Nvidia Installs Flagship Graphics Chip onto Tesla Compute Card

Nvidia Corp. on Tuesday introduced its new high-end compute card designed for high-performance scientific computing. The novelty utilizes the company's latest code-named GF110 graphics processing unit, has 512 stream processors and delivers the industry's highest compute performance among specially designed accelerators.

Nvidia Tesla M2090 compute card is based on the code-named GF110 graphics processor clocked at 650/1300MHz that has all the 512 stream processors enabled. The board also has 6GB of GDDR5 memory with ECC clocked at 3.70GHz. Thanks to usage of the more advanced graphics processor for computing, the novelty delivers roughly 665GFLOPS of double precision performance, which is about 30% more horsepower than the predecessor M2070.

 

In the latest version of Amber 11, one of the most widely used applications for simulating behaviors of biomolecules, four Tesla M2090 GPUs coupled with four CPUs delivered record performance of 69 nanoseconds of simulation per day. The fastest AMBER performance recorded on a CPU-only supercomputer is 46 ns/day, according to Nvidia. In addition to AMBER, the Tesla M2090 GPU is well suited to a wide range of GPU-accelerated HPC applications.

Read More Here

Nvidia to release their GeForce GTX 590 dual-GPU in February

GeForce GTX 590 is the confirmed name of NVIDIA's coming graphics card flagship. With dual GF110 GPUs the beafy graphics card will offer 1024 CUDA cores and 3GB memory. Perhaps the most exciting is that the launch is set for February, in the same timeframe as AMD Radeon HD 6990 is expected to arrive.

NVIDIA's Fermi architecture have gone from being a laugh in 2009 to being the father of a very respectable graphics card family that boasts the fastest single-GPU graphics card around - GeForce GTX 580.

The company is on a roll now and is expected to follow up the launch of GeForce GTX 560 Ti with a new flagship. The new graphics card comes with dual graphics circuits and according to reliable sources the name is set to GeForce GTX 590.

Continue Reading...
www.customdesktops.moonfruit.com

Intel's Next-Generation Chip to Support DirectX 11

Intel Corp. plans to integrate a DirectX 11-class graphics processing unit (GPU) into its next-generation code-named Ivy Bridge central processing unit (CPU). The move will allow the accelerated processing unit (APU) to compete head-to-head with Advanced Micro Devices' Fusion-series of chips that already have DirectX 11 support.

"Intel will integrate the technology in next-generation laptop and desktop chips, as use of the technology in applications will spread by then. [...] When we look at the schedule, we did not think it was... the right time. There's not much usage." said Mooly Eden, vice president and general manager of the PC client group at Intel, in an interview with the IDG News Service.

The latest Intel Core i "Sandy Bridge" microprocessors feature outdated DirectX 10.1-class graphics core and does not support GPGPU technology. By contrast, modern GPUs from ATI/AMD or Nvidia Corp. fully support DirectX 11 along with a number of other advantages that improve quality of video games as well as allow to use graphics processor for general-purpose computing.

Since Ivy Bridge will be made using 22nm process technology, Intel will be able to integrate more logic into the chip. For example, it is rumoured that Ivy Bridge will have 24 stream processors instead of 12 on the Sandy Bridge.

But even Ivy Bridge may not remove the gap between Intel and AMD chips in terms of graphics performance. In its next-generation APUs the Sunnyvale, California-based company may further improve performance and add new features to its integrated GPUs.

www.customdesktops.moonfruit.com