Swissmade
SEMCAD X
semcadmatter01.png
HPC Solution

HPC SOLUTION

SEMCAD X Matterhorn offers high-performance computing (HPC) for investigation of complex and realistic models. Multi-threaded execution for modeling, meshing, voxeling, and postprocessing enables parallel processing of heavy tasks without disturbing the workflow. A fully integrated centralized task manager efficiently manages all computationally intensive tasks on the local machine or in the cloud. Two sets of CUDA-based libraries for GPU acceleration, Acceleware and SPEAG CUDA, are provided.

 

SEMCAD X Matterhorn features the fastest available GPU-enabled EM-FDTD solver. The MPI parallelization-based EM-FEM solvers (QS) optimally use multi-core processors, clusters, and supercomputers to guarantee extreme performance for demanding tasks. A unified interface supports cloud computing on any of the major providers, e.g., Amazon or Google.

 

 

Key Features: Acceleware

  • Advanced libraries for EM-FDTD
  • Unique FDTD/GPU subgridding scheme (structure-adaptive)
  • Support for this sheets, dispersive media and lossy metals
  • Multi-GPU support

 

 

Key Features: CUDA

  • Libraries for EM-FDTD
  • Multi-GPU support

 

Key Features: MPI

  • Libraries for FEM-based EM-QS
  • MPI/OpenMP support of various SEMCAD X Matterhorn embedded Framework features
 

Supported GPUs

  • NVIDIA Ampere architectures
    • GeForce RTX 30 series
    • Workstation RTX Ax000 series
    • Datacenter A series
  • NVIDIA GeForce 16 and 20 series series (Volta and Turing architecture)
    • NVIDIA TITAN V, NVIDIA Quadro GV100
    • GeForce GTX 1650, GTX 1660, GTX 1660 Ti
    • GeForce RTX 2060, RTX 2070, RTX 2080, RTX 2080 Ti
    • NVIDIA Titan RTX
    • Quadro RTX 8000, Quadro RTX 6000, Quadro RTX 5000, Quadro RTX 4000
    • NVIDIA Tesla T4
  • NVIDIA GeForce 10 series (Pascal architecture):
    • GeForce GTX 1050, GeForce GTX 1050 Ti, GeForce GTX 1060, GeForce GTX 1070, GeForce GTX 1080, GTX 1080Ti 
    • NVIDIA Titan X, Titan Xp
    • NVIDIA P5000, P6000, P620
    • NVIDIA Tesla P100
  • NVIDIA GeForce 900 series (Maxwell architecture):
    • First generation Maxwell: GeForce GTX 750Ti, GeForce GTX 750
    • Second generation Maxwell: GeForce GTX TITAN X, GeForce GTX 980, GeForce GTX 980Ti, GeForce GTX 970, GeForce GTX 960, GeForce GTX 960, GeForce GTX 950
  • Depreciated NVIDIA K-series (Kepler architecture):
    • NVIDIA Kepler K10, K20/K20X, K40 and K80.
    • NVIDIA Quadro K6000

Hardware NVIDIA series 20 and older are no longer supported.

Important: In the next release of Sim4Life, the NVIDIA toolkit will be upgraded to support the most recent GPUs. As a side effect, some older GPU architectures will not be supported anymore. In particular, all cards with compute capability lower than 5.0 will be dropped, which includes the Kepler generation GPUs. Please check this link for details about the cards' compute capability and CUDA toolkits.

Supported OS

  • Windows 8.1/10/11
  • Windows Server 2008/2012
  • Ubuntu 18.04 64 bit (or compatible OS), or CentOS Linux 7 64 bit (deprecated support, to be discontinued)

Important: For the CUDA-based solver (CUDA P-EM-FDTD), SEMCAD X supports NVIDIA GPU models equipped with CUDA compute capability 3.5 up to 8.6, and requires device drivers that support the CUDA toolkit 11.0 and above. Note that NVIDIA may choose to discontinue certain drivers for specific operating systems, e.g., CUDA 11.0 drivers may not be available for Windows 8.1 for certain GPU series. In such a case, an update to Windows 10 or 11 may permit further usage of the GPU with Sim4Life.

Performance

Benchmark: Mobile phone simulation

 

Benchmark: Birdcage with ViP human model simulation

 

Benchmark: ViP human model with a mobile phone (with patch array antenna) next to his head.

 

Benchmark: SAM Head and Hand phantoms with a mobile phone.

 

Additional Benchmarks: Small Computational Domain