Simulation results differ with Simulation.md

[msaideroglu]

When I try example simulation with command which is described as in simulation.md:
./ci/blackbox.sh --clusters=1 --cores=4 --warps=4 --threads=4 --driver=opae --app=sgemm gives
Verify result PASSED! PERF: core0: instrs=75388, cycles=93187, IPC=0.808997 PERF: core1: instrs=75388, cycles=93189, IPC=0.808980 PERF: core2: instrs=75388, cycles=93188, IPC=0.808988 PERF: core3: instrs=75387, cycles=93347, IPC=0.807600 PERF: instrs=301551, cycles=93347, IPC=3.230431

However expected output which is shown in simulation.md is:
Verify result PASSED! PERF: core0: instrs=90802, cycles=52776, IPC=1.720517 PERF: core1: instrs=90693, cycles=53108, IPC=1.707709 PERF: core2: instrs=90849, cycles=53107, IPC=1.710678 PERF: core3: instrs=90836, cycles=50347, IPC=1.804199 PERF: instrs=363180, cycles=53108, IPC=6.838518

Expected performance is at least doubled with same config parameters. Is there any missing detail here?

[kmzaja]

My results also differ:

PERF: core0: instrs=112263, cycles=149410, IPC=0.751375
PERF: core1: instrs=112264, cycles=151896, IPC=0.739085
PERF: core2: instrs=112264, cycles=146323, IPC=0.767234
PERF: core3: instrs=112264, cycles=150096, IPC=0.747948
PERF: instrs=449055, cycles=151896, IPC=2.956332

I am also looking for the answer.

There is one more thing:
When I use --driver=opae, in the start I get overflowed values and then my cores run instructions one by one. The output results look something like this:

Are you also facing this problem?

[kmzaja]

For the --driver=opae prints, reason and workaround it can be seen in #127.