Nvidia is a master of marketing, so when they “quietly” launched the GeForce GT 610, 620, and 630 into the retail channel late last week, we knew something was up. It turns out of the three new cards, none of these are actually based on the most recently released Kepler architecture behind the GTX 670, 680, and 690, and are in reality based on the last generation designs. We knew Nvidia was already rebranding Fermi parts for use in OEM laptops and desktops, however it looks like the practice will again carry forward to the aftermarket parts as well.
As you all know, Intel announced the launch of its much anticipated 22nm Ivy Bridge processors earlier today. Intel is counting on these third-generation Core processors for the success of ultrabooks, which it hopes will be able to check the rampant growth of tablets and, in the process, conquer a large chunk of the mobile PC market. But ultrabooks will not be the only products to make use of Ivy Bridge chips; there will also be plenty of all-in-ones, desktops and notebooks. In all, over 570 Ivy Bridge-toting systems are expected to ship in 2012. The MSI GT70 and GT60 are two such products. Hit the jump for more.
Remember all the hoopla leading up to Nvidia's Fermi launch? We were teased with leaked photos, benchmarks, and several delays due to reported defects. Nvidia eventually ironed out whatever bugs it needed to in order to get Fermi to market in the form of a GTX 480, a fast videocard with a group of stream processors disabled. It also ran hot and a little bit loud, ultimately leading us to declare the the GTX 580 "the real Fermi" (see our review here). We're expecting a much smoother rollout to Fermi's successor, though it appears delays are still part of the game.
A new generation of GPUs from Nvidia and AMD has hit the streets. Both camps are offering incredible performance and the widest array of features ever before seen in graphics cards. But, inevitably, each side brings its own unique strengths and weaknesses. What better way to determine the performance champ than by letting this season’s new crop of cards duke it out in the various price categories?
Call this the littlest Fermi. The Asus ENGTS450 is built on Nvidia’s GTS 450 GPU, a scaled-down version of Nvidia’s flagship architecture. The GTX 450 is its own chip, not a feature-reduced version of a large GPU. Asus’s GTS 450 card includes only a single dual-link DVI connector, plus VGA and HDMI outputs; only two can be simultaneously active. This is a newer design than the XFX HD 5770, launched in September 2010.
Nvidia’s GeForce GTX 580 is what the original should have been: quieter, full-featured, faster and more efficient.
When Nvidia launched the GTX 480 -- code-named the GF100 -- early this year, the new GPU proved to be something of a mixed bag. It was undeniably fast, but also crippled – every GTX 480 GPU shipped with a full functional unit disabled. Whether that was because of yield or power issues wasn’t clear. Power clearly was a problem – Nvidia’s flagship ran hot and loud.
Given the competition, Nvidia had to get Fermi out the door. Even before the original Fermi left the building, Nvidia’s engineers were heads-down, respinning and reengineering the GF100. The result is the GF110. The new GPU is, as Emperor Palpatine might put it, “fully operational”, with all functional units now enabled.
Hit the jump for a detailed analysis of the GTX 580.
If NextComputing sounds familiar, it's because this is the same company that recently unveiled a monster rig consisting of three displays, 11TB of storage, two Xeon processors, and up to 16GB of ECC RAM. Now the company says it plans to use Nvidia's Quadro 6000 Fermi cards (PNY brand) in its line of small form factor workstations and servers.
"NextComputing's high-performance portable systems combine workstation-class performance with laptop-like mobility for users who need to run demanding applications in the field," NextComputing says. "Systems like the Radius can house a fast Intel Core i7 processor, up to 24GB RAM, multiple hard drives with RAID, and the Nvidia Quadro 6000 - all in a transportable system the size of a briefcase, and with or without a high-resolution 17-inch display. The NextDimension series and Vigor series portables add even more computing power with dual enterprise-class Intel Xeon processors, up to 48GB RAM, and further storage and PCI expansion capability, in both commercial and military rugged form-factors."
The Quadro 6000 line provides 448 CUDA processing cores, a 6GB frame buffer with GDDR5 ECC RAM, 1.3 billion triangles per second of raw computing power, and more.
Just because you might be saddled with a tight budget doesn't mean you have to resort to integrated graphics. Hell, you can even equip your system to handle DirectX 11 eye candy without spending big bucks. Enter Nvidia's GeForce GT 430, the latest addition to its Fermi family that costs less than a C-note.
For around $80 (give or take depending on make and model), Nvidia is pitching its GeForce GT 430 primarily at digital media PCs rather than a dedicated gaming box. The GPU comes built on a 40nm manufacturing process and includes 96 CUDA cores, a 700MHz core clockspeed, 1GB of DDR3 clocked at 900MHz, a 128-bit memory bus, 4 ROPs, and 585 million transistors.
It also supports HDMI 1.4a for 3D content, HD 24-bit multi-channel audio up to 192KHz, and of course Nvidia's PhysX technology. And while it's not intended as a gaming powerhouse, Nvidia claims you can expect up to 1.5x the performance of "previous generation products" and "playable framerates in all of today's top 30 games, when compared to integrated graphics solutions."
Nvidia on Monday announced a couple of new additions to its Fermi-based Quadro professional graphics series, including the mid-range Quadro 2000 with 192 CUDA processing cores and the entry-level Quadro 600 with 96 CUDA processing cores.
According to Nvidia, the Quadro 2000 packs 1.5 times the geometry performance of the previous generation Quadro parts, resulting in "dramatically higher performance across leading CAD and DCC applications such as SolidWorks and Autodesk 3ds Max."
The less powerful Quadro 600 is a half-height card boasting the "industry's best performance per watt for applications such as Audodesk AutoCAD 2011," Nvidia says.
Both cards sport 1GB of dedicated graphics memory and include the usual modern-day marketing bullets, including OpenGL 4.1, DirectX 11, Shader Model 5.0, DirectCompute, and OpenCL.
The Quadro 2000 and 600 are available now for $600 and $200, respectively.
As a hobby, building PCs is no place for the fickle minded, not unless their wallets run deep. No sooner do you power on your DIY build, you find every part inside is already old news, or soon will be. And if you just picked up a Fermi card, well, this applies to you too.
During Nvidia's GPU Technical Conference, the graphics chip maker revealed its GPU roadmap for the next several years. About to go into production is "Kepler," Fermi's successor that will come built on a 28nm manufacturing process.
"We expect to go into production later next year, the design is progressing vvery rapidly. There are hundreds of engineers working on it," said Jen-Hsun Huang, CEO of Nvidia.
Barring any delays, Kepler is due to arrive in the second-half of 2011. Two years later, Nvidia plans to release a part codenamed "Maxwell," which the company says will offer two to three times the performance of Fermi.