[Beowulf] What is the status of "remote prefetching" in QPI?
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
richard.walsh at comcast.net richard.walsh at comcast.netFri Feb 20 15:50:10 PST 2009
- Previous message: [Beowulf] bonding.txt really confusing, why don't I get higher aggregate bandwidth from multiple TCP connections from multiple gigabit clients with balance-alb bonding on a server?
- Next message: [Beowulf] ifort and gfortran
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
All, In early descriptions of QPI, this capability (a remote QPI agent's [on say an FPGA or GPU accelerator] ability to stimulate a QPI processor to prefetch data to its cache, avoiding memory), was listed as a possible feature of QPI. This has obvious potential benefits in globalizing distributed memory and reducing remote memory reference data copies. I am guessing that such a feature would be important in SGI's Tukwila based UltraViolet system, but am wondering if it is part of what will be supported on the soon-to-be-released Nehalem EPs? A side questions would be "Does HyperTransport allow for something similar?" Thanks, rbw -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.scyld.com/pipermail/beowulf/attachments/20090220/180c12fb/attachment.html
- Previous message: [Beowulf] bonding.txt really confusing, why don't I get higher aggregate bandwidth from multiple TCP connections from multiple gigabit clients with balance-alb bonding on a server?
- Next message: [Beowulf] ifort and gfortran
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
