You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!

Achronix Chooses Rambus GDDR6 PHY IP for Next-Generation FPGA

AmandaK

Administrator
SUNNYVALE, Calif – June 4, 2019 Rambus Inc. (NASDAQ: RMBS) today announced that Achronix, a leader in FPGA-based hardware data acceleration devices and high-performance eFGPA IP, has selected the Rambus GDDR6 PHY for its next-generation Speedster7t FPGA family. Leveraging the top-end data rates delivered by the Rambus GDDR6 memory interface, the Speedster7t family is optimized for artificial intelligence (AI), machine learning (ML) and high-bandwidth data acceleration applications and workloads.

Designed for performance and power efficiency, the Rambus GDDR6 PHY is beneficial for advanced driver-assistance systems (ADAS), AI, ML, graphics and networking applications. Rambus’ GDDR6 PHY enables the communication to and from high-speed, high-bandwidth GDDR6 SDRAM memory, which is a high-performance memory solution that can be used in a variety of applications that require large amounts of data computation.

“GDDR6 is emerging as a mainstream memory solution for high-performance AI/ML and networking applications, and, in utilizing Rambus’ GDDR6 PHY IP, Achronix is an early leader in adopting the technology to differentiate its next generation of products,” said Hemant Dhulla, vice president and general manager of IP Cores at Rambus. “We are excited to be working with Achronix to expand its portfolio of high-performance FPGA solutions, including its latest Speedster 7t family.”

“New use cases for GDDR6 such as AI/ML require extremely fast data transfer between memory and compute,” said Steve Mensor, vice president of marketing, Achronix. “Rambus’ GDDR6 PHY IP will enable Achronix’ Speedster7t FPGA family to support these high-performance data acceleration applications at a low cost with low latency, allowing it to process complex data loads quickly and efficiently.”
 
Top