The title pretty much. See this Github workflow for an example . 1.12 and 1.11 take 8 minutes, vs 1.10 at 3 minutes. Does anyone know why this is happening?

This is indeed a problem, and I’ve noticed it too. But I’m not sure what’s behind it.

People have been wondering this for a while, there is even a precompilation etc. benchmark , where timings can vary with setup enough that it’s not as simple as your observed regression at 1.11. For 1.11 specifically, people often bring up moving Array code from C into Julia and stdlib excision from …

I think in this package specifically is hurt more than most by the memory changes. I was taking a look and the __generic_kernel! that is the core computational kernels of this function (which were already huge functions) went from around 22k lines of IR in 1.10 to 55k in 1.12. This means that Julia …

One datapoint/tool here is the new julia CI timing tracking site. If you just look at building julia master for CI (various platforms, configurations) over the last year, there have been some temporary blips but it’s trended upwards +33% on average. Building julia !== precompiling packages, and diff…

Is there anything that could be done to simplify the amount of work for the compiler? Otherwise it sounds like the only fix is to somehow do even more caching so stuff doesn’t get recompiled so often, and disable world splitting so compiled code doesn’t get invalidated :thinking:

Not much outside of optimizing the compiler more and more (not just caching but making it faster overall)

World-splitting is only a fraction of what causes invalidations, and that optimization would be sorely missed in moderately type-unstable code e.g. comparisons returning Bool (though these cases could also be solved by a hypothetical function-wise return type annotation). More generally, letting a c…

This case here doesn’t have anything to do with world splitting or changes, it’s just the compiler getting slower to generate better code

Similar experience in ModelPredictiveControl.jl: 41 min on LTS 1 hour and a half on v1.11 and v1.12 for example here (edit: it was not executed on v1.11 in this job, but I had similar results on it, I can provide them if it’s needed). It clearly worsens the experience of interactive programming. …

Precompilation much slower in 1.11, 1.12 than in LTS (1.10)

General Usage

zweiglimmergneis January 22, 2026, 1:22pm 2

Topic		Replies	Views
Package precompilation logs/profiling General Usage question	2	71	December 12, 2024
Terminal interaction and precompile times General Usage	3	156	January 23, 2026
Regression in Precompile time with 1.10.0-rc2 compared to the rc1 version Internals & Design question	6	501	December 11, 2023
Julia 1.11 is slower than 1.10? Internals & Design regression	48	4215	October 29, 2024
Compiler Performance Internals & Design	20	2174	June 2, 2018

Precompilation much slower in 1.11, 1.12 than in LTS (1.10)

Related topics