With the continual explosion of data, analytics needs are growing and the demand for responsiveness in achieving these insights are changing.
With these growing needs come unsustainably growing costs. A new form of processing technology is needed to harness the power of your data without the unreasonable increases to your infrastructure spend.
Accelerator technology is the new norm, but technology such as GPUs, FPGAs, TPUs, and the multitude of AI specific technology does nothing for this large segment of your current existing CPU-based applications. Intensivate’s accelerator technology allows you to get the most out of your scale-out applications without any needed changes to your software environment.
Architected specifically for memory intensive parallel processing applications and workloads requiring high compute and throughput.
The IntenScale card is configured in a 1U rackmount server node with a motherboard using a low power host CPU, 32GB of memory and a dual port 100Gb network card. The host CPU is used for local disk access and file system purposes only. The entire application workload is loaded and runs onto the accelerator card, keeping the host system cost low and eliminating the need for any major changes to application software.
Enterprise businesses are becoming more analytically data driven in order to compete and advance. The software being used to create this data is currently being run on server nodes designed for older, traditional data processing needs. The result: excess, inefficient amounts of servers creating operational nightmares and massive costs. The situation has been made worse with the consistent performance increase reductions being seen in every new release of the general purpose CPU. With the end of Moore’s Law, hardware system architecture trends have taken a turn. The next evolution of computing is to leverage standardized servers with general purpose CPUs, equipped with accelerator cards to handle workload specific applications.