Hi Brad, Thanks a lot for your reply and also for the explanation.
I am adding a part of the trace output from Bus trace flag. 2265338330001: system.piobus: recvTiming: src 28 dst -1 ReadReq 0x801fc0003fa 2265338330001: system.piobus: The bus is now occupied from tick 2265338330001 to 2265338332000 2265338331001: system.piobus: recvTiming: src 24 dst 28 ReadResp 0x801fc0003fa BUSY 2265338331001: system.tsunami.uart-pioport: Send failed, waiting on retry 2265338332000: system.piobus: Sending a retry to system.tsunami.uart-pioport 2265338332000: system.tsunami.uart-pioport: Received retry 2265338332000: system.piobus: recvTiming: src 24 dst 28 ReadResp 0x801fc0003fa 2265338332000: system.piobus: The bus is now occupied from tick 2265338332000 to 2265338334000 2265338332000: system.piobus: Remove retry from list 24 This is getting repeated for 4 memory addresses : 0x801fc0003fa 0x801fc0003fd 0x801fc0003fe 0x801fc0003f8 Except the address, the rest of the messages are the same, the source and destination nodes are identical for all 4. These keep repeating in an endless loop and never proceed further. I am running using the following command. ./build/ALPHA_FS_MESI_CMP_directory/m5.opt --trace-flags=Bus ./configs/example/ruby_fs.py -t -n 4 -r 1 Pritha On Wed, Jul 14, 2010 at 11:05 AM, Beckmann, Brad <[email protected]>wrote: > Hi Pritha, > > > > The warnings you mentioned are expected and can be ignored as long as they > only complain about Ruby Sim Objects. The reason for them is the checkpoint > you created using normal ALPHA FS, only checkpointed the Sim Objects used by > normal ALPHA FS, but when a Ruby enabled system loads that checkpoint, it > will search for information associated with Ruby Sim Objects. > > > > The fake ports you mentioned are created to approximate the Tsunami chip > set. See src/dev/alpha/Tsunami.py for more info. > > > > I have never seen instructions being retried when using Ruby. To my > knowledge, RubyPort will never ask for any request to be retried. Could you > provide more information? > > > > Brad > > > > > > *From:* [email protected] [mailto:[email protected]] *On > Behalf Of *Pritha Ghoshal > *Sent:* Tuesday, July 13, 2010 10:27 AM > *To:* M5 Developer List > *Subject:* [m5-dev] Using Ruby to run run from PARSEC checkpoint > > > > Hi, > > I am trying to use the parsec benchmark to create checkpoints using normal > ALPHA FS (without ruby), and then trying to run with > ALPHA_FS_MESI_CMP_Directory or ALPHA_FS_MOESI_Hammer binaries. Though the > checkpoint complete execution successfully if I run with normal ALPHA_FS > without ruby, they seem to hang while running ruby based Alpha. Also in the > beginning I get the a lot of warnings like the following: > > warn: Not unserializing > 'system.ruby.network.topology.ext_links2.ext_node.L2cacheMemory': no section > found in checkpoint. > For more information see: http://www.m5sim.org/warn/dc3e87f8 > warn: Not unserializing > 'system.ruby.network.topology.ext_links2.ext_node.sequencer.icache': no > section found in checkpoint. > For more information see: http://www.m5sim.org/warn/dc3e87f8 > warn: Not unserializing > 'system.ruby.network.topology.ext_links2.ext_node.sequencer.dcache': no > section found in checkpoint. > For more information see: http://www.m5sim.org/warn/dc3e87f8 > warn: Not unserializing 'system.ruby.network.topology.ext_links1': no > section found in checkpoint. > For more information see: http://www.m5sim.org/warn/dc3e87f8 > warn: Not unserializing 'system.ruby.network.topology.ext_links1.ext_node': > no section found in checkpoint. > For more information see: http://www.m5sim.org/warn/dc3e87f8 > > I think the components are not being set up properly, is that what it > means? > > When I enabled the config flag, I realised these are the ports being > created: > > 0: system.piobus-p3: setting peer to system.tsunami.fake_sm_chip-pioport > 0: system.tsunami.fake_sm_chip-pioport: setting peer to > system.piobus-p3 > 0: system.piobus-p4: setting peer to > system.tsunami.fake_uart1-pioport > 0: system.tsunami.fake_uart1-pioport: setting peer to > system.piobus-p4 > 0: system.piobus-p5: setting peer to > system.tsunami.fake_uart2-pioport > 0: system.tsunami.fake_uart2-pioport: setting peer to > system.piobus-p5 > > Could someone help me regarding this? Why are the ports created being fake? > Also for this reason, the instructions are not going through and the same > instruction keeps on looping in retry mode. > > Thank you > > Pritha > > _______________________________________________ > m5-dev mailing list > [email protected] > http://m5sim.org/mailman/listinfo/m5-dev > >
_______________________________________________ m5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/m5-dev
