Julia v1.6.0 has been released

ararslan · March 25, 2021, 3:27pm

The Julia developers are pleased to announce the release of Julia v1.6.0, the sixth minor release in the 1.x series. Some folks put together a blog post highlighting some of the most exciting new features in 1.6. Check it out!

As usual, binaries are available for all of your favorite platforms (Linux, macOS, Windows, and FreeBSD) at https://julialang.org/downloads.

As a minor release, v1.6.0 contains no breaking changes, only new features, performance improvements, and marginal, undisruptive changes in behavior. You can also see the NEWS file for the full set of changes.

It is currently undecided whether 1.6 will have long term support; that determination will come after 1.6 has seen more use. Julia 1.0 is currently still the only long term support version. Note that as of this release, 1.5 has been effectively superseded by 1.6, which means that there will not likely be any further 1.5.x releases.

We encourage everyone to give it a try. Packages can test with 1.6.0 on CI by specifying 1.6 on Travis, AppVeyor, Cirrus , and soon GitHub Actions. As always, let us know in the issue tracker if you run into any issues.

Enjoy!

affans · March 25, 2021, 4:02pm

Last time I was this excited was the 1.0 release. Great job everyone. Downloading it right now!

rdeits · March 25, 2021, 4:04pm

FYI, it looks like plotly isn’t happy with the embedded graph in that post:

Edit: It’s fixed! Thanks for the great blog post!

mbauman · March 25, 2021, 4:06pm

Already fixed and slowly coming down the tubes.

TheLateKronos · March 25, 2021, 4:32pm

Read the blog-post, and the changes look pretty amazing. Thank you to everyone who contributed!

I am going to try and install julia using the chocolatey package-manager for the first time. Do anyone know how long it usually takes for the julia-package over at chocolatey to get updated?

rashid · March 25, 2021, 4:34pm

I am not sure whether I should ask this question here, but anyway.

It is mentioned in the blog post that precompilation is…

…multi-processed, as opposed to multi-threaded, so is not limited by Julia’s thread count. By default, Julia will spawn a maximum CPU-balanced load of package precompile jobs at once based on the number of CPU cores

Is there any way to limit the number of cpu cores being used for this. Asking because when I was installing a package (using julia-1.6_rc3, in my low end system), the precompilation package completely hanged the system for sometime.

kristoffer.carlsson · March 25, 2021, 4:38pm

There is the JULIA_NUM_PRECOMPILE_TASKS env variable that you can set. Unfortunately, it doesn’t seem to be documented right now.

rashid · March 25, 2021, 4:44pm

Thanks. Will try that.

marllos · March 25, 2021, 4:46pm

Excellent job! Julia is very good. I already updated.

jishnub · March 25, 2021, 5:00pm

I am facing issues using this with gcc 4.9.3 which comes pre-installed on our HPC cluster.

(base) jb6888:~/ $ module load gcc
Loading module 'gcc/4.9.3'

(base) jb6888:~/ $ julia
ERROR: Unable to load dependent library [...]/julia-1.6.0/bin/../lib/julia/libjulia-internal.so.1
Message:[...]/gcc/4.9.3/el7/lib/gcc/x86_64-unknown-linux-gnu/4.9.3/libstdc++.so.6: version `CXXABI_1.3.9' not found (required by [...]/julia-1.6.0/bin/../lib/julia/libjulia-internal.so.1)

In all likelihood this is because of the paths that gets prepended to LD_LIBRARY_PATH.
I wonder if I might require a different gcc version to fix this issue?

giordano · March 25, 2021, 5:06pm

Very likely

You don’t need GCC at all to use official binaries of Julia (only libc), they come with the libraries of GCC 9. Your system libraries are overriding them

jishnub · March 25, 2021, 5:10pm

Prepending the julia lib directory to LD_LIBRARY_PATH fixes this for the time being

export LD_LIBRARY_PATH=[...]/julia-1.6.0/lib/julia:$LD_LIBRARY_PATH

Reading through a similar issue posted last year it might be an issue with HPC clusters using dated CentOS

brianguenter · March 25, 2021, 7:14pm

Fantastic job! Kudos to all the contributors to 1.6.

Amazingly fast precompilation. We were using precompiled sysimages in 1.5.x because otherwise load times were unbearably slow.

1.6 is so fast this no longer seems necessary.

BLI · March 25, 2021, 7:17pm

I deleted all of past Julia, and installed from scratch.

Very snappy installation! And actually entertaining to listen to my little Lenovo media PC start panting while using all the i9 cores. Now: installing it on my laptop.

Running code also seems much quicker.

Thanks for you hard work!!

mbauman · March 29, 2021, 2:16pm

3 posts were split to a new topic: Pkg.add updates other packages too?

StatisticalMouse · March 25, 2021, 8:55pm

@kristoffer.carlsson “For Julia release 0.6, formatting of stacktraces…”

lostella · March 25, 2021, 8:58pm

Fantastic job, thank you everyone

Minor: the link to the PDF in the documentation appears to be broken: https://raw.githubusercontent.com/JuliaLang/docs.julialang.org/assets/julia-1.6.0.pdf

Edit: works now

mbauman · March 25, 2021, 9:04pm

That’s actually intended! Julia 0.6 was the last time stack traces were significantly changed:

See the fifth bullet point here: https://github.com/JuliaLang/julia/blob/master/HISTORY.md#library-improvements-1 (and search elsewhere in the history for stacks).

kristoffer.carlsson · March 25, 2021, 9:20pm

I should probably have worded that differently though, many people have had the same confusion.

Edit: tweak wording to not make it look like a typo by KristofferC · Pull Request #1242 · JuliaLang/www.julialang.org · GitHub should improve the wording.

FHell · March 26, 2021, 9:26am

Just gave it a try. Early days but wow! The precompilation changes, and the compilation speed improvements are absolutely wonderful. This is really a massive quality of life improvement. Excellent job!