Communications of the ACM
Proceedings of the Workshop on Micro-kernels and Other Kernel Architectures
The V Kernel: A Software Base for Distributed Systems
IEEE Software
Distributed and multiprocessor scheduling
ACM Computing Surveys (CSUR)
A system software architecture for high-end computing
SC '97 Proceedings of the 1997 ACM/IEEE conference on Supercomputing
Designing a highly-scalable operating system: the Blue Gene/L story
Proceedings of the 2006 ACM/IEEE conference on Supercomputing
The blue gene/L supercomputer: a hardware and software story
International Journal of Parallel Programming
Early evaluation of the cray XT3
IPDPS'06 Proceedings of the 20th international conference on Parallel and distributed processing
Hi-index | 0.00 |
This article presents an overview of PUMA (Performance-oriented,User-managed Messaging Architecture), a message-passing kernel formassively parallel systems. Message passing in PUMA is based onportals - an opening in the address space of an applicationprocess. Once an application process has established a portal,other processes can write values into the portal using a simplesend operation. Because messages are written directly into theaddress space of the receiving process, there is no need to buffermessages in the PUMA kernel and later copy them into theapplications address space. PUMA consists of two components: thequintessential kernel (Q-Kernel) and the process control thread(PCT). Although the PCT provides management decisions, the Q-Kernelcontrols access and implements the policies specified by the PCT.