Re: [Gems-users] fastpath option in Ruby

[Dan Gibson <degibson@xxxxxxxx>]

Did you issue
instruction-fetch-mode instruction-fetch-trace
istc-disable
dstc-disable
cpu-switch-time 1

to Simics before loading Ruby?

Also, the TSO flag hasn't been used by us for some time... it may not be 
working properly. Does the same thing happen with TSO off?
Regards,
Dan

salil pant wrote:
> Hi, I noticed that turning fastpath on for RUBY with FFT , actually
> increases execution time for my experiments.
> Time without fast path - 2762532
> Time with fast path on - 3419833
> I am not able to figure out the reason for this.
> I am using the MOSI_SMP_bcast config with TSO. with 4 processors.
> There is some latency for accessing L1s during hits which the fastpath
> gets rid of.
> I notice no difference in the statistics output by ruby except a
> slight difference in the number of L1 Dcache misses.
> Misses without Fastpath - 4392
> Misses with fastpath  - 4666
> Although average miss latencies are still very similar
> I dont know if the increased misses can be a reason for this
> performance difference ?
> I cant think of a scenario where misses would increase because of the
> fastpath option.
>  Any ideas ? or is something wrong with my setup ?
>
> salil
> _______________________________________________
> Gems-users mailing list
> Gems-users@xxxxxxxxxxx
> https://lists.cs.wisc.edu/mailman/listinfo/gems-users
> Use Google to search the GEMS Users mailing list by adding "site:https://lists.cs.wisc.edu/archive/gems-users/"; to your search.
>
>
>