Highest memory bandwidth, Lowest cost per core
AMD Opteron 6100 CPUs in 4P configurations:

• 4-Socket servers for 2-Socket prices
• Provide large SMP memory capability and easy management
• Exceptional memory bandwidth at over 100GB/sec
• Available with the latest ConnectX-2 QDR IB & 10G ethernet interconnects.
• Offered in 1U-4U footprints starting below $5,000.

Faster clock speeds at lower cost
Intel Xeon 5600 CPUs in 2P configurations:

• Fastest performance per core - maximize computational throughput of your nodes
• Larger 12MB L3 cache and more banks of DIMMS running at 1333MHz
• Similar price points to Xeon 5500-series CPUs
• Available in 1U-4U servers, 1U Twin, or 2U Twin² with the latest ConnectX-2 QDR IB & 10G ethernet interconnects.

Balanced, Cost-Effective Clusters
AMD Opteron 6100 CPUs in 2P configurations:

• Maximize the core count of your cluster with 8-core and 12-core processors
• Available in 1U-4U servers, 1U Twin, or 2U Twin² with the latest ConnectX-2 QDR IB & 10G ethernet.

Contact us for a quote on these configurations; all are available with GPU compute processors.

On Monday, June 21st at 11am Pacific/2pm Eastern Microway and NVIDIA will host a webinar. Stay tuned for an e-mail next week with more details and a link for registration.

After July, NVIDIA will no longer offer a 25% discount to academic institutions.

Microway will continue to quote and sell NVIDIA Tesla products with the 25% university discount through the month of July. Please work with your purchasing department to have all orders in to Microway by July 29th. Late orders cannot be accepted, as NVIDIA has placed a strict cutoff date for these discounts. Fax and email orders will ensure prompt processing.

Academic customers may pre-order not-yet-released Tesla products with the 25% discount - please contact your Microway sales representative to make the arrangements.

Tesla S2050 GPU Compute Servers

• 4 GPUs in a 1U footprint total of 2 TFLOPS double- and 4.13 TFLOPS single-precision floating point performance
• Total of 12GB GDDR5 memory (4x 3GB GPUs) with optional ECC
• Order now - a 6-week lead time is expected due to demand

Tesla C2070 GPUs - Pre-order Now

• 448 CUDA cores provide 515 GFLOPS double- and 1.03 TFLOPS single-precision floating point performance
• 6GB GDDR5 memory with optional ECC

We've just returned from a successful exhibit with NVIDIA at HPCS '10 in Toronto. Microway enjoyed meeting present and future Canadian customers and we particularly valued the high quality of discussions and inquiries.

Successfully accelerating your applications using GPUs requires the right hardware platform. Learn more in our GPGPU Architecture Camparison.

Scores of customers anecdotally share their cluster successes on Microway technology with us. If you would like to share your story or projects that leverage Microway hardware either privately or with others who read this newsletter, please email sales@microway.com