After treating the enthusiast community to the Republic of Gamers (ROG) ARES Dual HD 5870 graphics accelerator, ASUS isn’t wasting any time is designing its successor, referred to (for now) as “MARS II”. This graphics accelerator uses two NVIDIA GeForce GTX 480 (GF100) GPUs on one board, that’s right, the first dual-GPU accelerator based on GF100, which is dreaded for its thermal and electrical characteristics so much, that NVIDIA is content with having the second-fastest graphics card in the market (GTX 480), with no immediate plans of working on a dual-GPU accelerator.

ASUS’ ambitious attempt is in the design stage deep inside its R&D, where the design is in an evaluation state. The R&D gave us some exclusive pictures of the MARS II PCB to treat you with. To begin with, the card’s basic design is consistent with almost every other dual-GPU NVIDIA card in recent past. There are two independent GPU systems, each with its own VRM and memory, which are interconnected by an internal SLI, and connected to the system bus by an nForce 200 bridge chip. On this card, two GF100 GPUs with the same configuration as GeForce GTX 480 (GF100-375-A3) are used, each having 480 CUDA cores, and connecting to 1536 MB of GDDR5 memory across a 384-bit wide memory interface.

Germany-based Aqua Computer is ready with its water-block for the NVIDIA GeForce GTX 480 graphics accelerator, the AquagraFX GTX480. The block is full-coverage and monolithic in design, and cools all vital heat-producing parts of the card. Its main unit is cut from a block of copper. The block weighs 850 g (1.87 lbs). It has a stainless steel top with laser-etched lettering on it. Aqua Computer used distance pieces for all threads. They are pre-assembled so users won’t have to fumble around with them. All surfaces which have contact to the GPU/RAM/VRM are high gloss polished (mirror-finish). The block also comes with a small backplate that cools voltage regulators on the reverse-side of the PCB. The company will start shipping these from next week, at a price of 89.90 EUR or around US $101 excl. VAT.

Many would be familiar with these pictures of a generic graphics card based on NVIDIA’s GF100 GPU which was spotted at this year’s CES. A company slide sourced by DonanimHaber reveals the final iteration of NVIDIA’s reference design GeForce GTX 480 graphics accelerator, and what it looks like from the outside. A set of slightly more recent pictures showed its cooling assembly from inside. The protruding heat pipes intrigued us as they were inconsistent with the cooling assembly on the card NVIDIA showed off at CES, which we then believed to be the top-end GTX 480 part. The company slide confirms what the cooling assembly looks like when it’s all put together.

The cooler is highly ventilated, with vents all over the cooler’s shroud. There are vents on the top, on the sides, apart from the usual obverse fan air intake. To increase its intake, the PCB is further cut to help draw air from the reverse-side of the PCB. The cooler’s four large (we reckon 8 mm thick) heat pipes protrude about a centimeter out of the card’s periphery, increasing its height by that much. The cooler itself respects the 2-slot thickness limit which is most conventional. A table in the slide also confirms some details we already know: the card has 1536 MB of GDDR5 memory across a 384-bit wide interface. It has a TDP of under 300W, which a recent report reveals to be a hairbreadth under 300W, at 296W. Power is drawn in from an 8-pin and a 6-pin PCI-E power connector. The card is 10.5 inches long, the same length as its reference-design GeForce GTX 280. The card supports 3-way SLI. It will be unveiled on the 26th of March.

Source: DonanimHaber

Czech technology website PCTuning confirmed a few details about NVIDIA’s upcoming performance graphics accelerator, the GeForce GTX 470. It was found out that a beta driver by NVIDIA, GeForce 196.78 supports GeForce 400 series accelerators, and was able run a qualification sample of GeForce GTX 470. The card was using A3 revision GF100 silicon. The driver’s System Information dialog revealed that the card indeed has 448 CUDA cores (SIMD units). Further, it has 1280 MB of memory, and a 320-bit wide memory interface. NVIDIA also changed the way it represents memory clock speeds. Since it is using GDDR5 memory, while the memory has an actual clock speed of 1000 MHz, the data rate (DDR speed) is represented first, as 2000 MHz, and “effective speed” next, which is 4000 MHz.

Given these speeds, at 1000 MHz GDDR5, the GPU has a memory bandwidth of 160 GB/s. Without compromise on looks and quality, NVIDIA kept the cooler design basic. It has a matte finish. Display outputs include two DVI-D, and one mini HDMI. It supports NVIDIA 3D Vision Surround (a technology competitive to ATI Eyefinity, to span a display head across multiple physical displays), just that NVIDIA requires at least two accelerators in SLI to use it. NVIDIA’s GeForce 400 series graphics accelerators will be launched on the 26th of this month.

Source: PCTuning

NVIDIA’s recently released GeForce 196.75 WHQL drivers are causing problems with some users reporting “intermittent low FPS”, in the words of Blizzard’s technical staff. It notes that the drivers have a bug in its fan-speed control that causes graphics cards to overheat when running applications such as Warcraft 3, World of Warcraft and StarCraft 2 Beta. Not only does bad fan control result in bad application performance, but can also damage the graphics accelerator. Blizzard on its part advised users to revert to older versions of the GeForce driver. Version 196.21 (older WHQL-signed driver) can be downloaded from here.

Update (3/5): NVIDIA has withdrawn the 196.75 WHQL drivers from its website. The company is working to resolve this issue.

Source: inc gamers

It looks like Sapphire and ASUS will not be the only AMD board partners out with a custom design Radeon HD 5970 graphics accelerator that has higher clock speeds and twice the amount of memory. XFX seems just about ready with what it calls the XFX Radeon HD 5970 Black Edition 4GB. This limited edition graphics card borrows heavily from AMD’s reference design, but adds some changes:

• The leaf-blower is centrally located, with a GPU system on its either sides.
• GPUs use higher clock speeds of 850 MHz (core), 1200 MHz (memory).
• Twice the amount of memory, 2 GB per GPU, 4 GB total.
• ATI Eyefinity6 display output configuration: 6 mini-DisplayPort connectors to support six physical displays (dongles may be included for DVI users).
• Power is drawn from two 8-pin PCI-E power connectors.

XFX may manufacture upto 1500 pieces in all, and could price each piece around US $1000, with sales starting later this month.

Source: Tweakers.net

Here are some of the first pictures of a “living, breathing” ASUS ROG Ares graphics accelerator. This ASUS’ latest creation packs two AMD Cypress GPUs running at 850 MHz (core), 1200 MHz (memory), with twice the amount of memory (2 GB per GPU, 4 GB total). It also packs an enthusiast-grade voltage circuitry that gives it an amount of overclocking headroom, as well as a complex cooling assembly.

At its CeBIT exhibit, ASUS also showed off the card’s 3DMark Vantage score in the Extreme preset. The card secured a score of X14416, compared to a typical score of the reference Radeon HD 5970 to be around the X11000 mark. The bench was driven by an Intel Core i7 965 XE processor, ASUS Rampage II Extreme motherboard, and 6 GB of triple-channel memory. The ROG Ares should be out in a few weeks time. Its price and availability remain a mystery.

Source: PC Games Hardware

A PCB shot of the upcoming ASUS ROG Ares dual GPU graphics accelerator made it to the media this day. The picture reveals what could be a very complex single-PCB dual-GPU board, with perhaps the strongest VRM to drive the CrossFire-on-board setup. The ROG Ares uses two AMD Cypress GPUs that run at high clock speeds, with even more overclocking potential on offer. The picture reveals that ASUS has made extensive use of digital PWM circuitry, giving each GPU a 4-phase vGPU, 2-phase vMem, and uncore phases. Each zone has its own voltage controller. Power is drawn in from three inputs: two 8-pin and a 6-pin, though the tracks show that the PCB is capable of using three 8-pin inputs. At source, the inputs are fused as a surge-protective measure.

Each GPU is wired to 16 GDDR5 memory chips, 8 on each side of the PCB. The PCB itself is roughly an inch taller than full-height addon-cards. Display connectivity includes one each of DVI-D, DisplayPort, and HDMI connectors. The lone CrossFire finger provides CrossFireX support with another Ares card – or probably other Radeon HD 5800 series products. ASUS in a statement says that all heat-producing components other than the GPU – VRM chips and memory – will be cooled by a copper heatspreader that covers almost all such components. Each VRM chip gets its own copper heatsink. These parts will be anodized in red for the black+red livery characteristic to the ROG series. Earlier, a CAD drawing of the cooling assembly made news.

Source: NordicHardware

About a month ago, it surfaces that ASUS was working on a limited-edition extreme high-end graphics accelerator that uses two Radeon HD 5870 GPUs, in essence an overclocked custom-design Radeon HD 5970, called the Republic of Gamers (ROG) Ares. The Ares builds on the legacy of the ROG Mars. It uses two AMD Cypress GPUs with 1600 stream processors, each, core and memory clock speeds on par with that of the Radeon HD 5870 (850 MHz / 1200 MHz), and double the amount of memory (2 GB per GPU, 4 GB on the card).

A CAD drawing of the Ares surfaced on Plaza.fi, which shows a single-PCB accelerator. The cooling design borrows a little from that of NVIDIA’s second edition GeForce GTX 295, in having a centrally located blower that drives air onto copper GPU blocks on its either sides. The cooler assembly, however, seems much larger at 2.5 slots’ thickness. ASUS also claims that the fan will be quieter on load than AMD’s reference HD 5970 leaf-blower. A table given out lists its important specifications, which shows it to have the same clock speeds as the single-GPU Radeon HD 5870 (850/1200 MHz), versus those of the HD 5970 (725/1000 MHz), twice the amount of GDDR5 memory, and results of an internally conducted 3DMark Vantage benchmark which shows a 28.2% increment over the HD 5970 on the Ares. The card is powered by three PCI-Express power connectors (8-pin + 8-pin + 6-pin), and may have significantly higher power draw. It has also been designed for record-setting scores in graphics benchmark competitions. Being a limited edition product, we expect productions in the tens of hundreds only. If the price of ROG Mars is anything to go by, this one will be an expensive product, too.

Source: Plaza.fi

ASUS has developed its own take on AMD’s mid-range Radeon HD 5750 graphics accelerator, with the EAH5750 Formula 1 GB. Being a non-reference design product, it makes use of both cooler and PCB indigenously developed by the company. The black PCB seems slightly shorter than the reference PCB, and has its power connector located on the top, rather than on the end. The Radeon HD 5750 GPU features DirectX 11 compliance, 720 stream processors, and a 128-bit GDDR5 memory interface to connect to the 1 GB of memory on this card. It sticks to AMD’s reference clock speeds of 700/1050 MHz (core/memory).

The other, more important selling point of this card is its custom designed GPU cooler. Apparently the design involves a GPU contact block from which metal fins project radially, on which a PWM-controlled fan circulates air. Characteristic to ASUS’ Formula series graphics cards, the cooler shroud coarsely resembles a Formula One racing car. The connectors are all rounded off onto one metal bracket, although the card still needs two expansion slots. Connectors include one each of DVI-D, D-Sub, and HDMI. The price is expected to be under € 120.

Source: TechConnect Magazine