Scalable Machines conducts cutting-edge research in making software achieve scalable peak performance on modern multicore, accelerated, shared- and distributed-memory clusters and HPC establishments.
We are experts in design and development of performance measurement, analysis, and visualization tools. Using our performance-analysis tools and performance-enhancement techniques, we have established a strong track record of tuning complex software to achieve top performance on a variety of platforms: mobile, commodity and enterprise servers, HPC clusters, and custom hardware. We are domain experts in high-performance computing, parallel computing, parallel algorithms, synchronization techniques, computer architecture including accelerators (GPU, FPGA, ASIC), large-memory machines, memory hierarchy, NUMA machines, non-traditional architectures (e.g., ARM servers), and hardware-software co-design. We a have strong presence both in industry and academia.
We offer professional consulting and training in the following areas:
Choosing the right hardware, programming model, and toolchain to best meet your performance, efficiency, and cost criteria.
Developing and deploying software performance measurement and analysis tools best suited for your development and production environments.
Understanding the behavior of complex software on target machine(s).
Pinpointing performance and scalability bottlenecks in your software.
Guiding software design and development for top performance.
Tuning complex software to achieve peak performance on target hardware.
Overhauling legacy software and algorithms to match modern hardware.
Contact: support@scalablemachines.co