The documentation has improved significantly, but it still can be much better, the simulation speed side is particularly vague in the slides. It would be nice if somebody could also write about how the gem5 works in implementation details and its design philosophy, to help another generation of simulator builders as well.
On 9 Jun 2011, at 20:51, Richard Strong wrote: > Hi all, > > First, I wanted to share my appreciation to the gem5 community for putting up > the great documentation and slides. However, the example I tried has a few > problems that may send users for a loop so I figure I would post it here to > see if we can get some updated slides. For 64 socket SMP, 2-Level on-chip > Caches, MOESI protocol, 32K L1s, 8MB L2 per chip, mesh interconnect the > tutorial says to do: > > ./build/ALPHA_FS/m5.op configs/example/ruby_fs.py -n 64 -l1i_size=32kB > -l1d_size=32kB -l2_size=512MB -num-l2caches=64 -topology=Mesh -timing > > But this leads to complaints with the options compiler, because "-" should be > "--". If you fix this, Mesh.py on line 80 throws an exception because it > expects a DMAController but finds a DirectoryController. The change here > after looking in the mailing list is to add --num-dirs=64. The final > functional version I show below: > > ./build/ALPHA_FS/m5.opt configs/example/ruby_fs.py -n 64 --l1i_size=32kB --l > 1d_size=32kB --l2_size=512MB --num-l2caches=64 --topology=Mesh --timing > --num-dirs=64 > > Also, please note that the example for "8 core CMP, 2-Level, MESI protocol, > 32K L1s, 8MB 8-banked L2s, crossbar interconnect" has a similar problem with > how the options were added where only a single '-' appears in the slides as > opposed to '--'. I have not tested this example yet, but saw a similar > format. > > Best, > Rick _______________________________________________ gem5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
