iWARP Performance with PAM-CRASH* Virtual Solution
The Intel® iWARP (Internet Wide Area RDMA Protocol) Ethernet Performance Series of test results seeks to help close the gap between real-world user requirements and the micro-benchmarks promoted by other RDMA (Remote Direct Memory Access) vendors. Each paper in the series demonstrates the real-world performance of NetEffect* 10 Gigabit Ethernet Server Cluster Adapters from Intel on an industry-standard application.
This paper reports on performance testing of ESI Group’s PAM-CRASH* physics-based simulation software. It demonstrates that iWARP is a viable alternative to proprietary fabrics in this test case, providing the opportunity for HPC operators to benefit in terms of cost, complexity, and flexibility.
Overview
DMA enables direct, zero-copy data transfer between RDMA-capable server adapters and application memory, removing the need in Ethernet networks for data to be copied multiple times to operating system (OS) data buffers.
The mechanism is highly efficient and eliminates the associated processor-intensive context switching between kernel space and user space. HPC applications can therefore reduce latency and perform message transfer very rapidly and consistently by directly delivering data from application memory to the network.
Both iWARP and InfniBand* use RDMA and a common API for HPC applications, but iWARP has the advantage of enabling the use of RDMA over the familiar Ethernet fabric. Because iWARP runs over Ethernet TCP/IP, it enables both application and management traffc to operate over a single wire.
Read the full iWARP Performance with PAM-CRASH Virtual Solution White Paper.
1876KB
Lo sentimos, este PDF sólo está disponible en formato para descarga
iWARP Performance with PAM-CRASH* Virtual Solution
The Intel® iWARP (Internet Wide Area RDMA Protocol) Ethernet Performance Series of test results seeks to help close the gap between real-world user requirements and the micro-benchmarks promoted by other RDMA (Remote Direct Memory Access) vendors. Each paper in the series demonstrates the real-world performance of NetEffect* 10 Gigabit Ethernet Server Cluster Adapters from Intel on an industry-standard application.
This paper reports on performance testing of ESI Group’s PAM-CRASH* physics-based simulation software. It demonstrates that iWARP is a viable alternative to proprietary fabrics in this test case, providing the opportunity for HPC operators to benefit in terms of cost, complexity, and flexibility.
Overview
DMA enables direct, zero-copy data transfer between RDMA-capable server adapters and application memory, removing the need in Ethernet networks for data to be copied multiple times to operating system (OS) data buffers.
The mechanism is highly efficient and eliminates the associated processor-intensive context switching between kernel space and user space. HPC applications can therefore reduce latency and perform message transfer very rapidly and consistently by directly delivering data from application memory to the network.
Both iWARP and InfniBand* use RDMA and a common API for HPC applications, but iWARP has the advantage of enabling the use of RDMA over the familiar Ethernet fabric. Because iWARP runs over Ethernet TCP/IP, it enables both application and management traffc to operate over a single wire.
Read the full iWARP Performance with PAM-CRASH Virtual Solution White Paper.
