论文信息 - Cray Advanced Platform Monitoring and Control (CAPMC)

Cray Advanced Platform Monitoring and Control (CAPMC)

With SMW 7.2.UP02 and CLE 5.2.UP02, Cray released its platform monitoring and management API called CAPMC (Cray Advanced Platform Monitoring and Control). This API is initially directed toward workload manager vendors to enable power-aware scheduling and resource management on Cray XC-series systems and beyond. In this paper, we give an overview of CAPMC features, applets, and their driving use cases. We further describe the RESTful architectur e of CAPMC and its security model. Finally, we preview future enhancements to CAPMC in support of in-band control and additional use cases. Keywords-Power monitoring; power capping; RAPL; energy efficiency; power measurement; Cray XC30; Cray XC40

Steven Martin | Matthew Kappel

[1] Russ Housley,et al. Internet X.509 Public Key Infrastructure Certificate and CRL Profile , 1999, RFC.

[2] Douglas Crockford,et al. The application/json Media Type for JavaScript Object Notation (JSON) , 2006, RFC.

[3] James H. Laros,et al. Power/Energy Use Cases for High Performance Computing , 2013 .

[4] Martin Schulz,et al. Exploring hardware overprovisioning in power-constrained, high performance computing , 2013, ICS '13.

[5] Steven J. Martin. Cray XC30 Power Monitoring and Management , 2014 .

[6] Laxmikant V. Kalé,et al. Analyzing Energy-Time Tradeoff in Power Overprovisioned HPC Data Centers , 2015, 2015 IEEE International Parallel and Distributed Processing Symposium Workshop.