3 level nested structs leading to out of memory

3 views (last 30 days)
I'm initializing nested structs as:
struct('nestedLevelOne', repmat({struct('nestedLevelTwo',repmat({zeros(1,1080)}, [1,540]))} , [1,100]));
When I access and store data to these structures this is leading memory issues with 192gb ram on my workstation. Is there I am doing wrong with this?
  2 Comments
Science Machine
Science Machine on 14 Apr 2022
Edited: Science Machine on 14 Apr 2022
@Akira Agata when i start reading and writing to the structures with parfor loops, then this quickly leads to memory issue. So i was thinking that was due to
(1) how the structure was initialized (you just checked if the structure 'could' be initialized as such, which is not my issue),
(2) or maybe the fact i'm using nested structures (perhaps another data structure type would be better for lots of data?

Sign in to comment.

Accepted Answer

Chunru
Chunru on 14 Apr 2022
It takes ~472MB memory as shown below.
a = struct('nestedLevelOne', ...
repmat({struct('nestedLevelTwo',repmat({zeros(1,1080)}, [1,540]))} , [1,100]));
whos
Name Size Bytes Class Attributes a 1x100 472192864 struct
a(1)
ans = struct with fields:
nestedLevelOne: [1×540 struct]
a(1).nestedLevelOne(1)
ans = struct with fields:
nestedLevelTwo: [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 … ]
size(a(1).nestedLevelOne(1).nestedLevelTwo)
ans = 1×2
1 1080
  1 Comment
Science Machine
Science Machine on 14 Apr 2022
Edited: Science Machine on 14 Apr 2022
I see, a structure i have is 47.69 gb, so I should save that piecemeal. Also, using parfor loops would create copies of this, which would explain why 20 workers crashed it

Sign in to comment.

More Answers (0)

Categories

Find more on Loops and Conditional Statements in Help Center and File Exchange

Tags

Products


Release

R2021b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!