Master doesn't properly measure runtime, reports wrong simulation frequency

Open davidbiancolin opened this issue 8 years ago • 4 comments

For long running workloads we stuff like this:

==> spec-test/473.astar.test.err <==
SEED: 7282986
time elapsed: 18446744072974.2 s, simulation speed = 0.00 KHz
*** PASSED *** after 74901751853 cycles
Runs 74901751853 cycles
[PASS] MidasTop Test
SEED: 7282989

real    130m58.391s
user    0m0.237s
sys     0m0.526s

Feb 24 '17 18:02 davidbiancolin

Is that a 32-bit casting issue?

Feb 24 '17 22:02 ccelio

It certainly smells like it.

Feb 25 '17 00:02 davidbiancolin

I had to change the coremark source code (bad on me, I know) that handles printing out the timer... even though they provided define hooks to provide your own timer, they still fucking hardcoded the printf statement and got it wrong.

Feb 25 '17 00:02 ccelio

This is only a problem on zynq platforms; not F1.

Jan 31 '18 19:01 davidbiancolin