草榴社区 Cloud
Cloud native EDA tools & pre-optimized hardware platforms
Request a Free Trial →
Runtime Manager
The LynxNXT Automation System includes a patented GUI that simplifies and automates flow creation, configuration, and maintenance to improve productivity.
Intuitive and easy-to-use, LynxNXT Automation’s Runtime Manager provides the ability to graphically edit, execute and monitor customer flows as a design progresses from RTL to tape-out. Creating a new design flow or modifying existing flows is accomplished using intuitive edit operations. Flat or hierarchical design flows for a single block or for an entire chip, are defined by connecting various tasks together based on their execution dependencies. Parallel task execution, design exploration involving multiple runs of the same task or experimenting with different parameters in the design flow are also easily accomplished within LynxNXT Automation System’s Runtime Manager.
Figure 1: Patented GUI enables easy creation, configuration, and debug of design flows at multiple levels
Adaptive Resource Optimizer
Included with LynxNXT Automation System’s Runtime Manager is a distributed high-performance computing (HPC) workload management optimization feature called Adaptive Resource Optimizer (ARO). ARO monitors usage patterns of submitted jobs and determines a more optimal value to be used for reserving compute resources (e.g. memory, CPUs) of future job submissions based on historical use data.
Figure 2: Adaptive Resource Analyzer
ARO can substantially reduce job pending time, the time a job sits in queue waiting for a resource and improve compute resource utilization by ensuring that jobs are assigned to the best queue/machine combination. The busier the HPC environment, the greater the benefit realized from ARO compared to traditional fixed resource management practices. Improvements of 10% or more in each job’s turnaround time, the time it takes from job submission to getting results, can be achieved using ARO. When aggregated across thousands of jobs, ARO significantly improves the productivity of the design team.
ARO currently supports LSF, SGE and UGE, but can be adapted for proprietary environments as well.
Podcast