Hi,
I am running my application on 16 processors with ruby for a while and then
load opal and step 100million instructions. Now I can dump ruby and opal
statistics.
Both seem to agree on number of cycles, but there is something that I don't
understand with instruction count.
Opal says that
[0] total number of instructions committed: 100000002
[0] total number of instructions decoded: 105486715
[0] total number of instructions executed: 101600065
[1] total number of instructions committed: 117935284
[1] total number of instructions decoded: 121728002
[1] total number of instructions executed: 119055476 so on.
I understand that different processors may be idling so on and we can have
difference in instructions committed.
Ruby says that
R uby_current_time: 58905273
Ruby_start_time: 11538060
Ruby_cycles: 47367213
instruction_executed: 1223603037 [ 81105382 80324304 75240703 84349669
69569396 74585663 78035505 85433618 71846017 75168266 80232199 76919609
67920704 72447556 75125355 75299091 ]
simics_cycles_executed: -738438776 [ -46152236 -46152236 -46152236 -46152236
-46152236 -46152236 -46152236 -46152236 -46152236 -46152236 -46152236
-46152236 -46152236 -46153236 -46153236 -46153236 ]
Here it seems that even though ruby is started earlier it has less
instruction executed count. Why is that?
And why are the reported cycles are negative?
Thanks
Berkin
|