Laboratorio PARCO (Parallel Computing)

rr
The PARCO (Parallel Computing) lab is devoted to the development, analysis, and tuning of parallel code for multi-core and many-core (GPU) architectures.
It provides HW and SW environment for developing in CUDA, OpenCL, OpenACC, OpenMP, MPI targeting performance, power, and energy consumption design constraints.

Servizi

- 6 workstations Linux

- 5 server CPU/GPU:
1. AMD Phenom II X6 1055T (6x 2.8Ghz - Turbo @ 3.3GHz), 8GB RAM, NVIDIA GeForce GTX 780 (2304 cores - 3GB)
2. AMD FX-4100 (4x 3.6GHz - Turbo @ 3.8GHz), 8GB RAM, NVIDIA GeForce GTX 460 336 cores 1GB + GeForce GTX 570 (480 cores - 1.28GB)
3. AMD APU A8-7600 (4x 3.1Ghz - Turbo @ 3.8GHz), 16GB RAM, NVIDIA GeForce GTX 980 (2304 cores - 3GB)
4. AMD Phenom II X6 1055T (6x 2.8Ghz - Turbo @ 3.3GHz), 8GB RAM, NVIDIA Tegra K40 (2880 cores - 12GB)
5. AMD APU A8-3870 (4x 3GHz) , 8GB RAM, GeForce GTX 570 HD (480 cores - 2.5GB).

- 2 low-power heterogeneous embedded systems:
1. NVIDIA Jetson TK1
2. NVIDIA Jetson TX1

- Powermon2 power monitoring device

- 1 cluster (Linux, Torque, 20 x86 64bit nodes)
 
AVVISI

Nessuna categoria trovata.

Organizzazione

Strutture del dipartimento