Seg fault when using julia (pakcage installation, profiling)

General Usage
, , ,
1

Hello,

I’m new to Julia so I don’t know if seg faults are a classic thing but anyway, I have one and I don’t know why.

I installed Julia using Juliaup

I’m on linux, so I just ran the command

curl -fsSL https://install.julialang.org | sh

While trying to understand somethings, I wanted to run the following code, which is a simple profiling code :

using Profile

function myfunc()
    A = rand(200, 200, 400)
    max_val = maximum(A)
    return max_val
end

println("function def ok") 

# First run to compile first 
myfunc()

println("first run ok") 

Profile.clear()

println("Profile clear ok") 

@profile for i in 1:5
    myfunc()
end

println("profiling ok") 

Profile.print()

When I tried to execute the code, by doing “julia code.jl” I got a seg fault.

Here’s the exact output :

function def ok
first run ok
Profile clear ok
Segmentation fault (core dumped)

As you can see, the execution reached the point where the profiling starts.

So I wanted to try another method for profiling and found a package called BenchmarkTools. When installing the package I also got a seg fault :

julia> import Pkg; Pkg.add("BenchmarkTools")
  Installing known registries into `~/.julia`
Segmentation fault (core dumped)

Even simple commands such as “pkg> st --manifest” give a seg fault :

(@v1.12) pkg> st --manifest
  Installing known registries into `~/.julia`
Segmentation fault (core dumped)

I’m using Julia version 1.12.1 (2025-10-17), installed using Juliaup
I’m on Ubuntu 24.04.2 LTS

Please tell me if you need more info, and thanks in advance for any help.

2

Your code can be run without any error on my linux machine. And I also installed julia via juliaup (the recommended way of installing).

Segfaults are not usual for me when writing daily code.

3

There is nothing normal about this.

It seems to me like a common factor for your segmentation faults are attempts to write files, possibly to /tmp. Can you successfully perform

julia> write("/tmp/test.txt", "test")

and

julia> write(expanduser("~/.julia/test.txt"), "test")

?

4

Hi,

Thanks for the replay.

When I executed the commands that you gave me, I got ‘4’ for each, which means the commands that you gave me worked 'since write function returns the number of bytes that have been written :

julia> write("/tmp/test.txt", "test")
4

julia> write(expanduser("~/.julia/test.txt"), "test")
4

Like you said, there’s nothing normal about this.

A colleague of mine suggested that it might be related to the C libraries that I have in my machine and the possibility that they are not compatible with what julia uses but everything else that I use that uses the C libs works fine so I don’t know.

5

Yes, that disproves my hypothesis. I also rather doubt that C libs should have anything to do with it. Ubuntu 24.04 is a very common platform to run Julia on.

You could try to run Julia in a debugger or with strace and see if that provides any hints.

6

Perhaps that is a hardware issue, either with your RAM or with your HD drive?

I’m not sure how to debug that, though. Are you overclocking your PC?

7

Can you run versioninfo(true) in Julia and send the result