"Helping our customers break the business performance barrier every day"
The choice of a new compute platform is a daunting decision for any engineering organization. The choice represents a large investment in an engine which must perform reliably, predictably, for a reasonable price. Often, Tier 1,2,3 suppliers to OEMs need access to small scale HPC facilities but cannot afford the overhead of staff with such expertise. Mach1 Computing can be engaged for precisely this reason and act as our customer's agent. Such an engagement usually starts with a customer personnel facing requirements gathering and validation exercise. Then we can interact with hardware vendors, VARs, ISVs, and integrators to solicit submissions from each to build a solution which matches the customer's requirements.
Often in HPC, validation is also required in the form of benchmarks which some vendors run themselves and others lack the ability or capacity to do so. Mach1 can orchestrate and perform the benchmarks whenever required.
In the final analysis, this kind of engagement minimizes the customer's risk for his HPC investment. Mach1 Computing can provide sample ROI calculations from post-analyzes of other customer environments to prove not only the risk minimization, but the financial gains to be had when HPC experts act as agents for you.
Savvy customers have recognized that CAE software licensing costs are rivaled only by the real and opportunity cost of engineering time. Therefore the simplest method to maximize the investment in CAE software licenses is to assure those licenses run their corresponding application at peak efficiency (this minimizes engineering time spent waiting for results and provides maximum innovation thought process continuity).
Mach1 Computing offers services around the top MCAE applications which goes to the heart of this notion. Applications such as MSC and NX Nastran, LS-Dyna, ABAQUS, Fluent, Star-CD, MSC Marc, and many others are prime targets for our system tuning techniques.
Customers interested in this service are not required to have current support contracts with Mach1 or to buy hardware to achieve additional performance (We can provide hardware upgrade recommendations for additional improvements).
No HPC provider has in-depth industrial and domain experience as our team and no other services vendor has the systems knowledge and experience our team has developed. These services provide the customer with the confidence that their hardware is performing at the highest level it can achieve, and also provides a road map for further improvements should they be needed.
Workload management or "batch queuing" refers to the use of middleware which collects users submitted job, resource requests and distributes them amongst resources according to a well defined strategy. This concept is extremely useful in a resource poor environment where the demands of a user or users are significant compared to the amount of available resource. By distributing job load according to a strategy, a batch queuing system can maximize throughput (the number of jobs a resource can run in a given period of time) which minimizes TCO.
A very large number of virtual product design customers in the marketplace currently run their application environments without the benefit of queuing in favor of ad-hoc queuing dependant on email, voice, or even white-boards. As your environment grows these methods, which may work for small 1-4 person environments with long wait times between job launches, will quickly breakdown.
Therefore we offer a set of inclusive services which can bring different workload management systems online. Multi-platform, multi-architecture, multi-processor, multi-core, multi-host shared and distributed memory environments are supported out of the box.
The real world provides several situations where a little help from a trusted partner can be the difference between confidence and nightmarish worry.
Several customers have leveraged our expertise and ability to travel to minimize downtime, get to the heart of a problem quickly, help with that upgrade which is too daunting to do on your own, or provide clear concise direction to far reaching questions. This is just a small portion of the value our team can provide.
Several companies that have heavy demands for high performance computing have already started to explore distributed computing clusters as a production platform. The vendor landscape continues to change as the economics of HPC clusters on commodity hardware prove some business models while defeating others. This changing landscape has left several companies with unsupported HPC systems, even though these systems have not reached their expected end of life.
With some qualification, Mach1 Computing can extend various levels of support to customer who have solutions not designed or integrated by ourselves. The primary initial customer targets for this support service are the former customers of MSC Software HPC and Paracel, though we welcome engagement from other HPC consumers as well.