Enterprise-Grade HPC – From workstations to Cloud

Accelerate AI and HPC workloads with our innovative multi-architecture solutions that offers up to 30% better performance/watt than GPU-only solutions.

Brane LLC at SC24: Showcasing Workstations with Triple Accelerators (AMD GPUs, Kalray Manycore, Xilinx FPGA) in a Desktop Form Factor for your lab, and an SDK for Unified Programming of heterogeneous systems.

Featured Products

Brane HX-9000 Triple Accelerator Workstation
Brane Unified SDK
Kalray TC4 Data Processing Accelerator
HX-9000 Triple Accelerator
ComponentSpecs
ProcessorAMD Threadripper Pro
Accelerator 1AMD Instinct GPU
Accelerator 2Kalray Coolidge v.2
Accelerator 3AMD Alveo accelerator
  • 15% faster AI model training compared to equivalent GPU-only systems
  • 2.2x better performance/watt for data preprocessing workloads
  • Unified programming environment across all accelerators
  • Complex AI model development and training
  • Real-time defense simulation and analysis
  • Medical imaging and genomic sequencing
Brane Unified SDK
  • Supported Languages: C++, Python, OpenCL
  • Integrated Framework Support: TensorFlow, PyTorch, RAPIDS
  • Deployment Targets: On-premise workstations and cloud environments
  • Licensing: Perpetual and subscription options available
  • Single programming environment for CPU, GPU, FPGA, and Kalray DPUs
  • Built on IntelliJ with specialized debugging and profiling tools
  • Automatic workload distribution and optimization
  • Performance analytics and bottleneck identification
  • Reduced development time by 65% for multi-architecture applications at a leading defense contractor
Kalray TC4 Accelerator
  • 80 processor cores with VLIW architecture
  • 128 GB/s memory bandwidth
  • PCIe Gen4 x8 interface
  • 70W maximum power consumption

Key Advantages:

  • Superior performance/watt compared to GPUs
  • Optimized for real-time data processing and analytics
  • Programmable using standard frameworks
  • Lower thermal footprint in dense computing environments

Ideal Applications:

  • Network Processing and Real-time data analytics
  • AI inference at the edge
  • Signal and image processing

Key Markets and Applications


Connect With Us