NVIDIA's new family of graphics processors
(gizmag) -- NVIDIA’s new family of GeForce GTX 200 graphics processors (GPUs) - which includes the GeForce GTX 280 and GeForce GTX 260 GPUs - include NVIDIA CUDA technology and the new CUDA runtime for Windows Vista, which allows programmers to offload the most intensive processing tasks from the CPU to the NVIDIA GPUs. The GPUs also deliver 50% more gaming performance over the Company’s previous NVIDIA GeForce 8800 Ultra GPU with impressive shading horsepower at resolutions as high as 2560 x 1600.
Rendering 3D images in real-time is just about the most mathematically intensive task a PC will ever undertake, but as PC applications become increasingly visual, many ordinary tasks will benefit from the graphics horsepower provided by the GPU, including editing photos, simply running a new operating system such as Windows Vista or encoding and playing high-definition videos. Recognizing this the new cards feature NVIDIA’s PureVideo HD Technology, which delivers a combination of HD video decode acceleration and post-processing for unprecedented picture clarity, smooth video, accurate color, and precise image scaling for movies and video. The GPUs also take advantage of massively parallel, general-purpose computing architecture to transcode high-quality video 18 times faster than with CPU-only implementations.
An increasing number of applications are also being written that use the GPU for straight, non-graphical computational tasks. Stanford B-School-Isnt-What-It-Used-To-Be University’s distributed computing computational program Folding@Home, combines the computing horsepower of millions of consumer GPUs to simulate protein folding to help find cures for diseases such as Alzheimer’s or Parkinson http://en.wikipedia.org/wiki/Parkinson ’s. With the computing processing power of the GeForce GTX family, applications such as Folding@Home and others can run upwards of 140 times faster on an NVIDIA general-purpose parallel processor than on some of today’s traditional CPUs.
But gamers need not fear NVIDIA is ignoring its gaming roots. NVIDIA believes the CUDA general-purpose parallel processing mode will usher in a new generation of ultra realistic games by using the power of the “CUDA Computing” mode to simulate realistic, physically accurate effects then render images using the “GeForce GPU” mode. NVIDIA GPUs are also the only ones to support PhysX technology, the world’s most pervasive physics engine that allows developers to incorporate effects such as rigid body dynamics, collision detection, and cloth simulation. The cards’ 3-way NVIDIA SLI Technology also offers impressive performance scaling by implementing 3-way alternate frame rendering (AFR) to provide what it claims is the world’s fastest gaming solution under Windows Vista.
Graphics cards featuring GeForce GTX 280 GPUs are available now while Graphics cards featuring GeForce GTX 260 GPUs will be available starting on Thursday, June 26, 2008. Suggested retail pricing for the GeForce GTX 280 and GeForce GTX 260 GPUs are USS$649 and $399, respectively. For further info visit NVIDIA.
18 June 2008
NVIDIA unleash GeForce GTX 200 GPUs
AMD's new Firestream chip tops 1 teraflop
(reghardware) -- The latest round of graphics card dueling between AMD and Nvidia isn't just over high-end gamers. The vendors will also exchange blows for the hearts and wallets of your friendly neighborhood medical imagers, seismic modelers, and computational fluid dynamisists.
AMD is refreshing the FireStream processor line with a new general purpose GPU (GPGPU) that boasts more than one teraflop of processing power.
Instead of handling gaming or graphics operatins, GPGPUs are built to crunch hundreds of parallel calculations per clock cycle. They promise massive speed improvements over a CPU in mathematical workloads of the scientific, educational and high performance computing variety.
AMD claims developers are reporting up to a 55x performance increase on financial analysis codes as compared to using a CPU alone.
AMD's new card succeeds the FireStream 9170, released last November and capable of 500 gigaflops for single precision performance. The FireStream 9250 is capable of 1 teraflop for single precision calculations, or 1 trillion floating-point operations per second.
Nvidia is similarly rolling out its 240-core Tesla-10 Series GPGPU chip that's also capable of 1 Teraflop of computational muscle.
12 June 2008
AMD announces physics partnership with Havok
(custompc) -- Surprise move sees AMD cooperating with Intel’s physics branch to get Havok acceleration on both CPUs and GPUs
Many people predicted a sticky situation for AMD when it came to physics acceleration after Intel bought Havok, and AMD even admitted that talks had broken down between Havok and AMD after the acquisition. It looked even worse when Nvidia bought Ageia, and AMD even started talking about using Stanford’s Brook language for physics. However, in a surprise move, AMD has now announced that it’s back on track to offer gaming physics with Havok.
In a statement, AMD put aside its rivalry with Intel and claimed that ‘by working together, both companies are demonstrating their commitment to open standards and continued support for the needs of the game community.’ AMD’s senior vice president and general manager of AMD’s graphics products group, Rick Bergman, described Havok as ‘the clear market leader in physics software’ and said that AMD would be working to optimise its ‘platforms to consistently deliver the best possible visual experience to the gamer.’
AMD is hoping to accelerate Havok Physics on both its multi-core CPUs and GPUs and claims that it’s ‘striving to deliver the best of both worlds.’ However, the main focus at the moment appears to be AMD’s CPUs. AMD and Havok say that they’re planning to optimise the ‘full range of Havok technologies on AMD x86 superscalar processors,' and AMD claims that ‘Havok Physics scales extremely well across the entire family of AMD processors.’
Havok’s managing director, David O’Meara, explained the priority for CPUs, saying that ‘the feedback that we consistently receive from leading game developers is that core game play simulation should be performed on CPU cores.’ However, he added that GPU physics acceleration could become a feature in the future, saying that ‘the capabilities of massively parallel products offer technical possibilities for computing certain types of simulation. We look forward to working with AMD to explore these possibilities.’
So where does this leave Nvidia with its GPU-acceleration of PhysX via CUDA? Will game developers happily work with two physics APIs? Let us know your thoughts.
09 June 2008
NVIDIA Introduces GeForce 9M Graphics Processors
(efytimes) -- NVIDIA has introduced GeForce 9M series of graphics processors and Hybrid SLI technology. NVIDIA GeForce 9M family of GPUs, is the next-generation mobile graphics processor, that delivers up to 40 per cent faster performance than the previous generation and 10x the performance of integrated graphics solutions.
In addition, the new GPUs also feature Hybrid SLI technology, that enables two NVIDIA GPUs, one low-power and one high-performance, to work cooperatively in the same PC to deliver two features —GeForce Boost and HybridPower.
"Beginning this summer, GeForce 9M GPUs and Hybrid SLI, paired with AMD and Intel CPUs, will enable a new breed of notebooks," said Jeff Fisher, senior vice president, GPU business, NVIDIA. "These new notebooks will be optimised to deliver a visual experience and raw computing performance that traditional cookie-cutter notebooks with integrated graphics simply can't touch."
NVIDIA GeForce 9M family of GPUs also include PureVideo HD video processing features for improved colour and contrast; full support for the latest Blu-ray advancements including BD-Live and dual-stream video playback; multi-display connectivity with support for DVI, HDMI 1.3, Display Port 1.1, and VGA; and support for MXM version 3.0 graphics module specification.
02 June 2008
Sparkle Introduces GeForce 9600 GSO 768MB
Taipei, Taiwan (hardwarezone) -- SPARKLE Computer Co., Ltd. introduces the Sparkle GeForce 9600 GSO 768MB Graphics Card for mainstream market. Sparkle GeForce 9600 GSO 768MB Graphics Card doubles the video memory capacity, providing better DX10 gaming performance for mainstream DX10 gamers.
SPARKLE GeForce 9600 GSO 768MB Graphics Card is the first product to be introduced in the GeForce 9600 family gives gamers the 768MB video memory to play cutting-edge DirectX 10 games at a price they will love and can afford. It features 550MHz core speed, 1800MHz memory speed with 192-bit 768MB GDDR3 video memory. With 96 extreme fast stream processors at 1375MHz and doubled 768MB video memory, so compared with ordinary GeForce 9600 GSO grahics cards with 384MB video memory on market, the SPARKLE GeForce 9600 GSO 768MB Graphics Card can deliver much more amazing floating point processing power for unmatched DX 10 3D performance in latest DX10 games, such as Gears of War, Crysis, Bioshock and so on.
Also with the new compression technology, the SPARKLE GeForce 9600 GSO 768MB Graphics Card will lead about 20% than previous G80 products in latest DX10 games. The SPARKLE GeForce 9600 GSO 768MB Graphics Card features dual heatpipe fan cooling system using high technology. Compared with conventional fan cooling system, it features higher throughput thanks to high efficiency heat-pipe optimized for mass heat volatilization. It has high performance thermal compound ensures optimal thermal dissipation even after years of use.
The SPARKLE GeForce 9600 GSO 768MB Graphics Card improves high-definition video playback on everyday PCs by leveraging NVIDIA PureVideo HD technology to deliver highquality playback of HD DVD and Blu-ray movies. The new programmable video-processing engine takes on all of the high-definition H.264 video decoding, freeing the CPU to perform other tasks, while significantly reducing power consumption, heat, and noise. Spectacular picture clarity and vibrant color is achieved with advanced video processing technology. The SPARKLE GeForce 9600 GSO 768MB Graphics Card also supports dual-stream decode accelerations, which lower the average CPU usage to below 8% when using picture in picture feature.
Ahead for next generation data transportation standard, the SPARKLE GeForce 9600 GSO 768MB Graphics Card complies with PCI Express 2.0 standard, it can maximize the usage of 5GT/s bandwidth ( doubled than 2.5GT/s PCI Express 1.1 can provide) provided by PCI Express 2.0, to be full ready for future games and applications which have strict requirements on bandwidth. At the same time, the SPARKLE GeForce 9600 GSO 768MB Graphics Card is also
compatible with with present PCI Express 1.1 motherboards with widest scope of support. "With doubled capacity of video memory, the SPARKLE GeForce 9600 GSO 768MB Graphics Card deliver much more powerful graphics processing for the most demanding DX 10 games.
Now mainstream gamers have a new weapon for handling DX10 games, such as Gears of War, Crysis, Bioshock and so on."Said Precilla Wu, sales manager of SPARKLE Computer Co., Ltd. "The new SPARKLE GeForce 9600 GSO 768MB Graphics Card shows an improved performance-per-watt ratio compared to its predecessor as well as improved compression efficiency. It further extends SPARKLE’s ability to deliver improved 3D performance at an affordable price point, so that mainstream users can have an amazing gaming experience.”
For more information, please visit www.sparkle.com.tw
27 May 2008
NVIDIA to Announce Tegra: Computer-on-a-Chip solution
(hardwarezone) -- Yup, you've read right. NVIDIA is cooking up a new series that's not in their GeForce, nForce, Quadro or Tesla series and it could be a big stir for both AMD and Intel. From what we found out and hints from NVIDIA, the newly trademarked name "Tegra" seems to point towards a SOC (system-on-a-chip) solution. What we know to-date is that NVIDIA and VIA are joining hands to debut a sub US$45 platform, but it's really unclear if this is going to be named Tegra or some other new-fangled development that's been under-wraps.
Fortunately, we needn't wait long as come Computex 2008 on 3rd June 2008, NVIDIA will announce this new development of the Tegra branded family. Will it be a CPU-GPU combo chip or a new platform for the mobile computing space or a new device for handhelds or something completely unfathomed? The curtain will unveil soon enough in less than two weeks. So stay tuned as the Hardware Zone team prepares itself to set a temporary base in Taipei for all the latest developments direct from Computex.
Visit Sapphire Technology at Computex 2008 Taipei
Taipei, Taiwan (hardwarezone) -- This year, the Asia’s biggest exhibition “Computex 2008” is being staged at the new exhibition center—TWTC Nangang Exhibition Hall and TWTC Hall1, TWTC Hall3 and Grand Hyatt Taipei.
Sapphire Technology will also join the world’s leading IT show, and Sapphire will show you the newest series gaming graphics and professional workstation graphics, mainboard, digital photo frame, TV tuner etc….
This is the first global exhibition of Sapphire workstation graphics group after being the AMD authorized exclusive partner in Asia Pacific (except Japan), it is a big step of Sapphire and it also means Sapphire is not only the gaming graphic’s leader but also eligible to involve in this professional high-end workstation graphics market.
In the Computex show, Sapphire will display the most high-end graphics ATi FireGL V8650 which built-in 2GB high speed memory with a huge 56” inch LCD monitor along with double duel-link (Maximum Resolution: 3840X2160). This is the world superior and the debut of 56” quad full high definition LCD monitor collocate with ATI FireGL V8650 solution demo. The unprecedented accuracy must makes you fully enjoy in this truly visual experience. ATI FireGL is completely designed for CAD/CAM industry design, Digital Content Creation, GIS analysis and high- precision medical equipment.
"ATI FireMV multi-view is the best solution for the “one card-quad monitor display”, and it can also process with four different works independently. The special features such as low power consumption and half length have been got lots of internal enterprise’s affirmation. For instance, NASA, UPS, American Airline, Bloomberg, United State Postal Service… are all ATI FireMV long-term customers .
The other hot subject will be the newest display interface ”Display Port”. Sapphire will show you the first workstation graphics which are available to support Display Port?
Want to see more exciting new products!! Visiting our Sapphire Technology Suite in Grand Hyatt Taipei Room 2401
Date: June 3-June 6
Time: 9:00-18:00
Venue: Grand Hyatt Taipei Suite 2401
For more information, please visit Sapphire website
http://workstation.sapphiretech.com/computex2008/computex2008.html
26 May 2008
Thermaltake DuOrb AX VGA Cooler
(ubergizmo) -- Graphic cards have gotten more and more complex in recent years, and these tend to let out a whole lot of heat as well. Thermaltake has a solution to that problem with its DuOrb AX VGA cooler that brings a couple of 80mm fans to fight the heat, boasting a maximum airflow of of 23.8 CFM at 2500 RPM when connected to an aluminum heatsink with a copper base and a couple of 6mm heatpipes. There is no word on pricing, but it ought to arrive in store shelves within two weeks' time. The DuOrb AX VGA cooler is compatible with the following video cards :-
* NVIDIA GeForce 9600 GT
* NVIDIA 8800 GTX/GTS/GT
* NVIDIA 8600 GTS/GT
* NVIDIA 7900 GTX/GT/GS
* NVIDIA 7800 GTX/GT/GS
* NVIDIA 7600 GT/GS
* NVIDIA 6800 GT/GS
* AMD/ATI Radeon HD 3870
* AMD/ATI Radeon HD 3850
* AMD/ATI Radeon X1950 PRO/XTX/XT
* AMD/ATI Radeon X1900 GT/XTX/XT
* AMD/ATI Radeon X1800 GTO/XL/XT
21 May 2008
Supposed pictures of the Nvidia GTX280 appear
(n4g) -- Chinese website itocp.com has scored what is said to be the first pictures of the upcoming Geforce GTX280, although it's hard to see any details as the pictures are tiny. There's a larger picture of the GPU itself and it lists the model number as G200-300-A2.
This is of course an engineering sample as final production cards are almost a month away. If these pictures are of a real card, then they confirm the earlier mechanical drawings that turned up and the card does indeed feature an additional chip towards the front of the card.
It looks very similar to the 8800 GTX, but it uses one 6-pin and one 8-pin power connector, so it's quite clear that this is a different card. The memory layout is also different as the card has eight memory chips per side for a total of 16 compared to 12 for the 8800 GTX. The SLI connectors also appear to be slightly different, as well as the power regulation area.
18 May 2008
Preview: Upcoming graphics chips from ATI, Nvidia
(crave.cnet.com) -- AMD-ATI and Nvidia are preparing for the next graphics chip showdown. And there is already a good deal of information (and rumor) on the two chips due in June.
The names of the two upcoming product families have been widely reported: The ATI line is branded as the Radeon HD 4800, while the Nvidia is dubbed the GeForce GTX 200.
Advanced Micro Devices is expected to launch the HD 4850 (price estimates of graphics boards range between $189 and $219) and then follow with the 4870 (estimates range between $199 and $279). In the fourth quarter, AMD plans to add the dual-chip ATI Radeon 4870 X2.
Nvidia will respond with the high-end GeForce GTX 200 family. Initial products will be the GeForce GTX 260 and GTX 280.
VR-Zone has already gotten its hands on some preliminary performance numbers for the HD 4850 and 4870. German-language site Hardware-Infos has posted a table with specifications of the HD 4850 and 4870.
Tech site tg daily said "that card vendors will start printing their boxes next week, which means that the specifications are final at this time."
Less seems to be known about the Nvidia GTX 260 and 280, though a Turkish site is claiming to have all the specifications.
23 April 2008
NVIDIA GeForce 9800 GX2 - Commemorative Box Edition!
BFG Technologies recently teamed up with NVIDIA to create a limited run of only ten of these boxes (and donated the graphics cards!), featuring the image of Huang taken by Rick Dahms, to honor NVIDIA’s selection by Forbes as “Company of the Year.” Each of these graphics cards has been autographed by both Huang, as well as Scott Herkleman, the President of BFG Technologies. Packaged in this limited-edition box, only 10 of these examples exist in the world.
19 March 2008
Nvidia's GeForce 9800 GX2 Launches
(WP) -- This is the day that all PC gamers dread. The graphics card they just sank a couple hundred dollars into is no longer the graphical king of the hill. The new heir to the throne: Nvidia's GeForce 9800 GX2. And it'll only cost you between $600 and $650.
So what makes this card such a big deal? The 9800 is a powerhouse. In fact, it physically feels like two 8800 GTX's sandwiched into a single card and the specs justify that notion -- it has two 128 Processor Cores (256 advertised on the box), two times the 512MB GDDR3 memory (512MB per GPU). With a 600-MHz core clock, 1500-MHz shader clock and 1000-MHz memory clock, it promises to crush benchmark scores. full story
02 March 2008
New Features for ATI Catalyst Drivers
Hybrid CrossFire with the 8.3 Catalyst release!
The season for new features added to graphics drivers seems to have begun. First was Nvidia with the High-Definition Video feature added to its ForceWare release, and now we learn that ATI also has in mind some changes for its future Catalyst Software. It seems that the Catalyst 8.3 will bring major changes to the driver and software included in the suite, something more than just the usual monthly updates that we are used to. [more...]
25 February 2008
ATI Radeon HD 3870 X2 Video Card Review
(Legit Reviews) -- The ATI Radeon HD 3870 X2 video card (known internally as the R680) was first seen by the world right here on Legit Reviews a few months ago and was again spotted at CES 2008, where the the retail card was seen. Everyone knew that the launch was close, but today is the day that the card is officially announced. The ATI Radeon HD 3870 X2 is a single-board with multiple-GPUs on it, which is why the card is so interesting. ATI took two RV670 cores (Radeon HD 3870) and put them on the same PCB with the same amount of memory with the bridge chip already on the video card that enabled CrossFire automatically. The result is a graphics card that has two RV670 cores with 1GB of GDDR3 memory at a price point of $449, which puts it in competition with the GeForce 8800 GTX and the GeForce 8800 Ultra graphics cards by NVIDIA.
The ATI Radeon HD 3870 X2 graphics card delivers up to twice the Gigaflops per watt of previous generations of high-end AMD GPUs as it is basically two ATI Radeon HD 3870 cores on one card. This means that the ATI Radeon HD 3870 X2 video card should deliver twice as much performance and all the specifications remain the same. It should be noted that the Radeon HD 3870 X2 has enough math processing muscle to break the 1 TeraFLOPS mark!
Just looking at the video card one can tell just how beefy the cooling system is and see that it must consume some power as it has two power connectors located along the top edge of the card. The ATI Radeon HD 3870 X2 graphics card requires a minimum 500W or greater system power supply for single card operation. Since the board power usage is less than 225W only the two 2x3 PCI Express power connectors are requierd in normal operation. The 2x4 PCI Express power connector may be needed for overclocking, so keep this in mind if you plan on overclocking. For running a pair in CrossFireX it is unknown what the requirements will be as ATI has not made this information public since the CrossFireX drivers are still in development.
The core frequency on the R680 is higher than that on an RV670, which is a shocking move. The ATI Radeon HD 3870 has a core clock of 775 MHz while the ATI Radeon HD 3870 X2 has been increased up to 825 MHz. This might have been down as the memory clock was decreased from 2.25GHz GDDR4 to 1.8GHz GDDR3. (note the above slide is dated and that the memory clock frequencies have changed) This slide does point out that the ATI Radeon HD 3870 X2's cooling fan operates at 36dBA, which is just slighting higher than the 34dBA seen on the Radeon HD 3870. The boards peak power usage is also 91W higher than the Radeon HD 3870 at 196W under full load. Having two RV670 cores on one board doubles many things like the frame buffer size and the math processing power, but the power consumption does not double.
Read More......GeForce 9600 GT Hands-On Preview
Nvidia released the GeForce 8 GPU series in late 2006 and reigned over the enthusiast graphics world essentially undisputed until AMD responded with the ATI Radeon HD 3800 series, all of which threaten the GeForce in more price-conscious segments. AMD's assaults on Nvidia's bread-and-butter midrange cards forced the GeForce-makers into action. The company's response to these new threats is the GeForce 9600 GT 512MB. With an MSRP close to $180, the GeForce 9600 GT 512MB slides into the mainstream category with some seriously great performance at an affordable price.
The GeForce 9600 GT 512MB comes with 64 stream processors, double the number found in the GeForce 8600 GTS but not quite as many as in the original GeForce 8800 GTS 320MB. To make up for the shortfall in stream processors, Nvidia gave the GeForce 9600 GT quicker clock speeds, pairing the 650MHz core clock with a speedy 1.62GHz shader clock. The 512MB of GDDR3 memory has an effective speed of 1.8GHz. Nvidia states that the card uses 95W of power and recommends a 400W power supply for the standard GeForce 9600 GT 512MB.
| GeForce 9600 GT | GeForce 8800 GTS 320MB | GeForce 8600 GTS | |
| Price | $169-$189 | $220 | $135 |
| DirectX | DirectX 10, SM4 | DirectX 10, SM4 | DirectX 10, SM4 |
| Core Speed | 650MHz | 500MHz | 675MHz |
| Stream Processors | 64 | 96 | 32 |
| Shader Speed | 1.62GHz | 1.2GHz | 1.45GHz |
| Memory | 512MB | 320MB | 256MB |
| Memory Speed | 900MHz (1.8GHz GDDR3) | 800MHz (1.6GHz GDDR3) | 1GHz (2GHz GDDR3) |
| Memory Interface | 256-bit | 320-bit | 128-bit |
Expect to see several GeForce 9600 GT 512MB variants out there, because video card manufacturers can adjust memory and GPU clock speeds to create premium models with higher prices.
The introduction of the GeForce 9600 GT 512MB does not bring with it a whole new architecture or feature set. However, according to Nvidia, the new GPU benefits from process and architectural improvements that allow it to handily outperform an equivalent older-model GeForce 8 GPU, and it also comes with compatibility for PCI Express 2.0.
The GeForce 9600 GT 512MB comes with the same accoutrements as other GeForce 8 series GPUs. You'll get the unified shader architecture, full DirectX 10 compatibility, as well as hardware accelleration and image-quality improvements for H.264 and VC-1 high-definition video content.
Conclusion
The GeForce 9600 GT 512MB performs spectacularly for a card in the mainstream category. It thoroughly trounced prior-generation parts like the GeForce 8600 GTS and Radeon HD 2600 XT, outperformed the more expensive Radeon HD 3870 in Crysis and Call of Duty 4, and waltzed past the comparably priced Radeon HD 3850 in all of our tests. The GeForce 9600 GT didn't beat the GeForce 8800 GT 512MB card or XFX GeForce 8800 GTS 320MB XXX Edition, which is considerably more expensive and overclocked.
The competition in this price range is fierce, as ATI just announced price cuts for both of their competing GPUs. Choosing a video card got a little harder, as it depends on how the price war plays out. Regardless of card, you win. Lower prices and higher performance doesn't exactly hurt the consumer.