GraphicsCard
Latest
NVIDIA GeForce GTX 650 and 660 review roundup: hitting the sweet spot, sometimes
If you're building or upgrading a budget gaming rig, it'll be hard to ignore the GeForce GTX 650 and 660. Whether or not NVIDIA's new chipsets are worth the glance is another matter, and early reviews suggest that a sale depends on just which market you're in. The GTX 660, by far the darling of the review crowd, competes solidly against the Radeon HD 7850 by outrunning AMD's hardware in most situations while undercutting on the official price. Only a few have taken a look at the lower-end GTX 650, but it's not as much of a clear-cut purchasing decision -- the entry-level video often slots in between the performance of the Radeon HD 7750 and 7770 without the price edge of its bigger brother. Either card is much better value for the money than the GT 640, however, and looks to be a meaningful upgrade if you're trading up from equivalent prior-generation gear. Read - AnandTech (GTX 660) Read - Benchmark Reviews (GTX 660) Read - Bit-Tech (GTX 660) Read - Guru 3D (GTX 650) Read - HardOCP (GTX 660) Read - Hot Hardware (GTX 660) Read - PC Mag (GTX 660) Read - PC Perspective (GTX 660) Read - Tom's Hardware (GTX 650 and 660)
NVIDIA Quadro K5000 GPU for Mac offers significant Premiere Pro performance boost, we go hands-on
NVIDIA just announced that its new Quadro K5000 GPU will be available on Mac Pros, offering 4K display compatibility and support for up to four displays, not to mention 4GB of graphics memory and about 2x faster performance than the Fermi-based Quadro 4000. While the Kepler-powered chip won't actually hit Apple systems till later this year, we got a first look at the K500 on a Mac here at IBC. NVIDIA demoed Adobe After Effects and Premiere Pro CS6 on a Mac Pro with dual K5000 GPUs. As you'll see in the video below, with 11 streams of 1080p video at 30 fps in Premiere Pro (and one overlay of the NVIDIA logo), GPU acceleration handles the workload seamlessly, letting us add effects in real time without any processing delay. Switching to software rendering mode in the editing program shows a night-and-day difference: video playback is extremely choppy, and processing moves at a crawl. Even with two K5000 chips in this desktop, Premiere Pro utilizes just one, but After Effects takes advantage of both GPUs. In this program, NVIDIA showed us ray-tracing, a computationally intensive 3D imaging feature, which only became available in After Effects with the release of CS6. Like in Premiere Pro, the program runs smoothly enough to let us edit images in real time. Take a look for yourself by heading past the break.
NVIDIA working on Linux support for Optimus automatic graphics switching
Linux godfather Linus Torvalds may have a frosty relationship with NVIDIA, but that hasn't stopped the company from improving its hardware's support for the open-source operating system. In fact, the chipset-maker is working on the OS' compatibility with its Optimus graphics switching tech, which would enable laptops to conserve power by swapping between discrete and integrated graphics on the fly. In an email sent to a developer listserv, NVIDIA software engineer Aaron Plattner revealed that he's created a working proof of concept with a driver. There's no word on when the Tux-loving masses may see Optimus support, but we imagine that day can't come soon enough for those who want better battery life while gaming on their mobile machines.
NVIDIA GeForce GTX 660 Ti review roundup: impressive performance for around $300
No one's saying that $300 is "cheap," but compared to the GTX 670 and GTX 680 before it, the newly announced GeForce GTX 660 Ti is definitely in a more attainable category. The usual suspects have hashed out their reviews today, with the general consensus being one of satisfaction. A gamechanger in the space it's not, but this Kepler-based GPU managed to go toe-to-toe with similarly priced Radeon GPUs while being relatively power efficient in the process. That said, AnandTech was quick to point out that unlike Kepler reviews in the past, the 660 Ti wasn't able to simply blow away the competition; it found the card to perform around 10 to 15 percent faster than the 7870 from AMD, while the 7950 was putting out roughly the same amount of performance as the card on today's test bench. HotHardware mentioned that NVIDIA does indeed have another winner on its hands, noting that it'd be tough to do better right now for three Benjamins. Per usual, there's plenty of further reading available in the links below for those seriously considering the upgrade.
NVIDIA announces $299 GeForce GTX 660 Ti, lets Kepler walk among the people
It's taken NVIDIA a mighty long time to squeeze its Kepler GPU into something more affordable than the GTX 670, but it's finally happened -- the mid-range GTX 660 Ti is out and available to purchase for $299 on boards from EVGA, Gigabyte, ASUS and the usual suspects. Some buyers may complain that's $50 more than the 560 Ti, while others will no doubt be reeling off their CVV codes already. For its part, NVIDIA claims the 660 Ti is the "best card per watt ever made" and that it beats even AMD's higher-priced Radeon HD 7950 at 1920 x 1080. Check out the slide deck below for official stats, as well as for examples of what the card can do with its support for DirectX 11 tessellation, PhysX (particularly on Borderlands 2, which you may well find bundled free) and NVIDIA's TXAA anti-aliasing. We'll wait for independent benchmarks in our review round-up before making any judgment, but in the meantime it's fair to say that this 150-watt card comes fully featured. For a start, it has just as many 28nm CUDA cores as the GTX 670, the same base and GPU Boost clock speeds, the same 2GB of GDDR5 and indeed the same connectivity. The only sacrifice is memory bandwidth: all that computational performance is limited by a 192-bit memory bus, compared to the 256-bit width of the 670. Judging from those specs, we'd expect it to be almost 670-like in performance, and that's going to be pretty impressive.%Gallery-162585%
SIGGRAPH 2012 wrap-up
Considering that SIGGRAPH focuses on visual content creation and display, there was no shortage of interesting elements to gawk at on the show floor. From motion capture demos to 3D objects printed for Hollywood productions, there was plenty of entertainment at the Los Angeles Convention Center this year. Major product introductions included ARM's Mali-T604 GPU and a handful of high-end graphics cards from AMD, but the highlight of the show was the Emerging Technologies wing, which played host to a variety of concept demonstrations, gathering top researchers from institutions like the University of Electro-Communications in Tokyo and MIT. The exhibition has come to a close for the year, but you can catch up with the show floor action in the gallery below, then click on past the break for links to all of our hands-on coverage, direct from LA.%Gallery-162185%
AMD launches its next-gen FirePro graphics card lineup, we go hands-on at SIGGRAPH (video)
Just as you've cozied up with "Tahiti" and "Cape Verde," AMD has returned to grow its "Southern Islands" family of graphics cards with four fresh FirePros, offering up to four teraflops of graphics computing power. That spec can be found in the company's new W9000, which is capable of four TFLOPs single precision and one TFLOP double precision with a price tag just shy of $4,000. That behemoth of a card offers 6GB of GDDR5 RAM and requires 274 watts of power. More humble members of the fam include the W8000, which has the same form-factor as the higher-end W9000, but eases back on the specs, consuming 189 watts of power and carrying a $1,599 price tag. We had a chance to take a closer look at both cards at SIGGRAPH, and while they packed a significant amount of heft, you'll likely never take a second look once they're buried away in your tower rig. Fans of smaller housings (and price tags) may take notice of the W7000 and W5000, which are both considerably more compact and require less power to boot, with pricing set at $899 and $599, respectively. Those cards were also on hand for our demo, and can be seen along with the top two configs in our gallery below. You can also sneak a closer peek in the hands-on video after the break, and glance at the full specs over at our news post from earlier today.%Gallery-161943%
We're live from SIGGRAPH 2012 in Los Angeles!
Most of us experience the Los Angeles Convention Center during one of its most chaotic weeks of the year, when tens of thousands of gaming industry manufacturers, video game designers and consumers descend upon downtown LA for the annual E3 expo, booth-babe radar tweaked to 11. There's a hint of graphics prowess amid the halls this week, too, albeit on a vastly smaller scale, and with a heavy heap of civility. SIGGRAPH is a trade event through and through, with attendees demonstrating their latest tech, taking in a handful of seminars or hunting for networking opportunities, in search of employment and partnerships. It's often also a venue for product launches, which is what's brought us out, along with the usual bounty of kooky creations that serve to entertain and lighten the mood. As always, we'll be bringing you a little bit of everything over the next few days, letting you sample the best of SIGGRAPH from the comfort of your own device -- head over to our SIGGRAPH 2012 tag to follow along.
AMD announces $4k FirePro W9000 GPU, entry-level FirePro A300 APU for CAD and graphics pros
After a brief tease earlier this summer, AMD just announced pricing and availability for its new market-leading FirePro W9000 graphics processing unit -- the $3,999 GPU is available now through AMD resellers, and is compatible with Supermicro SuperWorkstations. Joining that "world's most powerful" rig are the W8000, W7000 and W5000, which sell for $1,599, $899 and $599, respectively, and can each power six 30-inch 4K displays. Power-hungry pros will want to opt for the top-of-the-line model in order to take advantage of four TFLOPs single precision or one TFLOP double precision, along with 6 gigs of high-speed GDDR5 RAM. The W8000, on the other hand, offers 3.23 TFLOPs single precision and 806 GFLOPs double precision, followed by the W7000 with 2.4 TFLOPs / 152 GFLOPs, both with 4 gigs of RAM, along with the W5000, which packs 1.27 TFLOPs single and 80 GFLOPs double, with 2 GB of GDDR5 RAM. Design pros with slightly more modest demands may find the FirePro A300 APU more in line with their budgets -- we don't have precise pricing to share, since third parties will ship their own configs, but terms like "entry-level" and "mainstream" make it clear that you won't be drawing in more than a couple zeros in the checkbook to make your purchase. The integrated solution utilizes AMD's Turbo Core tech, supports Eyefinity and Discrete Compute Offload, and can power horizontal display arrays of up to 10,240 x 1,600 pixels. You'll find all the nitty-gritty in the pair of press releases after the break. Update: Our pals over at HotHardware have just pushed out a review of the W8000 and W9000, but found the results to be a bit of a letdown. Hit up their post for the full skinny.
First GHz Edition Radeon HD 7970 hits shelves, throws $699 wrench into AMD's pricing strategy
The GHz Edition was supposed to deliver a significant mid-cycle performance bump to AMD's flagship 7970, without any attendant rise in cost. Instead, according to AnandTech, it looks like third-party vendors are looking to exploit the GE has a chance to max out every other spec in addition to the updated silicon and then charge a premium. Sapphire's new Toxic card is a case in point -- a 6GB double-helping of VRAM and a "Lethal" BIOS mode that takes base clock up to 1150MHz and memory clock to 6.4GHz (compared to 6GHz on the stock card). Those who can splutter up $699 will surely love it, but it's no substitute for the $499 upgrade that AMD originally intended.
NVIDIA GeForce GTX 680 and 670 round-up: which overclocked card is the one for you?
Happy first-quarter birthday, Kepler. Your reference design is old enough for companies like EVGA, MSI, ASUS, Zotac to push your various clock speeds to the limit, while using custom coolers to avoid meltdown. Arguably the time to upgrade is now and that's why Hot Hardware has done a full-on round-up review, comparing some of the best GTX 680 and GTX 670 packages against each other and against the Radeon HD 7970 GHz Edition. If you prefer your morning news fresh and unspoilered, jump straight to the source link below. Or, if you just want the gist of it, click Read More.
AMD Radeon HD 7970 GHz Edition review roundup: a big, bad bruiser of a graphics card
You know the routine: we looked this revamped card's spec sheet a few hours ago, so now it's time to find out how it fared in independent tests and whether it's worth the $499 outlay. On the face of it, this powerhouse of a card ought to be a champ, since it comes $50 cheaper than the original's launch price (although that non-GHz Edition has now dropped to $449) and brings crucial improvements in clock speed and memory bandwidth. In practice? Well, it wins -- but only on points. Read on for more.
AMD launches Radeon HD 7970 GHz Edition, demands rematch with NVIDIA
If you've been missing out on the graphics card wars of late, then here's a quick rundown. AMD launched its high-end $549 Radeon HD 7970 at the end of last year, and it reigned comfortably for a few months until NVIDIA came out with the masterful GeForce GTX 680. That would have been the end of the matter, at least for this product cycle, except for one crucial factor: time. Having reached the market so much earlier, AMD has now had six months to not only tweak its drivers but also its 28nm silicon. That process has already culminated in 1GHz cards at the low- and mid-ranges, and today it leads to the (slightly predictable) announcement of a Radeon HD 7970 'GHz Edition' -- priced at $499 and expected to be available from a range of board makers from next week. To keep you amused in the meantime, there's plenty of detail in the gallery below and after the break. Update: review roundup added here.%Gallery-158843%
NVIDIA GeForce GT 640 review roundup: a $99 card that fails to keep Kepler's promise
NVIDIA blew reviewers away with its flagship GTX 680, but building cards for the low-end is a whole different challenge. The GT 640 hits the $99 price point, where it sits right between the AMD Radeon HD 7750 at $109 and the last-gen Radeon 6670 at around $80. As it turns out, these rivals pull off something of a pincer movement, leaving NVIDIA's card little room to retaliate. The GT 640 contains a 28nm Kepler chip, but it's a cutdown version of the silicon with just 384 CUDA cores, a poor memory bandwidth of 28.5GB/s and no GPU Boost feature -- which means the card can't boost its 900MHz to make use of thermal headroom. These limitations have a significant impact on the benchmarks, as you'll see in our little review roundup after the break.
AMD announces 900MHz version of the Radeon HD 7750, calls it 'the other 7750'
The original Radeon HD 7750 won plaudits for its low-wattage performance and $109 price tag, but now AMD's decided to put out an additional version with 100MHz more clock whizz. The new card also happens to be called the Radeon HD 7750, and it'll sit alongside its doppelganger on the shelves of the world's computer stores just daring you spot the difference. How will you do that? By checking if it requires a separate power supply: the original 7750 was powered entirely through the PCIe slot, limiting its power to 75W and hence its clock speed to 800MHz, while the new one accepts additional power to deliver slightly more grunt and stay ahead of looming competition from NVIDIA. XFX has already jumped aboard with the 7750 Black Edition Double Dissipation priced at $115 after a rebate, and AnandTech says other vendors will follow suit. So, unless you're totally out of power jacks, you know what to look for. Image courtesy of Newegg.
NVIDIA outs budget GeForce GT 610, GT 620 and GT 630, no Kepler in any of 'em
Look out, savvy graphics card buyer: just because it's labelled 'GeForce' and starts with a '6' doesn't necessarily mean it benefits from NVIDIA's premium 28nm Kepler architecture. We've already seen rebadged mobile chips with last-gen 40nm silicon, and now entry-level desktop cards are arriving on shelves that will stretch Fermi's expiry date even further. There are no price tags as yet, but according to AnandTech the 'new' GeForce GT 610 is a repackaged GT 520 with 48 CUDA cores and an ever-so-polite 29-watt power draw. The GT 620 is a GT 530 with a 49-watt TDP and twice as many CUDA cores as the 610 -- although a meager 64-bit memory bus will put a cap on any performance gains. Finally, the GT 630 is a 65-watt GT 440 in all but name, with a 128-bit memory bus width allowing its 96 CUDA cores to be fully exploited. This latter card shouldn't be confused with the OEM version of the GT 630, which does actually pack Kepler. Bewildering, right? We've quizzed NVIDIA over its strange rebadging tradition and were told that the company simply numbers its products according to raw performance, rather than freshness or chip type -- which sort of makes sense so long as you don't dwell on it.
NVIDIA GeForce GTX 670 review round-up: 'just get here if you can'
While the world still waits for the GTX 680 to reach Newegg, NVIDIA has pushed ahead with the next card down in its stack: the $399 GTX 670. This more affordable option keeps most of the main Kepler credentials intact, but it necessarily makes a few compromises on the computational side, with fewer processing cores (1344 instead of 1536) and texture units (112 instead of 128) as well as slower base clock speed (915MHz instead of 1006MHz). Is that likely to be a problem? Judging from reviewers' responses published today, which cover cards from a range of vendors, probably not. In fact, as TechSpot puts it, "there's very little to critique," because the GTX 670 matches the performance of AMD's flagship Radeon HD 7970 at a much lower price. AnandTech's benchmarks put the reference board only ten percent (or a handful of fps) behind the GTX 680 in many recent games, leaving it "nipping at the 7970's heels," but it was still plenty powerful enough to play Arkham City or Battlefield 3 at 5760 x 1200 with high settings. PCPer's stats put the new card 15 to 20 percent behind the 680, but found good scaling in SLI mode. The Tech Report found the the GTX 670's cheap stock cooler let it down slightly, with a "friction-filled" idle noise well above the top-end Radeons and even above the dual-GPU GTX 690 -- but under load it conducted itself relatively well. We could go on, but ultimately if you're looking to buy this card then you'll want to do your own research at the links below, and then do a raindance. Read - TechSpot Read - AnandTech Read - The Tech Report Read - PC Per Read - HotHardware Read - Tom's Hardware Read - Hexus
NVIDIA outs GeForce GTX 670 GPU: it's Kepler without the mortgage
This'll be old news for some lucky folks, but NVIDIA has just unveiled the GeForce GTX 670 graphics card. It aims to bring Kepler to gamers who don't have off-shore bank accounts, with a price tag of $399 (or £329 in the UK, and €329 in Europe). What sacrifices will be made to reach that bracket, compared to the flagship GTX 680? A loss of 192 CUDA cores, for starters, plus a slightly slower 915MHz base clock speed, which will no doubt have an impact on benchmarks -- we'll do a review round-up shortly to figure out just how much. Nevertheless, you'll still get the same 28nm chip architecture and 2GB of DDR5 RAM, along with NVIDIA's GPU Boost technology that autonomously overclocks the processor to make use of available headroom. In terms of official performance claims, NVIDIA has chosen to compare its benchmarks to AMD's high-end Radeon HD 7950 and boasts that the GTX 670 comes out on top every time by a margin of 18 to 49 percent. Of course, the war of words is little more than performance art at this point, so stay tuned for independent tests. Meanwhile, gaming-friendly manufacturers like Origin and Maingear have declared that they'll carry the card alongside the 690 in its desktop offerings -- you can learn more about that after the jump.
ASUS Zenbook UX32VD unveiled, packs discrete graphics to flaunt that 1080p screen
New high-resolution Zenbooks are on their way, but ASUS has now revealed another addition that's packing its own discrete graphics. Like its recently unveiled siblings, the Zenbook UX32VD still packs a 1920 x 1080 IPS panel and "the latest Intel Core processor" -- likely something from the Ivy Bridge family, although ASUS is keeping schtum on the details for now. Graphical horsepower is delegated to NVIDIA's GT620, while the UX32VD also had a hybrid HDD on show at the event, not the standard SSD -- explaining why this new model is slightly thicker than the UX31A. Alongside a lack of processor specifics, there was no info on prices or release dates at the Taiwan press event but for now, Engadget Chinese have posted their full impressions at the source below. Ultrabook and graphics mavens can hit up the Google translation for a few more details. %Gallery-155120%
AMD Radeon HD 7970 could get 'GHz Edition', put the hurtz on NVIDIA
AMD's Radeon HD 7770 and 7870 reference cards already sport 1GHz clock speeds, but so far the high-end flagship 7970 has been stuck at 925MHz. That'd be no big deal, perhaps, were it not for rival NVIDIA's benchmark-stealing GeForce GTX 680, which autonomously adjusts its clock speed on the fly and easily hits 1.2GHz under the right conditions. But while NVIDIA has yet to roll out its full stack of 28nm cards, AMD is finding plenty of time to play catch-up. According to Australian site Atomic MPC, the company has revealed that the manufacturing process of its next-gen GPUs has improved to the point where the same average voltages can yield much higher clock speeds. Recent chips can reach 1.25GHz without struggling, which means a conservative "GHz Edition" of the 7970 can now safely be rolled out, of course with scope for much higher overclocking on third-party boards with more robust coolers. By the time the battle between Red and Green reaches full-swing, it might not be so easy to call a winner.
