HW
- HPE DL380 Gen10
CPU : Xeon-G 5215(2.5GHz/10core)*2P
Mem : 256GB PC4-2933Y-R
GPU : Tesla V100-16GB*2
HDD : 240GB*2, 3.8TB SSD*4
NIC : 10Gb 2P NIC*1
Power : 1600W PWR*2
PCIe : x16 FIO Rise Kit
SW(운영환경)
- WinPro 10 SNGL OLP NL Legalization GetGenuine
- WINVDAPerDvc SubsVL OLV D 12Mth AP PerDvc
- Red Hat Enterprise Linux Server, Standard
- Quadro Virtual Data Center Workstation * 10copy
- 기 운영 중인 Vmware 6.7 솔루션에 클러스터 구성
SW(코딩툴킷)
- Orange3(Python 3.6 integrate)
- KNIME
- Ubuntu 18
- CUDA Toolkit 10.1 Update 2
(Toolkit Subpackges/C,C++,JAVA,Python;runable)
- CUDNN(v7.6.3 이상)
Deep learning Primitives, Network library
- TensorRT
Deep learning Inference Engine, TensorRT 6.0 GA
- DeepStream
Deep learning for Video Analytics(NVIDIA SDK)
DeepStream 4.0 for T4 and V100
- NCCL
Multi-GPU Communication
(NVIDIA Collective Communications Libray)
NCCL v2.4.8 for CUDA 10.1 2019.
- JAVA RUNNING(ECLIPSE, DL4J, AlexNet)
- Python 3.6 running