Re: [hatari-devel] Very slow emulation when enabling Cycle Exact

[Jerome Vernet <vernet.jerome@xxxxxxxxxx>]

Hi,

Le 13/02/2018 à 22:34, Eero Tamminen a écrit :

During bootup, CPU utilization was somewhat higher.

I profiled it with Valgrind Callgrind tool.  Attached is
a callgraph of where most of the PC CPU *instructions* are spent
according to it.  Cache prefill emulation seems to cost a lot.

Yeah, that's what I can see here: fill_prefetch_030, fill_prefetch_030_ntx are using most of the CPU (about 70 %).

  

There are 2 Core used, both about 50%, so things can be improved.

In Linux you can bind process to a single core with:
    taskset 0x1 <program>

Unfortunatly, it's impossible on MacOsX.

On MacOS, SDL has been traditionally a performance hog.
While that improved with v2, maybe it's still somewhat an issue.

I can't see any SDL function in the profiler, nothing, so SDL is now for negligeable part in performances. BUT i found something insteresting: the second Thread at 50% is not Hatari, but an Apple thread between Hatari Audio part and MacOsX, called com.apple.audio.IOTHreadClient. This thread use 50% of CPU, even if there is no sound playing.
In any other mode (ST/TT/...), this thread use less than 1% of CPU...

    - Eero