[¿fixed?] Serialosc Mountain Lion Compatibility?

  • I retired my 24ch A&H a few years back because of repeated instances of possession. I would often be working away at the other side of the room from it and the faders would move on their own. I wasn't sending any control messages to it... In fact I didn't even have MIDI cables hooked up most of the time.

    Other times, I'd turn it on and no sound would pass, even though the bridge was happily dancing away. It needed to be power-cycled to fix it but it sometimes took a half dozen tries.

    And, finally, it would mute and unmute single channels of it's own accord.

    I thought I had a lemon, or that it somehow took offense to the sounds I was pumping through it, but I've come across several other folks that had similar problems with different models in the same (vague) price range.

    Shame. They sound real nice.

  • Ok just noticed that it was pushing 100% CPU on my iMac again today. Doesn't seem to happen on my laptop (at least not yet).

  • > unplug my device, and all of a sudden one of the 2 remaining serialoscd processes shoots to 99% and stays there.

    okay, cool, i can work with that.

    the code related to this problem is really messy and honestly i'm not surprised we're seeing problems in it. i'll try and reproduce this.

  • Another bit of info, I had killed serialoscd a few hours ago (I guess I only killed the one) but checking now I had serialoscd (I'm guessing the remaining one) running at around 60% CPU. So I manually quit that one too.

  • Thought I'd check this thread to see how the progress of updating SerialOSC is going. A bit disheartening to hear that so many of you gents are having problems. Hopefully it gets sorted soon, particularly for those that are having issues on performance machines.

    I'm sticking it out with entering "Serialosc" in a open Terminal window.
    It's been working perfectly so far.

    Seeing as more people will eventually be upgrading to ML perhaps we should make this thread sticky?

  • i've seen the 99% cpu come up in earlier versions of serialosc, typically right before an FTDI crash.

  • I haven't had any FTDI crashes since I was on Snow Leopard. hope it stays that way.

  • noticed a cpu spike tonight for the first time. stopped patching, unplugged the monome and noticed fan wouldn't calm down, checked the activity monitor and one of the two instances of serialoscd was running between 65 - 80%.

    is there any difference between closing serialosc in the activity monitor and doing it from terminal?

  • i am having an absurd time trying to reproduce this CPU spike.

    myr, did the spike happen spontaneously, or did you unplug your monome first?

  • might have a lead here.

    were any of you guys using serialosc with more than one monome?

  • i didn't notice the spike for quite sometime, at least half an hour since i'd stopped programming and unplugged the monome. i then noticed the battery was running down pretty quickly and the fan was still making a bit of noise.

    if i remember correctly the fan had been making noise since when i was programming with the monome plugged in (had a lot of stuff open so didn't necessarily spike there), and continued after i'd stopped and unplugged monome. so i don't think it was totally spontaneous.

    i was using my spring 2012 128 this time, have used two recently with serialosc tho.

    doing a bit more work programming work today so i'll see if i can get it to reproduce.

  • I only have one monome (but also one arc). When I tested to see if serialosc was working I'm pretty sure I only plugged in my walnut64

  • i've found a bug that certainly could be the case. my suspicion is that i'm clobbering part of a data structure somewhere and it's causing one of my functions to go "what the hell, this is bullshit" and enter an infinite loop.

    i've found a bug that fits the description, but i need to test further. still, i'm on the trail of this one.

  • Good news indeed.

    I just noticed it going crazy on my laptop. I've not connected/disconnected a monome to it (for weeks, or actually, since upgrading to ML if I remember correctly).

  • Since updating to ML I've not plugged a monome into my laptop. I might have when I first did it a few weeks ago, but I don't think I did.

    I just noticed the fan going crazy even though I had quit Max.

    I should add that I was working in Max with serialoscs getting loaded (the subpatchers) and loadbanged, so maybe that triggered the bug.

  • k, think i've found at least a potential cause of the CPU pegging bug. here's a new build, see if this fixes it for you guys:


  • Will give it a spin. Got a gig tomorrow too so this is good timing.

  • Well, I haven't updated the latest serialosc yet (thanks btw visinin) but can confirm a crash when I unplugged my gs 128. My new MacBook Pro crashed completely requiring a restart. The error log showed a FTDI issue.
    Will update to the latest now.

  • ftdi issue is known and is a problem on their end. we've been in communication with them, and they have a monome to test with, but it's a problem we can't fix in serialosc.

  • So this may happen again? Hmmm - anyway to avoid such a crash? I never had this happen when unplugging a monome before. I can't remember if I had max running in the background or not...

  • if you really really really need to unplug a monome without it crashing, the best way to do it is to log out, unplug, and then log back in.

  • i just discovered it recently.
    serialosc runs per-user when you log in! which is dope.

  • I need to get home and troubleshoot to see Its something else but just did a gig and all my serialosc stuff was laaaagy as shit, including tons missed/double presses. Serialosc wasn't spiking the CPU but max was showing at 121% CPU in activity monitor. I have made sone other changes that I need to test first but figured I'd throw this in here in case someone noticed somethinf similar.

  • probably a max issue, man, especially if it was thrashing your CPU like that.

    what apps were you using?

  • My own (the party van). Max was only showing about 30% although activity monitor and max rarely agree. I have noticed general max/message stuff lag recently so it could be related to that. I need to sit down and delete bits to see of I can find out what's up. It almost feels like overdrive isn't turned on (though it is).

    (the main new things are a serial object (not being banged) and 'hints' on all GUI elements).

  • It just occurred to me (actually read something about it on the Max forum) that this might be related to javascript. In my last big update, I ditched OSC-route for a js based version. But from what I read js can get deprioritized and dumped onto a lower priority thread. It may also explain higher CPU usage(?).

    That would be a bummer as I was happy to live without OSC-route (trying to go external free, rather, trying to use only Alex Harker externals).

  • On Max6. I do have lots of instances of the js osc-route, any lag/overhead at all would be a compounded problem.

    Just got home from gig now so gonna take things apart to see what's up. (I might actually start with all the OSC objects).

    I wish it was a native object already, or if there was a simple workaround.

  • Yup, just tested it on a second computer and activity monitor is taking a shit with CPU usage (120%).

    Gonna try a regexp osc-route replacement to see if that fixes the problem.


  • Raja, I've got 10.7.4 running with the latest serialosc version (from this thread) with no probs... It seems it might be deeper than versioning to me. Could it be hardware specific?

    And yeah, Romney for president: I haven't slept in weeks.

  • what a total cluster fuck this has become... thanks to everyone who is continuing to test this stuff out so far. i'm on osx 10.8.1, and have serialosc version version c installed right now. i just don't unplug my 256 and haven't seen any problems with serialosc yet (knock on wood). i'm afraid to even try and uninstall the version i have and install another one, since that will probably fuck everything up. so glad i have a mountain lion partition on an external hard drive with everything that's installed on my computer now, minus serialosc and the ftdi driver. if worse comes to worse, i can just wipe my computer out and throw that partition right back on.

    at least raja is giving us some good laughs right?

  • > wanted to see if it would rewrite the org.monome.serialosc folder into my ~/Library/Preferences folder if i deleted it before installing(because this was a problem with an older Windows XP computer i had too...),
    it didn't rewrite the preferences folder... but worse yet, i get an error that reads: "The Application launchd quit unexpectedly"


    whoa whoa whoa, slow down. did you just delete the ~/Library/Preferences/org.monome.serialosc folder? did you do it //while serialosc was stopped//?

    serialosc creates that folder when it starts, it's not created in the installer at all. theoretically everything should work fine.

    it looks like you may have broke launchd. that's not good, and it's not something i have any experience with.

    //what did you do?//

  • @Raja
    I'm on MacBook Pro 8,1, 2.7GHz i7, 4GB RAM. Never had a monome related crash. I don't normally log out or shut down unless I have a software update that requires it.

  • hey, is anybody still experiencing the high CPU usage bug with 1.1d?

  • NNNNG. boy are kernel panics a pain in the fucking hole.

    not sure is this the right place, but here's another report.

    doing nothing special, few patches open. running on battery, unplugged nome.

  • lol. yeah i was a bit mortified about my visit count. fuck it.