The picture of the engineering drawing is garbled after solidworks renaming.

CAE (Computer Aided Engineering) is an approximate numerical analysis method to solve the mechanical properties of complex engineering and products, such as structural strength, stiffness, buckling stability, dynamic response, heat conduction, three-dimensional multi-body contact, elasticity and plasticity, and to optimize the design of structural properties.

Lianzhong cluster solution: 1, which adopts four E5-4600V2 series supercomputing platforms (the main frequency is higher than 2.5GHZ), which is suitable for molecular dynamics applications and realizes large-scale parallelism; It can also be used as a supercomputing center or a computing node.

2. For GPU computing platform, NVIDIA TESLA K80/40 series high-performance computing cards are recommended, which are suitable for molecular dynamics applications supporting GPU and applications developed by users themselves.

3. At the same time, it is equipped with a pre-processing workstation for grid division and visual display, and also equipped with a professional GPU card.

4. The storage system adopts 36-bit mass storage. Due to the high I/O pressure, the storage and reading performance requirements are fast.

5. Infiniband FDR(56GB) is adopted in the cluster computing network, and 10Gigabit network is adopted in the management network, which improves the network efficiency and reduces the time for CPU to wait for data.

6. Manage and apply software, install the system according to users' needs, and build the use environment. The software environment includes the operation and maintenance management system based on Linux core, the construction of MPI parallel environment, GCC series compilers, C, C++, fortran, java, etc. And scientific computing library BLAS/Atlas/LAPACK/Scalapack/FFTW. Senior engineers optimize PBS job management and job scheduling system.

7. All nodes of Lianzhong Cluster adopt 80PLUS high-efficiency (96%) power supply, which improves power conversion efficiency and reduces user operation and maintenance costs. Combined with cluster management monitoring software, idle resources are automatically set to standby or shutdown state, saving 20% overall energy.