Out of memory error

[celkins]

Robin,
I've just had what I think is my first bug report generating crash - an out of memory error, but there's some 40GB free in the system...
Entire log for the session is attached: I confess, I missed exactly what happened, since once both instances were live stacking successfully, I was sharing the previous image of the HorseHead & friends...
It was sitting with a "SharpCap has just crashed and will try to save your stack" message when I looked up.

Cheers,
Carl

[admin]

Hi Carl,

looks like the log may be incomplete as it cuts off without showing any sign of a crash report being processed. However, I can see at the beginning of the log that you have set your memory settings in SharpCap up for 8Gb for planetary/high speed cache and only 2Gb for Live stacking/Misc usage. I expect that what happened is that you hit the 2Gb limit at some point during the live stacking. Best thing to do is to turn that limit up in the settings to 4-6Gb since you are using a high res camera (particularly if you are using some of the more advanced image processing tools in live stacking).

cheers,

Robin

[admin]

Aha!

I think I have found the uploaded bug report associated with this issue. That contained more details and reveals that there is a memory leak when the colour calibration button is used. Once you have used colour calibration enough times, the leak means you hit this sort of out of memory issue.

I will fix the leak for the next update - in the meantime, increasing the limit as I mentioned will delay the problem significantly.

cheers,

Robin

[admin]

Update #2...

actually the memory leak only happens once per live stack reset/restart - it is related to the plate solve operation needed to locate which stars are which in the image, and that is only needed for the first Colour Calibration run of a particular stack, since once the data is calculated for a stack it can be re-used the next time you press the button.

cheers,

Robin

[celkins]

at least it has been of some use to you, Robin - that make it worthwhile posting

[mgCatskills]

I recently started using the 2600mc camera (much bigger images), and now am getting regular crashes with what seems to be a memory leak with the color calibration... Maybe half the crashes happen when I'm connected to the Internet, which means you got automatic reporting. The other half took place in star parties with no Internet.

I've read your suggestion re: memory allocation. What would you recommend for an 8GB miniPC (Mele 3C) running Windows 11 Pro?

[admin]

Hi,

firstly make sure you have the very latest version if you are using colour calibration in live stacking, as the older versions affected by the bug above were leaking one frame's worth of memory per stack if you used calibration - with a large image size from the camera that can build up fairly quickly!

If you are mostly doing deep sky live stacking then I would suggest putting the 'high speed frame cache' to the minimum and setting the 'live stacking and misc' as high as it will go.

Looking at my uploaded crashes, I have four that might be yours (right camera, right amount of memory), including two on October 4th (UK time). If those are yours then it would appear that you were using 4.1.13758, which is from before the memory leak bug fix above, so definitely would need to update if that is the case.

cheers,

Robin

[mgCatskills]

Thanks, Robin. Glad the memory leak bug has been swatted. I'm headed out to another star party tonight. Will make sure I update SharpCap prior to departure.