Hi all,
I've been trying to measure the impacts of increasing the latency in some
functional units but found something odd in MinorCPU based cores.
I took as an example the ex5_LITTLE.py model (which is an A7-like core in
gem5 repo) and increased the latency of all functional units by 900 cycles
(just extrapolating). Then, I simulated the same binary (in System
Emulation) using the original and modified processor scripts. The final
tick count is exactly the same for both simulations. I have doubled checked
the config.ini files and the simulated core latencies are correct.
I understand that the MinorCPU models an in-order processor, which should
be influenced by the higher latencies. Am I missing something in the
MinorCPU concept or maybe in my configurations?

PS: I've repeated the same experiment using a DerivO3CPU and in this case
the tick count difference is expressive.

Regards,
Jeckson Dellagostin Souza
PhD Candidate
Instituto de Informática
Universidade Federal do Rio Grande do Sul
_______________________________________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users

Reply via email to