Hello
I am running the TurboSmarts code, as directed in README.TurboSMARTsim. The
experiment of mcf over test data runs fine.
However, as shown in paper, Simulation Sampling with Live-points, the full
warming of SMARTS and Checkpointed warming of Live-points take different
times (7 hours and 91 seconds average). How do we interpret the data
obtained in step 1 and step 4, to see this fact. It seems the time taken in
step 4 is greater than that in step 1. Also let me know if there is mistake
in assumption/understanding.
Here is the part of output obtained in step 1:
sim_meas_insn               2077155 # total number of instructions measured
sim_detail_insn             6256061 # total number of instructions in
detailed simulation
sim_pop_insn              259643186 # total number of instructions in
functional and detailed simulation
sim_sample_size                2077 # sample size when sampling is enabled
sim_sample_period            125000 # sampling period when sampling is
enabled
sim_num_refs                1029693 # total number of loads and stores
committed
sim_num_loads                756879 # total number of loads committed
sim_num_stores          272814.0000 # total number of stores committed
sim_num_branches             351709 # total number of branches committed
sim_elapsed_time                 68 # total simulation time in seconds
sim_inst_rate            30546.3971 # simulation speed (in measured
insts/sec)
sim_pop_rate           3818282.1471 # simulation speed (in all simulated
insts/sec)
sim_total_insn              3364656 # total number of instructions executed
sim_total_refs              1476551 # total number of loads and stores
executed
sim_total_loads             1134981 # total number of loads executed
sim_total_stores        341570.0000 # total number of stores executed
sim_total_branches           660280 # total number of branches executed
sim_meas_cycle              3365061 # total number of cycles measured
sim_IPC                      0.6173 # instructions per cycle
sim_CPI                      1.6200 # cycles per instruction
totalSeconds                67.6300 # total user and system time (secs)

Here is the part of output obtained in step 4
sim_sample_size                5769 # total number of checkpoints processed
sim_meas_insn               5769128 # total number of instructions measured
sim_detail_insn            17375854 # total number of instructions in
detailed simulation
sim_pop_insn               47295021 # total number of instructions in
functional and detailed simulation
sim_num_refs                2863030 # total number of loads and stores
committed
sim_num_loads               2110342 # total number of loads committed
sim_num_stores          752688.0000 # total number of stores committed
sim_num_branches             977407 # total number of branches committed
sim_elapsed_time                191 # total simulation time in seconds
sim_inst_rate            30204.8586 # simulation speed (in measured
insts/sec)
sim_pop_rate            247617.9110 # simulation speed (in all simulated
insts/sec)
sim_total_insn              9399201 # total number of instructions executed
sim_total_refs              4138946 # total number of loads and stores
executed
sim_total_loads             3196010 # total number of loads executed
sim_total_stores        942936.0000 # total number of stores executed
sim_total_branches          1850179 # total number of branches executed
sim_meas_cycle              9379554 # total number of cycles measured
sim_IPC                      0.6151 # instructions per cycle
sim_CPI                      1.6258 # cycles per instruction
totalSeconds               181.5900 # total user and system time (secs)
-- 
Thanks and Regards
Sparsh Mittal
Graduate Student
Electrical and Computer Engineering
Iowa State University, Iowa, USA
-------------- next part --------------
An HTML attachment was scrubbed...
URL: 
<http://sos.ece.cmu.edu/pipermail/simflex/attachments/20090729/b3f62147/attachment.html>

Reply via email to