- Today the demands on volume processing and their HPC/Server Farm environments are enormous and it is not uncommon for some to have issues. For mission critical processing Cloud Based services may not always be the answer. Some may run their own HPCs while many CSPs acknowledge this and provide dedicated servers for this purpose which is a step in the right direction. However they are still restrained by the current generic transport technologies, top heavy XML/SOAP etc. flows, with bloated OS's.
- With our technology your parallel applications will execute and communicate faster, on die, in package, on processor, or remote. Our Technology is purpose built from the ground up for MPP we can deliver faster and cost effectively.
- Our many years experience building mission critical systems for Government and Carrier, underline our ability to deliver an industrial grade results.
- Today 5-10 microseconds latency is not uncommon at the signalling layer. While 10+ us application end to end (ping-pong) for some is considered pretty good. We ping-pong in ~1+/- nano seconds, and as mentioned above, it makes no difference if the nodes are on different dies, packages, processors or chassis cards data to and from your application is direct.
Fine-Grained Execution Control
Having massive processing power at your finger tips is one thing, getting actual compute time is another. We provide the technology for you to run immediate and/or scheduled job loads. In addition, our unique architecture gives you fine-grained execution control allowing ad hoc runtime design and testing, shortening your project design cycle, costs and time to market.
NOTE: the clock ndelta is a measure of the timer library call, because we are working at very high speeds even looking at your watch incurs a significant time cost. The intf_loss ~5ns inter-card.
- Whether your average message size is a few bytes or Gigs (our framing is 8/32K) our system fabric is cache and pipeline packed/aligned for optimum message processing