Cavium™ Announces Beta Support for PathScale EKOPath Compiler and Optimized Math Libraries on ThunderX Processor Family
HPC development environment for the most demanding computationally complex and scientific computing applications now available
Frankfurt, Germany, July 13, 2015 – Cavium, Inc. (NASDAQ: CAVM), a leading provider of semiconductor products that enable intelligent processing for enterprise, data center, cloud, wired and wireless networking, today announced the availability of the Beta release of PathScale’s optimized compiler and library software for the ThunderX ARMv8-A processor family.
ThunderX is Cavium’s 64-bit ARMv8 server processor family for next generation datacenter and cloud applications. With up to 48 high-performance custom cores, single and dual socket capability, high memory bandwidth and capacity, and integrated hardware accelerators, ThunderX enables best-in-class ARMv8 performance per dollar and performance per watt. The ThunderX family includes multiple SKUs that enable servers optimized for compute, storage, network and security workloads in the cloud, and is widely supported by industry-leading OS, hypervisor, software tool and application vendors.
The EKOPath compiler for High Performance Computing (HPC) supports advanced loop optimizations, SIMD vectorization and many-core support. This release of EKOPath includes optimizations specifically for the ThunderX microarchitecture as well as tuning tailored to typical HPC workloads, including computationally complex and data intensive applications as well as OpenMP. This Beta release from PathScale provides essential support for ThunderX to scale performance on both single SOC 48 core systems as well as dual socket configurations of 96 cores and delivers the features and capabilities previously available from Pathscale on other server architectures. Future releases of the PathScale EKOPath compiler will further optimize for Cavium’s memory controller architecture that targets latency and memory sensitive HPC applications.
The optimized PathScale BLAS libraries will help ensure scientists who rely on DGEMM, SGEMM and math library performance achieve the highest possible efficiency for Cavium's ThunderX processors.
“Pathscale’s software will provide essential highly optimizing compilers for C, C++ and Fortran HPC applications on ThunderX,” said Larry Wikelius, Director Thunder Ecosystems and Partner Enablement at Cavium™. “Today’s announcement is the latest example of Cavium’s leadership in the ARM server and HPC ecosystem. This release will dramatically accelerate the development, optimization and delivery of software on ThunderX based systems for the HPC market that are being provided by Cavium partners such as Cray, Lenovo, E4 and ASUS.”
“This Beta release of the PathScale software will enable the HPC market to fully utilize the features that ThunderX delivers in both single and dual socket configurations up to 96 cores,” said Christopher Bergström, Chief Technology Officer, PathScale. “Pathscale’s HPC industry and many-core expertise offers a tremendous combination to Cavium’s leadership in scalable SOC architecture and design.”
The ThunderX family includes multiple workload optimized processors that enable servers and appliances that are optimized for compute, storage, network and secure compute workloads. The ThunderX processor family is fully compliant with ARMv8 architecture specifications as well as ARM’s SBSA and SBBR standards and is widely supported by industry leading OS, Hypervisor and SW tool and application vendors.
ThunderX customer reference platforms will be on display at Cavium booth #1105 at the International Supercomputing Conference from Monday, July 13th to Wednesday, July 15th in Frankfurt, Germany. The PathScale team will be jointly participating with Cavium at ISC.
PathScale has developed industry leading high performance Fortran, C and C++ compiler products for ARMv8, AMD GPU, NVIDIA GPU and provides support to users desiring the highest level of performance from their applications. The PathScale Compiler Suite has the world's most advanced optimization infrastructure and can fully exploit the potentials of many-core architectures The company’s goal is to deliver robust and high performance compilers and GPGPU solutions tailored to clustered and multi-core computing environments.
Cavium is a leading provider of highly integrated semiconductor products that enable intelligent processing in enterprise, data center, cloud and wired and wireless service provider applications. Cavium offers a broad portfolio of integrated, software-compatible processors ranging in performance from 100 Mbps to 100 Gbps that enable secure, intelligent functionality in enterprise, data-center, broadband/consumer and access and service provider equipment. Cavium’s processors are supported by ecosystem partners that provide operating systems, tool support, reference designs and other services. Cavium’s principal office is in San Jose, CA with design team locations in California, Massachusetts, India and China. For more information, please visit: http://www.cavium.com.
Sr. Marketing Communications Manager,
Cavium, 2315 N. First Street,
San Jose CA 95131,
Telephone: +1 408-943-7417