Dell’s HPC solution for up to 16 GPUs
November 3rd, 2010 | knowledge | Comments Off
Dell’s HPC solution for up to 16 GPUs
The PCIe Expansion Chassis PowerEdge C410x holds up to 16 Tesla modules from Nvidia. The system has no CPUs or memory chips. It is designed for connection to a Dell PowerEdge C6100.
Dell has presented the PCIe Expansion Chassis PowerEdge C410x for the high-performance computing market (HPC). It was especially for complex compute-intensive demands of science and technology develops.
The PCIe expansion chassis provides 16 slots for GPUs. It is available with Nvidia Tesla GPUs M1060 or M2050. The maximum usable 16 Tesla M2050 modules is a computing power of up to 16.48 Teraflops available.
The system has no CPUs or memory chips and is specially optimized for recording and efficient cooling of GPUs or other PCIe products. This chassis offers an extremely high density with low power consumption and reduced costs.
The expansion unit is designed for connection to a Dell PowerEdge C6100, which has four server nodes, each with two sockets. But it can also be connected to a different server with the appropriate host interface.
Dell HPC is that a small but rapidly growing market. Particularly in relation to cloud computing and solutions for the scientific research in Cambridge about Dell and Intel in J眉lich Exaflops of so-called clusters. Floating Point Operations Per flops are Second, floating-point calculations per second – a measure of the performance of supercomputers. The term Exaflops describes the processing of a trillion – a one with eighteen zeros – instructions per second.
The PCIe Expansion Chassis PowerEdge C410x holds up to 16 modules from Nvidia Tesla (Photo: Dell).
Via: Dell’s HPC solution for up to 16 GPUs
Other Electronics News:
⇒⇒⇒ dell inspiron 6000 battery
