On Mon, 2009-12-21 at 15:58 +0100, Thomas Sailer wrote: > I can't currently give you the output of ./tbench --stats, as this > crashes after finishing the testbench but before dumping the statistics. > I need to debug that first. Thank you for your patience.
Sorry for the delay; I first had to find the cause of the segfault (see bug #15015). Here are the stats: total: wall: 1248.460 user: 1054.400 sys: 3.860 elab: wall: 0.230 user: 0.050 sys: 0.050 internal elab: wall: 0.130 user: 0.070 sys: 0.000 cycle (sum): wall: 1248.100 user: 1054.280 sys: 3.810 processes: wall: 116.370 user: 99.810 sys: 0.650 resume: wall: 0.740 user: 0.160 sys: 0.370 update: wall: 41.480 user: 34.540 sys: 0.810 next compute: wall: 1089.510 user: 919.770 sys: 1.980 Number of simple signals: 55525 Number of signals with projected wave: 12198 Number of non-driven simple signals: 240 Number of resolved simple signals: 131 Number of multi-sourced signals: 97 Number of signals whose activity is managed: 23110 Number of drivers: 58597 Number of direct drivers: 30039 Number of signals per mode: B2: 36 E8: 55386 E32: 0 I32: 14 I64: 3 F64: 86 Propagation table length: 33 Propagation table count: Drv (1 drv) : 16 Eff (1 prt) : 2 Eff 1 resolved : 5 Eff Actual : 2 end : 8 Number of delta cycles: 346722 Number of non-delta cycles: 114998 Nbr of events: 8052110 Nbr of active: 9531510 Number of processes: 26864 Number of sensitized processes: 26767 Number of resumed processes: 15726789 Average number of resumed processes per cycle: 34 This is with svn133, plus the --stats patch. It's however not for the full run, I'm getting the internal errors back: ./tbench:internal error: process in timeout Tom _______________________________________________ Ghdl-discuss mailing list [email protected] https://mail.gna.org/listinfo/ghdl-discuss
