Darwinbots Forum

Code center => Suggestions => Topic started by: Testlund on November 07, 2014, 02:29:04 PM

Title: The previous autosave feature was better.
Post by: Testlund on November 07, 2014, 02:29:04 PM
I think it was better when you could set your own save intervals and have multiple saves. Now there is only one autosave that is used and if that one is corrupted (which always happens when I get BSOD) then the whole sim is lost.

It's also less taxing on the hard drive to not save every 2000 cycles as it is now. Every writing to the hard drive shortens it's life span.
Title: Re: The previous autosave feature was better.
Post by: spike43884 on November 08, 2014, 08:17:39 AM
Well, I wouldn't keep all of the saves, but keep the last 3 saves...
Mhm, writing it to the harddisk....Mhm, I dont ever get BSOD so maybe have the option then the save the last save that was kept of that sim to the harddisk, rest to RAM and keep any manual saves on hard disk.
Title: Re: The previous autosave feature was better.
Post by: Testlund on November 09, 2014, 07:25:10 AM
It's my laptop that is beginning to fail I think. I can have a sim running fine for a week and suddenly I get BSOD.

Because DB autosaves so often, and to the same file, I think the risk is high that it tries to save just when a hardware error happens, which makes the save corrupt.

So now I have to try to remember to do manual saves several times a day.
Title: Re: The previous autosave feature was better.
Post by: spike43884 on November 09, 2014, 11:01:04 AM
Testlund, you dont happen to be a expert in DB2 code?
Title: Re: The previous autosave feature was better.
Post by: Testlund on November 10, 2014, 04:22:45 AM
Not at all.  :P
Title: Re: The previous autosave feature was better.
Post by: spike43884 on November 10, 2014, 11:28:17 AM
ahh, I just came up with a new fancy-pancy system to get bots to perform tasks, with the existing DB code.
Instead of 1 gene per task, its a lot of genes, and each perform little useless things alone, but depending on exact conditions etc., it will perform various different tasks...so you might have 30 genes active to do a task, and another 40 inactive as the conditions aren't perfect...instead of having 1 gene to do each possible full outcome.