Supressing output of mfm_emu

[hupfadekroua]

Dear all,

since I'm getting timeouts while trying to boot a PDP11/53 which kind of options are available to speedup teh response time of mf_emu?

From my knowlegde any output to screen or terminal is consuming a lot of time.

Is there a possibility to supress the informative output of mfm_emu completely?

Andreas

[David Gesswein]

On Sun, Mar 05, 2023 at 10:21:48AM -0800, hupfadekroua wrote:
>
> since I'm getting timeouts while trying to boot a PDP11/53 which kind of
> options are available to speedup teh response time of mf_emu?
>

Need to understand more on what the timeout problem is. If the issue is due to
the slow flash writes and lots of writing --pool can be used to try to hide it.
Better solution for this is fast USB stick or uSD card. Needs to be fast
for 4k random writes.

To turn off the messages you can change this line in
emu/parse_cmdline.c
// Enable all errors other than debug
msg_set_err_mask(~0 ^ MSG_DEBUG);
to
// Enable all errors other than debug and info
msg_set_err_mask(~0 ^ (MSG_DEBUG | MSG_INFO));
to turn off all the info messages also.
Run make to rebuild

> >From my knowlegde any output to screen or terminal is consuming a lot of
> time.
>

Don't think they take that much time. I also put a change in a while ago to
make writing the messages non blocking so they will be discarded if outputting
them gets behind.

Let us know if something helps.

[hupfadekroua]

The emu file is stored on the emmc up to now.

Would a uSD card be faster, e.g. using an industrial one or a SanDisk Extreme?

Best

[David Gesswein]

On Sun, Mar 05, 2023 at 12:46:45PM -0800, hupfadekroua wrote:
> The emu file is stored on the emmc up to now.
>
> Would a uSD card be faster, e.g. using an industrial one or a SanDisk
> Extreme?
>

Probably a good quality one will be faster. Some of the cheap ones are very
slow for random write. I think emulator writes are around 16k since it writes
entire track.

The eMMC is plenty fast for reads. Its only slow for writes.

Look at the emulator prints when you get the timeout.
This message will show if your getting behind writing.
Free buffers 74,0 delay 0.000

If free buffers is much lower than normal then the intentional delay will
get large. You can also look at the seek times around the error.

If free buffers hasn't changed much changing the storage is unlikely to help.

Did a little testing. Noticed that seek time went to 300ms for one seek.
This command will give emulator priority. With it max seek time was 4.9ms.

chrt -f 40 ./mfm_emu rest of options