The computational laboratories have local clusters with computing nodes available (948 CPUs & 4 desktop GPU nodes & 22 rack form factor GPU nodes), and have had regular access to the Supercomputers of the Consortium of University Services of Catalonia. Moreover, we have had ca. 40 projects of 4 months at the Red Española de Supercomputación for a total of ca. 8 million hours. To handle these facilities, the IQCC has one informatics technician and two laboratory technicians.
The Institute of Computational Chemistry and Catalysis owns a CPU-based cluster called Beta, and some GPU-based computing nodes called BioComp and a large GPU cluster called GALATEA.
CPU cluster (called BETA)
IQCC computational resources have been purchased thanks to the FEDER Funds (European Funds for Regional Development) in collaboration with “Ministerio de Economía y Competitividad” of Gobierno de España. In particular, the beta nodes gatchan1-24 and the gamma nodes gatchan25-44 were purchased thanks to the FEDER UNGI08-4E-003 and UNGI10-4E-801 grants, respectively. Beta nodes gatchan47-57 have been purchased thanks to the FEDER and MINECO by CTQ2014-54306-P, CTQ201569363-P and MPCUdG2016/096 resources.
The GPU-based nodes have been purchased thanks to the CIG project (DIREVENZYME: 2013-CIG-630978).
CPU Calculation nodes | Cores |
---|---|
gatchan1 to gatchan24 (@hpiqc3) | 288 |
gatchan25 to gatchan44 (@hpiqc4) | 320 |
gatchan45 to gatchan46 (@hpiqc5) | 32 |
gatchan47 to gatchan57 (@hpiqc6) | 308 |
gatchan58 to gatchan60 (@hpiqc7) | 120 |
gatchan61 to gatchan65 (@hpiqc8) | 220 |
gatchan66 to gatchan70 (@hpiqc9) | 240 |
TOTAL | 1528 |
Specifications:
For each gatchan calculation node (nodes from 1 to 24):
- 2 x Intel(R) Xeon(R) CPU X5650 @ 2.67GHz and 12mb of cache
- 12 cores per node (6 per CPU)
- 880Gb of hard disk for scratch (/sgetmp)
- 36Gb of memory (@1333mhz)
For each gatchan calculation node (nodes from 25 to 44):
- 2 x Intel(R) Xeon(R) CPU E5-2670 @ 2.60GHz and 20mb of cache
- 16 cores per node (8 per CPU)
- 440Gb of hard disk for scratch (/sgetmp)
- 64Gb of memory (@1600mhz)
For each gatchan calculation node (nodes from 45 to 46):
- 2 x Intel(R) Xeon(R) CPU E5-2650 v2 @ 2.60GHz and 20mb of cache
- 16 cores per node (8 per CPU)
- 440Gb of hard disk for scratch (/sgetemp)
- 256Gb of memory (@1866mhz)
For each gatchan calculation node (nodes from 47 to 57):
- 2 x Intel(R) Xeon(R) CPU E5-2680 v4 @ 2.40GHz and 35mb of cache
- 28 cores per node (14 per CPU)
- 5.5Tb of hard disk for scratch (/sgetmp)
- 512Gb of memory (@2400mhz)
For each gatchan calculation node (nodes from 58 to 60):
- 2 x Intel(R) Xeon(R) Gold 6248 CPU @ 2.50GHz and 30mb of cache
- 40 cores per node (20 per CPU)
- 8Tb of hard disk for scratch (/sgetmp)
- 196Gb of memory (@2933mhz)
For each gatchan calculation node (nodes from 61 to 65):
- 2 x Intel(R) Xeon(R) Gold 6248R CPU @ 3.00GHz
- 44 cores per node (22 per CPU)
- 16Tb of hard disk for scratch (/sgetmp)
- 386Gb of memory (@3200mhz)
For each gatchan calculation node (nodes from 66 to 70):
- 2 x Intel(R) Xeon(R) Gold 5318Y CPU @ 2.10GHz
- 48 cores per node (24 per CPU)
- 4Tb of hard disk for scratch (/sgetmp)
- 256Gb of memory (@3200mhz)
Biocomp mini-cluster. GPU nodes for testing
Specifications:
The GPU cluster is now composed by four machines with the following specifications:
– Intel i7-4770 V3 processor at 3,4 GHz (3,9 GHz turbo) 4 cores (8 processes)
– 2*Nvidia GTX780 Graphic Cards, 3 GB RAM DDR5, 2304 cores per card
– 16 GB DDR3@1600mhz Kingston RAM memory
– 2 SATA II Seagate Constellation hard drives
GPU cluster (called GALATEA)
GALATEA cluster resources have been purchased thanks to the “ERC: European Research Council” by ERC-2015-StG-679001.
Specifications:
Galatea GPU cluster is composed by 22 machines with the following specifications:
– Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz
– 8*Nvidia GTX1080 Graphic Cards
– 128 GB DDR4@2133mhz Micron RAM memory
– 4Tb for scratch
There’s also 2 administration nodes and an external storage for user’s data all running over 1Gb private network.
Here some pictures of BETA.
Other resources
3D printer
CCDC
The Cambridge Structural Database is an incredible community achievement as for over fifty years crystallographers have shared the results of every published crystal structure with the CCDC – allowing us to share them with the world. The CCDC is committed to enabling everyone to benefit from the world’s comprehensive and up-to-date database of crystal structures.
SciFinder
SciFinder® is a comprehensive chemistry information resource that enables you to make better, more reliable decisions from the beginning to the end of your research project.