News Posts matching "GPUDirect"

Return to Keyword Browsing

Tesla K20 GPU Compute Processor Specifications Released

Specifications of NVIDIA's Tesla K20 GPU compute processor, which was launched way back in May, are finally disclosed. We've known since then that the K20 is based on NVIDIA's large GK110 GPU, a chip never used to power a GeForce graphics card, yet. Apparently, NVIDIA is leaving some room on the silicon that allows it to harvest it better. According to a specifications sheet compiled by, Tesla K20 will feature 13 SMX units, compared to the 15 available on the GK110 silicon.

With 13 streaming multiprocessor (SMX) units, the K20 will be configured with 2,496 CUDA cores (as opposed to 2,880 physically present on the chip). The core will be clocked at 705 MHz, yielding single-precision floating point performance of 3.52 TFLOP/s, and double-precision floating point performance of 1.17 TFLOP/s. The card packs 5 GB of GDDR5 memory, with memory bandwidth of 200 GB/s. Dynamic parallelism, Hyper-Q, GPUDirect with RDMA are part of the new feature-set. The TDP of the GPU is rated at 225W, and understandably, it uses a combination of 6-pin and 8-pin PCI-Express power connectors. Built in the 28 nm process, the GK110 packs a whopping 7.1 billion transistors.


NVIDIA Releases CUDA 5

NVIDIA today made available the NVIDIA CUDA 5 production release, a powerful new version of the world's most pervasive parallel computing platform and programming model for accelerating scientific and engineering applications on GPUs. It can be downloaded for free from the NVIDIA Developer Zone website.

With more than 1.5 million downloads, supporting more than 180 leading engineering, scientific and commercial applications, the CUDA programming model is the most popular way for developers to take advantage of GPU-accelerated computing.

NVIDIA Partners Make Ultra-Low Latency a Reality with NVIDIA GPUDirect for Video

NVIDIA and industry leading I/O board partners such as AJA, Blackmagic Design, Bluefish444, Deltacast, DVS, and Matrox are providing unprecedented real time video production capabilities leveraging NVIDIA GPUDirect for Video. The technology provides application developers and their customers seamless, fast accessibility of the graphics and image processing power of NVIDIA Quadro and Tesla professional graphics processing units (GPUs), with ultra-low latency input and output across a wide range of I/O devices.

NVIDIA GPUDirect for Video technology is the fastest, most deterministic way to get video data in and out of the GPU. Software vendors are now capable of harnessing the graphics and image processing power of GPUs without the latency, often as many as ten frames, previously associated with 3rd party video I/O boards. With this wide range of I/O vendors, customers can choose the best system for meeting their needs.

Return to Keyword Browsing