Preparing for new release of FC host tools
mychaela.falconia at gmail.com
Tue Feb 25 02:25:45 UTC 2020
> I have tried what you suggest and programmed a 3MB firmware
> image with and without the patch. I get the same result which is a duration of
> 2m32s at 812500 baud.
Hmm, interesting. I have just pushed a set of changes to the
freecalypso-tools repository, adding the following two new items:
* fc-loadtool now measures and reports the operation time on dump2bin,
dump2srec, flash dump2bin, flash dump2srec and flash program-bin
operations; fc-xram does likewise on its single operation.
* the new doc/Loadtools-performance article lists the times which I get
on my Slackware system and explains some of the underlying theory.
The operation times I get on my Slackware system are the same whether
I include or revert the ASYNC_LOW_LATENCY patch, at least with FT2232D
- I haven't yet tested CP2102 and PL2303 with the patch reverted. And
while actual measurement and reporting of these operation times is a
very new addition as of today, their subjective feel (how fast or slow
they feel to me) has not changed since the beginning of FreeCalypso
back in 2013, running Slackware 13.37 at the time.
Given that you run what I presume to be a "modern" host system and yet
you see no difference in flash program-bin execution time with or
without this ASYNC_LOW_LATENCY patch, I am now very tempted to revert
this patch and leave it out of the upcoming release, possibly to be
reintroduced at an indefinite later time if and when someone else
(someone other than Serg L) reports an actual system in the real world
on which fc-loadtool and/or fc-xram produces noticeably poorer
performance than what I have just documented in the new
Before I go ahead and revert Serg's patch, it would be great if you
(DS) could perform one more test on your presumably-modern system:
compile a ramimage.srec build of FC Magnetite hybrid for target
fcdev3b, and try loading it into your FCDEV3B (via FT2232D) with the
current version of fc-xram (the version with transfer time reporting
added), with and without the ASYNC_LOW_LATENCY patch. On my Slackware
system this operation executes in about 2m54s.
More information about the Community