I am using superlab on a macbook, running system 10.4.9, and it keeps freezing in the middle of my experiment and losing all the data. Has anyone encountered this problem (or fixed it) before?
We’re in the middle of a large run, and after about 130-140 Ss, this has indeed happened to us a few times. Also, Superlab has “quit unexpectedly” after writing the data, or has died during startup, but in both of the latter cases, we didn’t actually lose any data. We’re doing this testing on three Macbook Pros.
We suspect temperature may be a factor, although we’ve done some testing trying to crash superlab deliberately by heating up a room without success. I just installed a program called “fan control” on our macbookpros: it does temperature and fan speed monitoring, and can also be used to boost the fan speed a bit if that turns out to be necessary. Next week, we’ll see if that provides any useful information.
Was the problem happening on every run for you? I’m not able to get past about 200 trials (and my experiment involves almost 1500!) without it quitting every time.
Could you both send me your crash logs? The operating system automatically places them in your home directory under Library/Logs/CrashReporter. The SuperLab file should be named “SuperLab 4.crash.log.”
The only explicit crash issue that comes to mind is one relating to presenting an event from a Macro. If the event needs a stimulus file, SuperLab crashes in 4.0.2 (unless the file was loaded for some other reason). Other than that, the obscure crashes are limited to 10.3.9.
The crash logs don’t exist on my computer. Possibly because Superlab doesn’t actually crash - it just freezes, starts using over 99% of CPU resources, and I have to use forcequit to do anything. I haven’t been presenting events from macros when the problems happen.
What I did notice is that when I reduced the size of my files (from .psd’s to jpegs) the program survived longer - from about 200 trials to about 300 trials. Could this be related to it?
It’s possible that this is relevant. If it is, it would imply that your issue might be memory related (i.e. not enough RAM).
When SuperLab stops responding, what does it do if you let it sit for a little bit (i.e. maybe five minutes)? Does it keep chewing up processor time, or does it eventually crash?
I’ve seen this
I have seen something very much like what you have been seeing, Rachel, but only a few times, and I don’t think I get a crash report for them either. The more frequent crashes I’ve seen do generate a crash report (and I’ve sent them all to Hank). We have 528 trials in the current experiment, it takes 30-40 minutes, and like I say, in maybe 98% or so of the runs, the experiment has gotten to the end without hanging (but after that, the program fairly often “quits unexpectedly”).
Our RAs report that the macbookpros get pretty hot sometimes, do you see the same thing?
Today we’ve started running the next phase of the study and now all the computers have the fan control software installed and operational, so it’ll be interesting to see if that has any effect, and also whether the internal temperature rises too much (the fan control program also monitors the temperature).
I’ve recently identified (and fixed) an issue that causes SuperLab to stop responding 71 minutes into an experiment, but this issue is very strictly limited to the Windows implementation. I’ve successfully left an experiment running on one of my Macs for five hours. This machine was left unattended, so there was no participant input, and the experiment was very small.
If either of you have an experiment that causes this crash that is small enough to e-mail, I would appreciate being able to run it. Alternatively, if you are familiar with Shark (a part of CHUD) and can get me a Time Profile of this 100% processor usage, that would probably also be useful in at least figuring out where it’s occurring.
Greg, are you using any event-based input devices for the experiments that are crashing after the experiment is done? After having perused through every one of your crash logs, I’m beginning to have a vague suspicion that something like String Input or Mouse Input could be related.
No, the only input I’m using is single keyboard events (via the X-Keys USB switch interface). Also, I’m attaching one of the experiment files from the set we’ve been running. There are 400 of them and each one is different, and I’m not 100% sure which ones have been associated with crashes, but the one I’m sending is typical. (They were all produced using sl4util.)
s14a0.zip (24.9 KB)
Other than the fact that it about put me to sleep, I haven’t yet had any problems with this experiment. On Monday, I’ll have to run it again under 4.0.2 to verify.
Hank, as I mentioned a while back, I installed a fan control program (http://www.lobotomo.com/products/FanControl/) last weekend. Well, according to the RAs, there have been no crashes this week. There was one strange case where Superlab seemed to “pause”, but after hitting escape and then resuming, the experiment completed normally (this particular bogosity never happened previously). Before last week, there were several crashes per day.
I’m hoping that there is something strange going on with MacBook Pros and their fans that may have been causing overheating, that is corrected by the fan control program. I think I also mentioned that the RAs complained about how the machines were very warm at the end of a session; now they seem subjectively to seem much cooler.
If this continues through next week, then I think I’ll be convinced that heat is at least part of the problem.
Sorry that the experiment is soporific, that’s not its main purpose.
Hank, I’m emailing a file of my experiment. I’ve been experimenting on different computers with huge amounts of RAM, and it seems to be causing the same problem no matter what. I’ve also left it sitting for over an hour, and it doesn’t crash, it just sits there … watching me.
Greg, I haven’t noticed any problems with temperature, although I’m using plain old Macbooks, not the pro’s. However, if it seems to help I might search out an equivalent program to test it out - thanks for the advice.
It was late, I was tired, ready to go home… and I my eyes were wandering… I also wasn’t getting up and taking the breaks that your experiment recommended. Don’t blame yourself. :o
Just wondering if you got the email and if it’s all going OK? Do you have any idea how long this problem may take to resolve? I was hoping to have all my pilot testing done by now, so this is seriously impacting on my thesis timeline.
Yes, I did receive your experiment, and it ran fine for me (as did Greg’s).
We expect to have the SuperLab 4.0.3 update out shortly (i.e. a few hours), so I’ll e-mail you the link as soon as it’s available.
I apologize for the delay in responding to you.