Resources

TUBITAK ULAKBIM High Performance and Grid Computing Center, which started its operations in 2003, is included in TRUBA. Today, TRUBA serves our researchers with ~ 17,500 processor cores, 80 GPUs and 4PByte Luster distributed file system.

 

Year#CPU/GPUCPU ModelSPECfp_rate_base2006Theoric GFlopsMemoryDescription
20031281 core x 1 CPUPentium IV 2.66 GHz--1 GBOut of usage
2006642 core x 2 CPUXeon 5110 1.60 GHz17-4 GBOut of usage
2006362 core x 2 CPUXeon 5110 1.60 GHz17-4 GBOut of usage
2007304 core x 2 CPUXeon X5355 2.66 GHz56.3-16 GBOut of usage
2008164 core x 2 CPUOpteron 2356 1.15 GHz81.336.816 GBOut of usage
20091544 core x 2 CPUXeon X5550 2.67 GHz18485.4424 GBOut of usage
20101312 core x 4 CPUOpteron 6174 2.20 GHz574422.4128 GBGrid Site
201119212 core x 2 CPUOpteron 6176 2.30 GHz294220.8128 GBMercan Cluster, Federated Cloud and Grid Site
20131288 core x 2 CPUXeon E5-2690 2.90 GHz487371.2256 GBLevrek Cluster
20144812 core x 2 CPUXeon E5-2680 v3 2.50 GHz763960256 GBLevrek v2 Cluster
20141612 core x 2 CPU + 2x nVidia M2090Xeon E5-2680 v3 2.50 GHz763960 GFLOPs + 1330 GFLOPs256 GB + 6 GB GDDR5Levrek-CUDA Cluster
2016114 core x 16 CPUXeon E7-4850 v3 2.20 GHz-7.9 TFLOPs4 TBOrkinos SMP Node
201712814 core x 2 CPUXeon E5-2690 v3 2.60 GHz9701164 GFLOPs256 GBSardalya Cluster
201812020 core x 2 CPUXeon Scalable 6148 2.40 GHz14002048 GFLOPs384 GBBarbun Cluster
20182420 core x 2 CPU + 2 x nVidia P100 GPUXeon Scalable 6148 2.40 GHz14002048 GLFOPs + 9400 GFLOPs384 GB + 2 x 16 GB HBM2Barbun-cuda Cluster