Apex Job Manager (AJM)
Optimize your EDA compute and personnel resources with Apex Semiconductor’s job management solutions.
The Apex Job Manager (AJM) is an EDA-centric grid-computing middleware platform that seamlessly integrates with industry-standard grid managers such as IBM Platform LSF™ and SLURM to provide sophisticated EDA job management tools. AJM’s database facilitates sophisticated reporting and analysis by storing result, resource, and performance data.
Apex Job Manager
- Traditional completion monitoring and pass/fail reporting
- Aggregate job management via AJM Flights
- Robust error handling for unusual events: Crashed jobs, host failures, memory and disk oversubscription, etc.
- Resource management workloads: High-volume/low-loads such as register-transfer level (RTL) simulation, low-volume/high-loads such as physical design runs, and everything in between
Optimize your EDA compute and personnel resources with Apex’s job management solutions.
AJM Database
- Traditional pass/fail reporting
- Support for storage of individual failure messages and aggregate failure signatures
- Automatic capture of compute and license usage as well as workspace and source-code-control state for resource reporting, analysis, and prediction
- Automatic seed generation and storage to facilitate pseudo-random job management
AJM’s database is customer-extensible — limited only by available storage and your imagination.
AJM Analysis
- Traditional, strategic resource management: What jobs use the majority of our resources? Are we bottlenecked by CPUs, memory, or licenses? When do our jobs run? Can demand be more evenly distributed?
- Tactical job/flight management: What queue should this flight of jobs be submitted to? Can this flight of jobs run as a job array?
- Unique predictive analysis: Real-time job augmentation. Predict memory and license requirements and submit jobs with appropriate resource requirements. Predict the likely root cause of a failure, and include it in the failure message and/or signature.