Can dtrace help with collecting any of the following, in otherwords are
there any existing counters to be exploited, or am I going to have to make
my own custom Solaris build with additional counters...
currently I can get this detailed trap information from our SPARC simulator,
for example here are some ints-per-sec during a benchmark-of-interest:
sending-source
recv network disk cross-cpu, by sources 0...7
cpu ints INO ints INO 0 1 2 3 4
5 6 7
0 2453 0x684 197 0x688 72 46 47 47 46
47 46
1 1445 0x610 278 0x780 102 46 48 47 47
47 47
2 2443 0x60c 101 73 48 47 47
46 47
3 2423 0x640 101 73 49 47 46
47 47
4 2457 0x680 101 72 47 49 47
46 48
5 2455 0x6c0 99 71 47 48 48
47 47
6 2430 0x600
+2394 0x6c4 97 70 46 47 47
48 47
7 2449 0x64c 8 0x604 99 71 47 47 47 47 49
I also want to break-down the cross-cpu calls by function, but have not
yet done
so on the simulator.
thanks,
Pete Lawrence.
Did you try trapstat(1M)? -surya Peter Lawrence wrote On 02/06/07 01:07,:>Can dtrace help with collecting any of the following, in otherwords are >there any existing counters to be exploited, or am I going to have to make >my own custom Solaris build with additional counters... > > > >currently I can get this detailed trap information from our SPARC simulator, >for example here are some ints-per-sec during a benchmark-of-interest: > > sending-source > recv network disk cross-cpu, by sources 0...7 > cpu ints INO ints INO 0 1 2 3 4 >5 6 7 > 0 2453 0x684 197 0x688 72 46 47 47 46 >47 46 > 1 1445 0x610 278 0x780 102 46 48 47 47 >47 47 > 2 2443 0x60c 101 73 48 47 47 >46 47 > 3 2423 0x640 101 73 49 47 46 >47 47 > 4 2457 0x680 101 72 47 49 47 >46 48 > 5 2455 0x6c0 99 71 47 48 48 >47 47 > 6 2430 0x600 > +2394 0x6c4 97 70 46 47 47 >48 47 > 7 2449 0x64c 8 0x604 99 71 47 47 47 47 49 > > >I also want to break-down the cross-cpu calls by function, but have not >yet done >so on the simulator. > > >thanks, >Pete Lawrence. > >_______________________________________________ >dtrace-discuss mailing list >dtrace-discuss at opensolaris.org > >