, and it’s all over the Internet, since our pictures are regularly, uh, “borrowed” by content farms and such. I’m so proud of little George there.

The cards

Since we’re spilling secrets, here’s something I’ve learned over too many years of doing this job. Did you know you can largely determine the class of video card cooler via a simple visual inspection? True story: I’ve been told each heatpipe on these things adds 50 cents to the manufacturing cost. If you watch, card makers will economize meticulously. For instance, we absolutely loved the five-pipe cooler on MSI’s R7870 Hawk, pictured farther down the page here. The thing is darn near silent, even when running a game. When they built the GeForce GTX 660 Ti Power Edition card, MSI used a cooler that shares the same shroud and looks nearly identical—but has one fewer heatpipe, presumably since the lower-power GTX 660 Ti doesn’t produce quite a much heat. Fiddy cents, people. That’s what it’s all about.

That’s also why it’s a bit surprising to see the cooler pictured above on Asus’ GTX 660. I count five heatpipes under that plastic shroud. Yet there’s only one six-pin power plug on the card, an indication it pulls less than 150W of power. You know what that means? Somewhere in cubeville, an Asus engineer won a fight with a bean counter. Maybe clocked the dude with his slide rule. Good for him. And good for us, since this puppy barely has to spin those fans in order to keep the GK106 beneath it cool.

Then again, I’m not sure the bean counters aren’t in cahoots with the engineers on this one. You see, this card is full of upgrades over Nvidia’s reference design. The circuit board is custom, with six power phases (up from four) and a host of premium “super alloy” electronics components, because apparently, the reference-grade capacitors are made out of tin foil and fermented apple juice. There’s even a digital VRM. Meanwhile, the GPU chips have been binned, so they start out at 8% faster than the reference clock and purportedly have extra overclocking headroom. Asus includes a GPU Tweak program that lets users adjust the GPU boost clock, voltage, memory clock, power targets, and fan speeds. The trick here—and this is where the bean counters get their satisfaction—is that Asus charges $249.99 for this baby, 20 bucks more than the GTX 660’s base price. Is it worth paying the extra for this card and cooler? Just wait ’til you see our acoustic and thermal results. You’ll be ready to Amazon Prime that thing, hard. Just remember, for Asus: fiddy cents. Maybe several times over, with all of the upgrades, but still.

The Asus card measures 10.5″ to the tip of its cooler, while the Zotac is just 7.5″

Zotac, meanwhile, has taken the exact opposite approach with this rendition of the GTX 660, following Nvidia’s reference design closely, presumably right down to the MOSFETs made from weasel whiskers coiled around mothballs and dipped in solder. (I dunno. They still seem to work.) Even the reference card’s pretender-extender plastic cooling shroud is gone, revealing the tiny PCB in all its 7″ glory. The benefit here is the price, which is bone stock, a penny shy of $230, even though Zotac ups the GPU’s base and core clocks from 980/1033MHz to 993/1050MHz. Another benefit is Zotac’s sweet-looking “angry bumblebee” cooler. Although it has only two heatpipes, that flat-black shroud is made of sturdy metal. In a pinch, you could seriously stab an intruder with its pointy tip.

These differences in clock speed and even GPU brands add up to only a few gigatexels of filtering power or fractions of a teraflop within the range of 30 bucks or so. That’s like half the price of a game, for goshsakes, for a single Xbox worth of texture filtering.

Speaking of games, Nvidia has decided not to bundle the GTX 660 with a free copy of Borderlands 2 like it does for its GTX 660 Ti cards. That’s unfortunate, because I plan to waste nearly a week of “testing” time on BL2 later this month—it’s the most anticipated game of the year, in my book. I’ll tell the guys I’m doing important innovation in the field of latency-based game testing, but in reality: pew pew. Then I’ll pull a mean squares-based overall performance metric outta nowhere, and nobody’s the wiser.

Meanwhile, as you can see in the table above, the GTX 660 is at least theoretically a very close match for AMD’s Radeon HD 7870 graphics cards in most of the key rates. The 7800 series ships with a coupon for GTA-alike Sleeping Dogs, the game that’s already kept TR’s Geoff Gasior up late enough that his news posts the next day required major rewrites. At least we love our work.

Here’s a picture of the MSI R7870 Hawk, one of the 13 different video cards we tested, since I’m evidently a masochist. Not that the R7870 is painful to use at all. In fact, it’s formidable competition for the GTX 660—like you couldn’t have guessed that.

The upgrade scenario

(Gpix/s)

Speaking of pain, some of you unfortunate folks are probably still saddled with a graphics card like Ye Olde 9800 GTX listed at the top of the table there, and you may be contemplating an upgrade. That’s good. You should know that your five Xbox 360s of texture filtering oomph are seriously dated. The GTX 660 offers twice as many Xboxes now.

To give owners of older cards a sense of what improvement an upgrade might bring, and to add some sense of actual drama to this otherwise-sorry exercise, we’ve included a handful of older cards in our testing. The really old pre-DX11 cards could only run a couple of the games at the settings we used, Skyrim and Arkham City. Since we tested those games at the sweet, sweet Korean monitor resolution of 2560×1440, I had to scrap plans to include a 9800 GTX and was forced to use a GeForce GTS 250, instead. The GTS 250 is based on the same G92 chip but has 1GB of RAM instead of 512MB, so it has at least a chance of handling the higher resolution without bumping into memory capacity issues. Would-be upgraders who own any G92-based card—including the GeForce 8800 GTS 512, 8800 GT, 9800 GT, and 9800 GTX—will want to watch how the GTS 250 stacks up against the newest GeForces and Radeons. Just assume your card would be panting even harder than the GTS 250, and you’ll have the basic idea.

The somewhat newer and more powerful GeForce GTX 260 is also on our slate, as are several DX11 cards initially from the GTX 660’s price range, including the Radeon HD 6870, GeForce GTX 460, and GTX 560 Ti. I’m sure I’ve slighted somebody who wants to know how his Radeon 9700 AGP compares, but before you post an angry comment, worry not: Damage Labs offers personal concierge comparison testing for the video card of your choice. Just ship the card and a valid check for $5,000 to our P.O. box, and we’ll take care of you ASAP. Or right after we finish Borderlands 2, at least.

Another new GeForce: the GTX 650

Believe it or not, the GTX 660 isn’t the only new GeForce poking its head above the ground today. Nvidia is also unveiling the GeForce GTX 650, a much cheaper card that, mystifyingly, still carries the premium “GeForce GTX” brand. Like the GeForce GT 640 that we reviewed a while back, the GTX 650 is based on the GK107 graphics processor, a Kepler derivative slimmed down in the extreme. The GK107 has only a single GPC, two SMX cores, eight pixels per clock worth of ROP throughput, and a 128-bit memory interface. Crucially, though, the GTX 650 comes with GDDR5 memory. The GT 640 costs $99 and uses only DDR3, with under half the bandwidth, and it fared rather poorly in our testing. At 10 bucks more, the $109 GTX 650 should be quite a bit faster—which is good, since it’s a direct competitor to the Radeon HD 7750, a card that clowned the GT 640 in our last round of tests.

Since it’s Kepler-based, the GTX 650 has an up-to-date array of features, including PCI Express 3.0. As a low-end part, however, it lacks fancy extras like GPU Boost and SLI support. Because the GTX 650 is based on such a teensy Kepler derivative, the card requires only 64W of power at peak—or so they say. A TDP that low raises the question of why a six-pin aux power connector is still onboard. Hrm. I’ll have to ask Charlie about that.

Anyway, we may have to take a look at a GTX 650 at some point soon. Like the GTX 660, they’re supposed to begin selling at online retailers today, but unlike the GTX 660, Nvidia chose not to supply advance review units to the press—so you know it’s gonna be compelling. Eh, works for me, since I can pawn off that review on Cyril.

Our testing methods

As ever, we did our best to deliver clean benchmark numbers. Tests were run at least three times, and we’ve reported the median result.

Our test systems were configured like so:

Thanks to Intel, Corsair, and Gigabyte for helping to outfit our test rigs with some of the finest hardware available. AMD, Nvidia, and the makers of the various products supplied the graphics cards 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:

• Fraps 3.5.6
• GPU-Z 0.6.4

Some further notes on our methods:

• We used the Fraps utility to record frame rates while playing either a 60- or 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 Yokogawa WT210 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 2560×1440 with the Ultra quality presets, 4X MSAA, and FXAA enabled.

• We measured noise levels on our test system, sitting on an open test bench, using an Extech 407738 digital sound level meter. The meter was mounted on a tripod approximately 10″ 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.

Battlefield 3

I cannot believe we are still talking about this stuff. I said I wasn’t gonna do this, but here we are. Might as well chat about the performance results, since I spent many hours slaving away to collect ’em.

We tested Battlefield 3 with all of its DX11 goodness cranked up, including the “Ultra” quality setting with both 4X MSAA and the high-quality version of the post-process FXAA. Our test was conducted in the “Kaffarov” level, for 60 seconds starting at the first checkpoint.

You can click on the buttons above to see frame-by-frame rendering times from a single, representative test run for each card. There’s a surprise horror show if you click the “legacy GeForce” button, because older, Fermi-based GeForces have a persistent problem with intermittent high-latency frames in certain areas of BF3, including the area we tested. These spikes probably wouldn’t show up so dramatically in an FPS-based plot that averages things out over one-second increments—which is why we’ve been avoiding that mistake for a while now. Notice that the newer GeForces and the Radeons don’t have the same problem; their frame time plots are smoother and more consistent, which is what you’d want to ensure smooth, fluid animation.

Incidentally, if you’re confused by our latency-focused benchmark results, read this for an intro to our approach.

I like to show the traditional FPS average results right next to our latency-focused 99th percentile frame time because they ought to track together. If they don’t, that usually means there’s a problem the FPS average didn’t capture. The 99th percentile frame time is a simple concept: for each card, 99% of all frames were rendered in x milliseconds or less. Done properly, with a large enough data set, this number can serve as decent shorthand for overall performance.

As you can see, the FPS average and 99th percentile results largely mirror each other. Today’s subject, the GTX 660, is represented by both the reference card and Asus’ TOP card. The two straddle the competition, MSI’s R7870 Hawk, in both performance metrics. However, the older GeForces look much worse when we consider their 99th percentile frame times. That’s not a big surprise given all of the spikes we saw in the frame-by-frame plots. Playing on those cards isn’t a great experience, a fact the FPS average doesn’t fully convey.

We can plot the full curve of frame times to get a sense of the larger latency picture. When we do, it’s apparent the older GeForces are somewhat slower than the new ones pretty much throughout the test run, but most especially in the last three perecent of frames or so—a consequence of those big latency spikes.

Flip over and compare the GTX 660 to the R7870, and you’ll see that the shape of their latency curves is incredibly similar—and quite good, with nice, low frame times into the last one to two percent. The GTX 660 reference card and the Asus GTX 660 TOP bracket the 7870 throughout.

We can quantify “badness” by looking at the time spent rendering long-latency frames beyond a certain threshold. We generally start with a 50-ms cutoff, because anything that takes longer than that is almost certainly not helping the cause of smooth animation. 50-ms frame times equate to a steady-state rate of 20 FPS.

As you can see, only a handful of the cards spend any time beyond that crucial threshold, and the worst offenders are those older GeForces. If we ratchet the threshold down to 33 ms—equivalent to 30 FPS—all of the current-gen cards are still incredibly competent. Dial down to 16.7 ms, and you’ll see that none of the cards produces a steady stream of frames at 60 FPS. However, the Zotac GTX 670 AMP! comes pretty close. The GTX 660 cards again surround the 7870, essentially tied. Yawn.

The Elder Scrolls V: Skyrim

Our test run for Skyrim was a lap around the town of Whiterun, starting up high at the castle entrance, descending down the stairs into the main part of town, and then doing a figure-eight around the main drag.

We set the game to its “Ultra” presets with 4X multisampled antialiasing. We then layered on FXAA post-process anti-aliasing, as well. We also had the high-res texture pack installed, of course.

Since Skyrim doesn’t use DX11, all of the legacy cards can participate here. However, we really are pushing all of the cards with 1GB of memory pretty close to the edge by testing at these settings. You can see the occasional spikes in the plots for cards like the GTX 460 and GTX 560 Ti, likely caused by memory capacity issues. Then again, I’m not sure it matters for cards like the GTS 250 and GTX 260, since their plots are riddled with latency spikes throughout. They’re just not up to this challenge.

The Radeon HD 7870 has a solid lead over the GTX 660 cards in this test scenario. Notice, however, how there’s a pretty good-sized gap between the top cards and the mid-pack entrants in the FPS sweeps? That gap shrinks to almost nothing, just a few milliseconds or less, with the latency-sensitive 99th percentile metric. That’s likely because we’re running into another limitation, like CPU performance. Now, we have one of the fastest gaming processors available in our test system, so this limitation is going to be pretty common no matter what your PC config. In scenarios like this one, spending more on a faster graphics card would arguably be a waste.

Don’t be fooled by the different scales on the vertical axis for the legacy cards versus the newer ones. We had to use a higher peak value to fit the curves for the GTS 250 and such on the plot. You can see the spread between the older cards and the GTX 660 here. The upgrade benefits are more pronounced the further back you go.

Click over to the newer GeForces, and you can see how they converge together in the last 10% of frames, so that the Asus GTX 660 TOP offers essentially equivalent performance to the much pricier Zotac GTX 670 AMP! in the toughest portions of the test run. Click to the next plot, and you’ll see the Radeons converging, too. At the very tail end of the curve, though, the R7870 and friends maintain lower frame times than the GTX 660 cards.

Here’s where we get an even bigger dose of perspective: none of the current cards waste any time beyond the 33-ms threshold, and most of them don’t spend a substantial amount of time working on frames beyond 16.7 ms. The “badness” numbers escalate quickly for the older cards, in part because we’re probably bumping up against a memory size limit, as we’ve noted.

Batman: Arkham City

We did a little Batman-style free running through the rooftops of Gotham for this one.

This game does have some DirectX 11 features, but personally, I don’t think the tessellated trash bags add much—and the DX11 mode tends to crush performance. We’ve stuck with testing the DX9 path, so the older GeForces are once again along for the ride.

All of the plots are pretty spiky here, since we’re moving through a large portion of the city, forcing the game engine to stream in new content periodically. Flip between them, and you’ll see that the Radeon HD 7870’s plot shows more and larger spikes than the GTX 660 cards’ plots do.

The 7870 doesn’t really pay for its spiky frame times in the FPS average results, where it just barely trails the two GTX 660 cards. The 99th percentile frame time is a different story, though.

The Radeon HD 7870’s curve shoots upward sharply for the last four percent of frames, while the GTX 660 cards’ curves are smoother and their rise more gradual.

Once again, we are largely splitting hairs between the current Radeons and GeForces, simply because they’re all quite capable of running this game smoothly. The only real loser here is the GeForce GTS 250, which is outmatched and overwhelmed.

Max Payne 3

We should note a couple things about Max Payne 3. As you’ll notice in the settings image above, we tested with FXAA enabled and multisampling disabled. That’s not the most intensive possible setting for this game, and as you’ll soon see, Max 3 runs quite quickly on most of the cards we’ve tested. We wanted to test with MSAA, but it turns out multisampling simply doesn’t work well in this game. Quite a few edges are left jagged. Even the trick of combining MSAA with FXAA isn’t effective here. Enabling both disables FXAA, somehow. We couldn’t see the point of stressing the GPUs arbitrarily while lowering image quality, so we simply tested with the highest quality setting, which in this case was FXAA.

Also, please note that this test session wasn’t as exactly repeatable as most of our others. We had to shoot and dodge differently each time through, so there was some natural variation from one run to the next, although we kept to the same basic area and path.

Click through ’em all, and you’ll see that nearly all of the cards deliver virtually all frames in 30 milliseconds or less. This is one of the best-looking PC games of the past year, but it’s not especially demanding, as these things go.

This one is a modest but straightforward victory for the 7870 over the GTX 660. All of the latency curves are nice and flat, but 7870’s is quite a bit lower than the 660’s across the plot.

Click buttons for the first two plots above. Go ahead, click ’em.

Yeah, zero cards above either threshold. I threw those results in for free. No extra charge. You’re welcome.

DiRT Showdown

We’ve added the latest entry in the DiRT series to our test suite at the suggestion of AMD, who has been working with Codemasters for years on optimizations for Eyefinity and DirectX 11. Although Showdown is based on the same game engine as its predecessors, it adds an advanced lighting path that uses DirectCompute to allow fully dynamic lighting. In addition, the game has an optional global illumination feature that approximates the diffusion of light off of surfaces in the scene. We enabled the new lighting path, but global illumination is a little too intensive for at least some of these cards.

This is a fantastic game, by the way. My pulse was pounding at the end of each 90-second test run.

Uh oh. The latency plots for all of the GeForces are squiggly messes. None of them fare well with this game’s advanced lighting path.

You can see the GeForce cards’ struggles illustrated above. In addition the stark turn upward in latency for the last five percent of frames, they deliver higher frame times generally, from the 50th percentile on up.

Nvidia`s Gtx 650 Ti Makes Too Many Inches

Despite these issues, the newer GeForces offer a decent and quite playable experience in this test scenario. Even the older ones don’t spend much time beyond our 50-ms cutoff. Do you . . . sense a theme developing here?

Crysis 2

Our cavalcade of punishing but pretty DirectX 11 games continues with Crysis 2, which we patched with both the DX11 and high-res texture updates.

Notice that we left object image quality at “extreme” rather than “ultra,” in order to avoid the insane over-tessellation of flat surfaces that somehow found its way into the DX11 patch. We tested 60 seconds of gameplay in the level pictured above, where we gunned down several bad guys, making our way across a skywalk to another rooftop.

Uh oh. This time, the latency spikes are on the Radeon side of the fence, while the plots for the GeForces are nearly pristine.

The FPS averages you’ll get from, well, almost every other source make it look like the performance of the 7870 and the GTX 660 is a close match. Our nefarious methods tell us otherwise, because the Radeons trip over more long-latency frames.

Yep, the curves illustrate the difference. Also, again, the GTX 660 is substantially quicker than any of the legacy GeForces we tested.

Nvidia's Gtx 650 Ti Makes Too Many Red Blood Cells

Even the spiky Radeons rise to the challenge of our 50-ms threshold, though. Poof. Mind blown.

Power consumption

This first measurement comes from when the display has been placed in power-save mode while the system is idle. When that happens, the newer Radeons invoke AMD’s ZeroCore power feature and drop into a very low-power state, spinning down their fans in the process. That’s why the Radeons draw so very little power here. Somewhat remarkably, the systems equipped with the reference and Zotac GTX 660 cards only draw about 4W more than the Radeon-equipped configs.

At idle, the GK106 draws very little power, although the Asus GTX 660 TOP pulls more juice than the other two cards. I guess weasel hair and mothballs can be pretty effective, huh?

When running Skyrim, the GTX 660 looks to be very efficient. The same system with a Radeon HD 7870 card installed draws 40W more than the Asus GTX 660 TOP.

Noise levels and GPU temperatures

Even though the Radeon HD 7000-series cards disable their fans entirely when the display is off, our measurements for them aren’t much quieter than the Asus GTX 660 TOP card, probably because we’re getting close to the ambient noise floor for our test room, which is nothing special in terms of sound isolation (unless you consider a basement special).

The reality is that many of these cards are quiet enough at idle that you’d barely notice them in the average room. We’ve seen real progress on this front in recent years, as the higher readings for the GeForce 200-series cards attest.

I knew the Asus GTX 660 TOP was quiet from using it and from the readings at idle, but its noise levels under load are something else entirely. Somehow, the card bends time and space, shaving 0.8 decibels off of its noise level at idle—and lowering the noise floor of the entire test system and environment at the same time. The really freaky thing is that this strange result pretty much matches my subjective impressions. I need to check the label on that new allergy medicine.

Oh, and the Zotac cooler is pretty good for its size. MSI’s 7870 Hawk registers just a little higher on the decibel meter, which is a surprise, since it seems blessedly quiet subjectively.

Here’s what an extra dose or two of fiddy cents buys. Asus’ five-pipe cooler is magically quiet under load, yet it maintains some of the lowest GPU temperatures of the bunch. MSI’s R7870 Hawk isn’t far off on either front, but the Asus card has less heat to dissipate due to its more power-efficient GPU.

Conclusions

So once again, it comes down to this. We bust out a scatter plot of price and performance, summing up all of that work in one or two images. Our overall performance numbers come from the geometric mean of the scores across five of the six games we tested. (We chose to exclude DiRT Showdown, since the results skewed the average pretty badly and since AMD worked very closely with the developers on the lighting path tested.) We’ve translated our 99th percentile frame times into their FPS equivalents for the sake of readability. As ever, the best values will trend toward the upper left portion of the plot, and the worst to the bottom right. Have a look:

If you flip between the two plots, you’ll see several clear outcomes. One is the close price-performance parity overall between Nvidia and AMD at present. AMD’s price cuts after the GTX 660 Ti launch last month saw to that. Another crystal-clear result is the marked improvement with this latest generation of GPUs. The leap from the GTX 460 to the GTX 660 is formidable, and the jump from the Radeon HD 6870 to the 7870 ain’t bad, either. We tested at 1920×1080 resolution and above, with very high (if not always peak) image quality settings in some of the most intensive recent games, and this year’s crop of GPUs aced the test. You saw it in how infrequently any of them wasted time beyond our 50-ms threshold in the prior pages, and you can see it the 99th percentile plot above. We’ve ruled out the toughest 1% of frames, and for everything else, these cards are pumping out a steady stream of frames at rates above 40 FPS. That’s pretty phenomenal—and, yes, it calls into question why anyone should spend 300 bucks or more on a higher-end video card. Seems to me like the GTX 660 should suffice for a great many folks.

The other unmistakable outcome is that the GeForce cards tend to be a little bit stronger in the latency-focused 99th percentile frame time results. The FPS scatter plot paints a picture of almost perfect parity, with a nearly straight line slicing through the various Radeons and GeForces of the current generation as we step up the price-performance ladder. That’s probably no surprise, since pricing on these things tends to be carefully calibrated by the GPU and video card makers. However, they’re using the wrong benchmarks. If you’re focused on smooth gameplay and not just nice frame rate averages, Nvidia’s Kepler-based GPUs currently have the upper hand.

Nvidia Geforce Gtx 650 Ti 2gb

Given all of that, I probably can’t avoid doling out at least one Editor’s Choice award, and I figure the Asus GTX 660 TOP is the best pick. Obviously, it’s very well situated on our value scatter plots. Asus slapped an excellent cooler onto this card, and combined with the GK106 chip’s modest power draw, it’s almost magically quiet. Frankly, I’m a sucker for that. The Zotac GTX 660 is very good and plenty quiet, too, but if I have to pick one card, I’d pony up the extra for the Asus. You know, if I were committed to doing a review.

Please, help me waste time by following me on Twitter.

Gtx 650 Ti 2gb

• 2PCS Cooler Fan for MSI RX470 480 570 580 GTX1080 1070 1060 960 GAMING Card Sets

  $18.98

  $19.98

  Shipping: + $0.98 Shipping

• 2Pcs DC Cooling Fan for MSI GTX1080 1070 1060 960 RX470 480 570 580 GAMING Card

  $18.98

  $19.98

  Shipping: + $0.98 Shipping

• 2PCS Cooler Fan for MSI RX470 480 570 580 GTX1080 1070 1060 960 GAMING Card VS

  $18.98

  $19.98

  Shipping: + $0.98 Shipping

• 2* Cooler Cooling Fan for MSI GTX1080 1070 1060 960 RX470 480 570 580 GAME Card

  $18.98

  $19.98

  Shipping: + $0.98 Shipping

• Certified RAM for Apple Mac mini Core i7 2.7GHz Mid-2011 BTO/CTO 16GB kit 2x8GB

  $98.83

  Shipping: + $14.32 Shipping

• AAAwave 14 GPU Open Air Mining Rig Case Frame BTC Ethereum Cryptocurrency Coin

  $42.99

  Shipping: + $55.95 Shipping

• 32GB (4X8GB) DDR4 2133MHz For Kingston HyperX FURY PC4-17000 DIMM Desktop RAM US

  $150.47

  $158.39

  Free shipping

• CORSAIR Vengeance 16GB (2x8GB) 204-Pin DDR3 SO-DIMM DDR3 1600 (PC3 12800)... New

  $83.53

  Shipping: + $12.88 Shipping

• Cooling Fan For Radeon Red Dragon AX RX 480 470 580 PowerColor Red Devil RX470

  $24.14

  Free shipping

• (2x8GB) Certified RAM for Apple Mac mini Server Core i7 2.6GHz Late 2012 BTO/CTO

  $98.83

  Shipping: + $14.32 Shipping

• For FURY Hyperx 16GB(2X8GB) PC4-17000 DDR4 2133Mhz 1.2V Desktop DIMM Memory RAM

  $79.79

  $83.99

  Free shipping

• 32GB(4X8GB) PC-17000 2133MHZ desktopory NON-ECC DDR4 for HyperX FURY

  $166.88

  Free shipping