Tesla M2090 Now Available For Purchase in Rackmount Servers
NumberSmasher 1U with 2 TFLOPS
NVIDIA has released the new Tesla M2090 GPUs designed for rackmount servers. Features include:
- 512 CUDA cores with a passive heat sink
- 6GB GDDR5 memory with greater memory bandwidth: 177GB/sec
- Up to 1.33 TFLOPS single and up to 665 GFLOPS double precision floating point throughput
- Compatibility with Microway's NumberSmasher and Navion 1U GPU Servers
- Full specifications and datasheet are available here
AccelerEyes LIBJACKET Now Available
LIBJACKET provides a high-level coding interface for Tesla GPUs. It allows for rapid development and seamless scaling through the utilization of LIBJACKET's pre-coded GPU data types.
- Support for single and double precision values, complex numbers, and booleans
- GPU accelerated functions for manipulating vectors, matrices, and N-dimensional arrays
- Useful for GPU accelerated arithmetic, linear algebra, statistics, imaging, signal processing, and related algorithms
- No low level GPU programming experience required
- Full specifications and a product tour are available here
Microway Sponsoring the Life Sciences Technology Insights Summit
Microway is a Silver sponsor for this premier Life Sciences event to be held at the Harvard Club in Boston next week. For more information, please click here
Win a Radeon 6870 at the HPC & GPU Supercomputing Group of Seattle Meetup:
June 15, 7:30-9:30pm
The HPC & GPU Supercomputing Group of Seattle June meeting will be held concurrently with the AMD Fusion Developer Summit. Microway will sponsor a door prize of an AMD Radeon 6870 GPU
at this valuable meeting. At the meetup Michael Fried, Microway's GP-GPU Business Manager, will present a talk entitled Benefits and Pitfalls of OpenCL.
We encourage GPU developers and HPC users in the Seattle area to attend: the Radeon 6870 is a great GPU for OpenCL development.
For customers in the Boston area, Microway is also a sponsor of the HPC + GPU Meetup group of Boston. They meet monthly on Thursday evenings in Cambridge.
Still Time to Sign Up: June 13-16, Bellevue, WA
More than 90 tech sessions, 5 keynotes led by leaders from across the industry, including Microsoft, ARM, Corel and AMD, hands-on labs, Pre-Summit Tutorials and headline-making networking events.
Here’s a sampling of the HPC sessions that enable you to achieve more:
- AMD’s x86 Open64 Compiler
- Migration of Legacy Applications to Heterogeneous Architectures using HMPP
- GMAC: Simplified CPU-GPU Logical Data Sharing in a Heterogeneous Computing System
- Heterogeneous Computing for Finance Using Mathematica and OpenCL™
- Scaling DGEMM to Multiple Cayman GPUs With Interlagos Many-core CPUs Focusing on HPL
- Using OpenCL for the Design of Molecular Dynamics Simulations Targeting Hybrid CPU/GPU Computing Architectures
- Leveraging Multicore Systems for Hadoop and HPC Workloads
- GPU Accelerated Databases