Despite earlier reports that Intel visited board vendors in China to discuss Larrabee, the firm’s PR claims that no such meetings took place.

Heading off inquiries from the press, Intel’s Nick Knupffer said via Twitter that Intel CEO Paul Otellini “did not talk LRB to AIB’s in China.”

Knupffer’s statements come in response to a Digitimes report that Otellini made rounds with Chinese board makers to promote the company’s upcoming Larrabee GPU. Digitimes further alleged that Intel was preparing to offer preferential pricing and bundling to allay fears that the GPU would not effectively compete against incumbent GPU firms AMD and NVIDIA.

Larrabee is Intel’s first foray into discrete, or standalone, GPUs in about ten years. The new design is said to use an array of x86-compatible execution engines which are harnessed to crunch data in parallel. Intel hopes to leverage this architecture to compete directly with AMD and NVIDIA in gaming, and later in other markets.

Intel maintains that Larrabee-based products are scheduled for early 2010, which means the February/March launch window. If that is the case, we find it dubious that Intel isn’t talking with board partners to produce rebranded parts.

Recent reports from unofficial sources contradict official statements from NVIDIA which say that the company’s upcoming Fermi/G300 GPU will be available by the end of the calendar year.

Unnamed board partners cited at DriverHeaven and [H]ard|OCP claim that NVIDIA is working to have “some” Fermi sample boards running at the January, 2010 Consumer Electronics Show in Las Vegas. The implication here is that NVIDIA may very well miss that window to demonstrate working silicon to the public.

"0935A1" - WW35/09 A1 silicon (Image credit: PC Watch)

Of greater concern is the overall state of the Fermi project. If the company is hoping, rather than guaranteeing, to show the first public demo of Fermi in the second week of January, then the G300 may be further behind than anyone had imagined.

After the project slipped its original September/October launch, many believed that G300 would launch in the November/December time frame. That date was based on a July tape-out of the company’s A1 (first) G300 revision.

A tape-out is the finalization of a PCB/IC design; it is at this point that the schematics for the product are sent for manufacture. It typically takes about five months to translate the specs into a physical product. This is why a year-end G300 launch was expected from a July tape-out.

However, a tape-out is not always the last stage of circuit design. The design could fail viability checks at the foundry, or the production board may fail functionality checks in its shakedown cruise. Should either of these scenarios unfold, the circuit designer must commit itself to a new revision–a respin–which takes about 4-6 weeks from debugging to the very first test boards. Should that revision get a green light, you can add another 6-8 weeks for production boards.

(more…)

Over the weekend EVGA announced the release of the EVGA GeForce GTX 275 CO-OP graphics card which places a GeForce GTX 275 and GTS 250 on the same board, with the latter serving as a dedicated PPU.

The EVGA GeForce GTX 275 CO-OP.

Technologically, the card is an interesting achievement in that “first and only of its type” sort of way.

But when it comes to value, the card takes a big tumble; there aren’t very many AAA titles with a notable NVIDIA PhysX implementation. We also have to consider the GTX 200’s march towards EOL and the introduction of the Radeon HD 5000 series, both which further erode this product’s value.

The GPU’s price is also of concern. With an MSRP of $349, the card is similarly priced to the GTX 285, an adapter which definitely has a leg up in the performance department (to say nothing of the 5850).

Maybe somebody really loves Arkham Asylum this much.

NVIDIA has taken the wraps off of the newest driver revision for GeForce 100, 200, 6000, 7000, 8000, 9000 GPUs. Weighing in at roughly 100MB, this beta Windows XP/Vista/7 driver delivers the following changes:

• Adds support for OpenCL 1.0 (Open Computing Language) for all GeForce 8-series and later GPUs.
• Adds support for CUDA Toolkit 3.0 features and performance enhancements. See CUDA Zone for more details.
• Adds SLI and multi-GPU support for Borderlands, Cloudy with a Chance of Meatballs, G-Force, FIFA Soccer 2010, League of Legends: Clash of Fates, NHL 2009, Order of War, Pro Evolution Soccer 2010, Race On, Star Trek: D-A-C. Improved SLI and multi-GPU support for Champions: Online and Dragon Age: Origins.
• Includes over 200 bug fixes. Refer to the release notes (PDF) for information about the key bug fixes in this release.

Download

Windows Vista & 7 x86-32: Here
Windows Vista & 7 x86-64: Here
Windows XP x86-32: Here
Windows XP x86-64: Here

Digitimes is reporting that the Taiwan Semiconductor Manufacturing Corporation (TSMC) is once again experiencing yield issues on the 40nm process node.

TSMC said it has seen yield rates for its 40nm node drop to 40% due chamber matching issues. Main customers for TSMC’s 40nm processes are GPU vendors AMD, which recently launched its new 40nm-based Radeon 5870 series, and Nvidia, who is scheduled to launch its 40nm-based GT300 series of chips in December, according to previous reports. TSMC’s recent issues may impact the shipping schedules of the GPU vendors, market watchers commented.

In addition to shipping schedules, the standing MSRPs of 40nm GPUs may also be compromised. Indeed, rumors have already begun to churn that the Radeon HD 5850’s price will soon increase as demand continues to outstrip supply.

Fresh out of the oven, ATI has just served up a brand new Catalyst driver for Radeon HD 3000, 4000 and 5000-series GPUs. Weighing in at about 100MB, the WHQL-certified Windows XP/Vista/7 driver delivers the following changes:

• Full GPU support for the Radeon HD 5000-series GPUs.
• Enables Super Sampling Anti-Aliasing (SSAA) in the Radeon HD 5000-series cards.
• Adds support for Ubuntu 9.10.
• Several fixes for CrossFire, HDMI and Avivo configurations.

Downloads

Windows 7 x86-64: Here
Windows 7 x86-32: Here
Windows Vista x86-64: Here
Windows Vista x86-32: Here
Windows XP x86-64: Here
Windows XP x86-32: Here

Hollywood is intense. Today, you can go see a film like District 9 or Transformers 2 and be awed by the computer generated visual effects–imagery that mimics the nuances of reality to an unbelievable degree–and not even realize the minute details going in to the production of a shot or, more importantly, the astounding hardware used to create those effects. Animation and VFX studios use impressive machines to bring these films to life. These artists are not relying on Radeon and GeForce GPUs to render their work. No, they pull in the big guns–the workstation boards–to do the heavy lifting.

The workstation GPU market is vastly different than the desktop GPU market, despite radically similar hardware configurations. ATI and NVIDIA both offer workstation solutions, named the FirePro and Quadro lines, respectively. The drivers set these units apart from their desktop gaming-centric cousins. These graphics cards are fine-tuned and optimized to work exceedingly well in 3D and graphics packages, such as Autodesk Maya, Softimage, and 3D Studio Max. They are built with the professional artist in mind, not the high-end gamer and, as a result, the target audience of a workstation board is radically different.

I/O ports on the V8750.

Studios today deal with intense graphics production pipelines. Any and all rendering time is valuable, and costs money. These artists don’t need machines that can render a low polygon scene with pixel shader effects at 60+ frames per second. What they need are machines with hardware that can supplement viewport performance in applications, calculate incredibly complex object physics, and render out an image with thousands of elements, anti-aliasing, shader effects, applied simulations, ray tracing, and whatever else the shot calls for, and it needs to do all of this really, really fast.

(more…)

NVIDIA has released an SLI profile update which adds SLI support for several new titles, while enhancing performance in Champions Online and Dragon-Age: Origins.

Titles with fresh SLI support include:

• Borderlands
• Cloudy with a Chance of Meatballs
• G-Force
• FIFA Soccer 2010
• League of Legends: Clash of Fates
• NHL 2009
• Order of War
• Pro Evolution Soccer 2010
• Race On
• Star Trek: D-A-C

The update weighs approximately 300kb and should be combined with ForceWare version 191.07.

ATI has just recently published a new beta for their Stream SDK, a package designed to assist developers in creating GPGPU apps for ATI graphics adapters. The new package primarily adds OpenCL 1.0 support and, while all of that is well and good, it has also tipped the firm’s hand by revealing evidence of the Radeon HD 5900 series.

A quick peek at the new INF file for the SDK-compatible GPU driver lists the following entries:

“ATI Radeon HD 5900 Series” = ati2mtag_Evergreen, PCI\VEN_1002&DEV_689C
“EG CEDAR” = ati2mtag_Evergreen, PCI\VEN_1002&DEV_68E8
“EG CYPRESS” = ati2mtag_EvergreenGL, PCI\VEN_1002&DEV_6888
“EG JUNIPER” = ati2mtag_EvergreenGL, PCI\VEN_1002&DEV_68A8
“EG REDWOOD    ” = ati2mtag_Evergreen, PCI\VEN_1002&DEV_68DE

What the 5900 constitutes is anyone’s guess. If history is any indicator, it should be a juiced-up version of the 5870 with higher frequencies as a result of improved yields.

Rounding out the list, we see Cypress (Radeon HD 5870/5850), Juniper (Radeon HD 5770/5750), and Cedar and Redwood which have not yet shown in the market. Those cards are believed to be low-end adapters suitable for the OEM channel.

Last Thursday we caught a glimpse of the Radeon HD 5770, but a few 3DMark runs didn’t really answer any questions. Wonder no more, however, as today is the day when the Radeon 5770 hits the streets.

The first thing that can be noticed is that the 5770’s specifications do not stray far from that of the 4870’s. When translating the paper to testing, the spread still holds true: The 5770 and 4870 trade blows, with the 4870 pulling a touch ahead. The 5770 lags behind when memory bandwidth is at a premium, while it pulls ahead when fill rate is at a premium.

(more…)

Earlier in the month we spied the Radeon HD 5750 at a DX11 event in China, but we were left to speculate over specs and speed. Guesstimating be damned, though, as the 5750 is now on the shelves of etailers everywhere. Let’s take a look at how it stacks up.

The first thing that can be noticed is that the 5750’s specifications fall short of the Radeon HD 4870. When translating the paper to testing, the spread still holds true: The 5750 frequently falls short of the 4870, sometimes by over 20 percent.

From the perspective of competition, the GeForce GTs 250 is the 5750’s main opponent. There are no solid victories here, though. The 5750 and the GTs 250 trade blows every time you change the title or the resolution.

(more…)

Ol’ Fuad Abazovic has been doing some digging and reports that ATI has already put a name to its next generation of GPUs.

According to the report, the next-gen family is collectively known as Hecatonchires, or the fifty-headed, hundred-handed beasts of Grecian myth that managed to best the Titans. The desktop lineup appropriately contains three cards named Briareos, Gyes and Kottos, or one for each of the Hecatonchires proper.

Rumor also has it that the card will be a completely new architecture, rather than a much-improved riff on prior silicon as the 5870 was to the 48xx architecture. A headlining feature is said to be Multiple Instruction/Multiple Data (MIMD) shaders, which should dramatically boost GPU parallelism beyond what is offered by today’s SIMD cores.