Hi Kostas, Thanks for the explanation. Very cool tool!
As I mentioned in my review, I think it would be great to integrate the fault and temperature inputs into the overall GARNET SimObject configuration. Please let me know if you have any specific questions making that work. Brad > -----Original Message----- > From: gem5-dev-boun...@gem5.org [mailto:gem5-dev- > boun...@gem5.org] On Behalf Of Konstantinos Aisopos > Sent: Friday, July 08, 2011 5:41 PM > To: gem5 Developer List > Subject: [gem5-dev] review request: integrating a NoC fault model into > GEM5's GARNET > > hi all, > > I am a grad student at Princeton Univ, advised by Li-Shiuan Peh (MIT), and I > have recently developed a fault modeling tool for NoCs, which I would like to > incorporate into GEM5. > > Overview: this tool accurately models circuit failures; it was developed by > synthesizing multiple routers' RTL and performing Monte Carlo simulations > on their SPICE model netlists. On the other hand, it is wrapped with an > architectural interface, which eases the evaluation of resilient NoC/system > designs by abstracting circuit-level complexity. The tool interfaces with > GARNET, gets as input a (fixed) temperature and each router's configuration, > and provides (for each > router) the probability of occurrence for a vector of architectural fault > types > (data corruption, misrouting, flit duplication / loss / split in pieces, > erroneous > credit generation, credit loss, erroneous allocation of virtual channels / > switch, unfair arbitration). GARNET users can read the probability of > occurrence for the faults their resilient architecture is tackling, inject > these > faults into GARNET at runtime, and evaluate realistically how fault tolerant > their architecture is. > > ...more about the fault model: > [publication] Konstantinos Aisopos, Chia-Hsin Owen Chen, and Li-Shiuan > Peh, "Enabling System-Level Modeling of Variation-Induced Faults in > Networks-on-Chip", DAC '11 [PDF] > http://projects.csail.mit.edu/wiki/pub/LSPgroup/PublicationList/FaultModel > _DAC2011.pdf > [webpage] http://www.mit.edu/~kaisopos/FaultModel > ...more about me: http://www.mit.edu/~kaisopos > > I ll go ahead and publish a review request and I am looking forward to your > comments! > > regards, > -Kostas > _______________________________________________ > gem5-dev mailing list > gem5-dev@gem5.org > http://m5sim.org/mailman/listinfo/gem5-dev _______________________________________________ gem5-dev mailing list gem5-dev@gem5.org http://m5sim.org/mailman/listinfo/gem5-dev
