NOTICE OF INTENT TO PROCURE. LINUX CPU/GPU CLUSTER UPGRADES FOR DEPT OF NAVY, THE UNITED STATES RADAR AND ANTENNA SYSTEMS ENGINEERING DIVISION, RADAR AND COMPUTATIONAL ELECTROMAGNETIC MODELING (RACEMM) LABORATORY, NAVAL AIR STATION PATUXENT RIVER, MD

The Government intends to sole source this requirement to Aspen Systems, Inc, 3900 Youngfield St Wheat Ridge, CO 80033-3865 USA.  The purpose of this notice is to perform market research to see if any other vendor’s exist that are capable of providing the equipment as specified below:

CLASSIFICATION:

Product Service Code:  7021     ADP Central Processing Unit (CPU, Computer) Digital

NAICS Code:   334111    Electronic Computer Manufacturing

Place of Performance: Naval Air Station Patuxent River, Maryland  20670

DESCRIPTION:

AIR-4.5.5 Radar and Antenna Systems Engineering Division, Radar and Computational Electromagnetics Modeling (RACEMM) Laboratory possesses a state-of-the-art facility providing detailed Modeling and Simulation (M&S) services for antennas and Radio Frequency (RF) Systems onboard Naval platforms. RACEMM performs specialized computational electromagnetics (CEM) modeling which require state-of-the-art computational facilities/platforms to accomplish such tasks.

In FY2019, the RACEMM Lab, through a competitive RAC, procured a 7-node Linux cluster from Aspen Systems.  With an additional $250k available funding in FY2020, through a sole-source RAC, we require Aspen Systems to add compute nodes to fill the existing equipment rack.  This sole-source requirement is due to the proprietary Aspen software currently installed on the existing system that the controls communication between the head and compute nodes.  No other vendor can provide guaranteed compatible software that would replace this Aspen proprietary software ensuring seamless cluster integration and operation.

GENERAL REQUIREMENTS:

The vendor shall provide the following systems and services:

• Additional compute nodes with at least the following configuration:
  • Dual Xeon Cascade Lake 6238 processors
  • Minimum 512 GB RAM
  • Minimum 3 NVIDIA Tesla Volta V100 GPU cards, with minimum RAM per GPU card = 32 GB each)
  • ≥ 8 TB HDD space per node
• If required, additional UPS/MPU hardware to power these additional compute nodes.
• Five-year on-site service maintenance agreement.
• Delivery and set-up of Cluster on-site.

Point Of Contact:  Andrew Johnson, Contract Specialist, [email protected]