On Fri, 15 Jan 2021, Filipe Costa wrote:
 Hi all,
 I have a script which deals with a large data sample of stock 
 prices. In general it takes some time to run (30m - 60m). I'm on a 
 linux machine. When using Gretl 2020d it crashes because it eats 
 my 8GB RAM plus swap space. But, when using Gretl 2019d it 
 completes the work relatively fast and doesn't crash as it doesn't 
 eat all memory. Any thoughts or ideas on what may be behind this? 
 I tried with different linux flavours and I'm pretty sure it has 
 to do with Gretl version and nothing else. 
If it is to do with gretl, it sounds like a "memory leak" -- 
progressively allocating storage without freeing unused memory.
I'm not aware of any changes that could have introduced a leak 
between gretl 2019d and 2020d but I guess it's possible.
Can you try gretl 2020e, or the current snapshot (2020f-git) and see 
what happens? That would be very helpful.
Another thought: if you could send me offlist the script you're 
running and some sample data, then I can use valgrind to see if 
there are any leaks.
Allin