# Understanding when GC.gc() frees memory and doesn't
**URL:** https://discourse.julialang.org/t/understanding-when-gc-gc-frees-memory-and-doesnt/51682
**Category:** General Usage
**Created:** [December 11, 2020, 7:53pm UTC](https://discourse.julialang.org/t/understanding-when-gc-gc-frees-memory-and-doesnt/51682 "2020-12-11T19:53:57Z")
**Posts on this page:** 2
**Page:** 1
### Author:  [@tbenst](https://discourse.julialang.org/u/tbenst)
#### Post date: [December 11, 2020, 7:53pm UTC](https://discourse.julialang.org/t/understanding-when-gc-gc-frees-memory-and-doesnt/51682/1 "2020-12-11T19:53:57Z")
Consider the two code chunks below. When run in vscode, the first fails to free memory as reported by `htop`, while the second frees memory successfully. This is not just a reporting issue–if I create another large array in cell one, I can crash my computer by depleting free memory, while this does not happen in the case of two cells. Any ideas what is going on?
For a script, I cannot follow the two-cell strategy, so how can I force garbage collection?
```julia
## does not free memory
tseries = zeros(1024,1024,5,2000)
tseries = nothing
GC.gc()
sleep(5)
GC.gc()
tseries = zeros(1024,1024,5,2000) # crashes computer
```
```julia
## Two cells queued sequentially (no delay): does free memory
tseries = zeros(1024,1024,5,2000)
tseries = nothing
GC.gc()
##
GC.gc()
tseries = zeros(1024,1024,5,2000) # works fine
```
---
### Author:  [@paulmelis](https://discourse.julialang.org/u/paulmelis)
#### Post date: [December 14, 2020, 8:11am UTC](https://discourse.julialang.org/t/understanding-when-gc-gc-frees-memory-and-doesnt/51682/2 "2020-12-14T08:11:30Z")
Did you try the first version (single cell) from the Julia REPL, i.e. outside of VS? For me, with 1.5.3, the first `GC.gc()` call definitely reclaims the unused array, looking at the memory usage of my system.