Thank you for replying.
My problem is simulation is not finished.(IPC is very small)
I tried the simulation following simulation environment once again.
I am trying to run the Splash2 in Chip MultiProcessors using SE mode.
I get the Splash2 binary in the following link.
http://www.m5sim.org/dist/m5_benchmarks/v1-splash-alpha.tgz
I try to run SPLASH2 in Detailed 4 CPU cores
using default .py file (m5_1.1/m5/configs/splash2/run.py)
But Simulation isn't finished. (IPC is very small, too)
Therefore cause for not finishing the simulation is not "prefetching".
I'm sorry for the mistake.
I have no idea what cause this problem.
Could you tell me how to run Splash2 benchmarks correctly?
following line is my comand line message and m5stat.txt
##### comand line messege ######
command line: ./m5.fast -ENP=4 -d output_splash_deffft/
../../configs/splash2/run.py -ESYSTEM=Detailed -EBENCHMARK=FFT -EROOTDIR=/hom\
e/fukumoto/m5_1.1/m5-test/splash2/splash2/codes/
warn: Entering event queue. Starting simulation...
##### m5stat.txt ######
cpu0.COM:IPC 0.012835 #
Committed instructions per cycle
cpu0.COM:count 154016099 #
Number of instructions committed
cpu0.FETCH:idle_cycles 11964858470 #
number of cycles where fetch stage was idle
cpu0.FETCH:idle_rate 99.71 #
percent of cycles fetch stage was idle
toL2Bus.data_idle_cycles 39429064 #
number of cycles bus was idle
toL2Bus.data_idle_fraction 0.003286 #
fraction of time data bus was idle
cpu1.COM:IPC 0.002342 #
Committed instructions per cycle
cpu2.COM:IPC 0.002343 #
Committed instructions per cycle
cpu3.COM:IPC 0.002344 #
Committed instructions per cycle
Thanks and Best regards,
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
