Tesla C2050
A Supermicro workstation with 2x 2 x Intel Xeon processor E5-6202 (Quad-Core/2.4GHz/12M L3 Cache) have been added with our computational facility for massive simulations. This server is a our first GPU based simulation server in our High Performance Computing(HPC) Lab. Sever is named as Tesla Server C2050 which has 64GB of DDR3 Memory and 6TB of disk space managed by SAS Raid Controller. We just moved from from cluster computing era to GPU computing from this server and performing simulations on Amber and Gromacs software. The special feature of the server is dual Tesla C2050 GPU card is installed with latest Cuda library. From the benchmark it is found that the server can complete 100ns simulation on gromacs in 10 hrs with parallel GPU acceleration. Here is the specification of the server.
| CPU Spefications | Details |
|---|---|
| CPU(s): | 2 x Intel Xeon processor E5-6202 (Quad-Core/2.4GHz/12M L3 Cache) |
| Memory (RAM) | 64 GB DDR3 |
| Storage space (HDDs) | 6 x 1000GB, 7200 RPM, SATA |
| RAID Card: | SAS RAID controller with 128MB cache supports Hardware RAID 0, 1, 10, 5 & 6 |
| Optical Disk Drives | DVDRW drive |
| Network Interface Card | 2 x Gigabit Ethernet on-board using Intel® i210 |
| PCI Express Slots | 2 x PCI-Express x16, 1 x PCI-Express x8, 2 x PCI-Express x4 |
| Chassis | Mid-Tower |
| Power Suply | 900W Single Power Supply (80PLUS Gold) |
| Display Unit | 23.8" TFT Color Monitor |
| Input and Output Units | 104 keys keyboard, Optical Scroll Mouse |
| GPU Card Specification | Tesla C2050 |
|---|---|
| CUDA Cores | 448 |
| Memory | 3 GB GDDR5 SDRAM |
| Bandwidth | 144 GBps |
| PCI | PCI Express 2.0 x16 |
| Core Clock | 1150 MHz |
| Bus Width | 384-bit |