Julia can hang when using threading, on 1.7 beta2 and 1.8 #41407
Description
A.
$ time ~/julia-1.8-DEV-7553ca13cc/bin/julia --startup-file=no -t4 -O2 fasta.jl 25000000 > /dev/null
real 0m3,032s
user 0m8,379s
sys 0m0,883s
then immediately after:
$ time sh -c ~/julia-1.8-DEV-7553ca13cc/bin/julia --startup-file=no -t4 -O2 fasta.jl 25000000 > /dev/null
^C
^C
^C
Terminated
real 4m25,588s
user 0m0,004s
sys 0m0,000s
Note, CTRL-C didn't work to kill, I had to use the kill command (did in another terminal tab). I previously ran into hangs with hyperfine, which I do not believe implicated, it does the latter, but wqith it I could kill with CTRL-C, meaning hyperfine got killed but julia was still running in the background.
I would mark this as a regression. I've done plenty of benchmarking before, often with hyperfine for 1.6, or time (or just @btime), and this never happened with hyperfine. I'm however new to using only "sh -c" (which is implied by hyperfine). See at #39598 (comment)
This affect 1.7.0-beta2 too (tested later):
$ time sh -c ~/julia-1.7-DEV-da96fef327/bin/julia --startup-file=no -t4 -O2 fasta.jl 25000000 > /dev/null
Terminated
real 1m31,985s
user 0m0,001s
sys 0m0,003s
and strangely doing "nothing":
Note, I believe I pressed ENTER despite of ERROR on the same line. At least it was quick:
$ ~/julia-1.7-DEV-da96fef327/bin/juliaERROR: TaskFailedException
Stacktrace:
[1] try_yieldto(undo::typeof(Base.ensure_rescheduled))
@ Base ./task.jl:705
[2] wait
@ ./task.jl:764 [inlined]
[3] wait(c::Base.GenericCondition{ReentrantLock})
@ Base ./condition.jl:113
[4] take_buffered(c::Channel{Any})
@ Base ./channels.jl:389
[5] take!
@ ./channels.jl:383 [inlined]
[6] repl_backend_loop(backend::REPL.REPLBackend)
@ REPL /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.7/REPL/src/REPL.jl:195
[7] start_repl_backend(backend::REPL.REPLBackend, consumer::Any)
@ REPL /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.7/REPL/src/REPL.jl:185
[8] run_repl(repl::REPL.AbstractREPL, consumer::Any; backend_on_current_task::Bool)
@ REPL /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.7/REPL/src/REPL.jl:317
[9] run_repl(repl::REPL.AbstractREPL, consumer::Any)
@ REPL /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.7/REPL/src/REPL.jl:305
[10] (::Base.var"#914#916"{Bool, Bool, Bool})(REPL::Module)
@ Base ./client.jl:394
[11] #invokelatest#2
@ ./essentials.jl:721 [inlined]
[12] invokelatest
@ ./essentials.jl:719 [inlined]
[13] run_main_repl(interactive::Bool, quiet::Bool, banner::Bool, history_file::Bool, color_set::Bool)
@ Base ./client.jl:379
[14] exec_options(opts::Base.JLOptions)
@ Base ./client.jl:309
[15] _start()
@ Base ./client.jl:492
nested task error: IOError: stream is closed or unusable
Stacktrace:
[1] check_open(x::Base.TTY)
@ Base ./stream.jl:386
[2] raw!(t::REPL.Terminals.TTYTerminal, raw::Bool)
@ REPL.Terminals /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.7/REPL/src/Terminals.jl:138
[3] prompt!(term::REPL.Terminals.TextTerminal, prompt::REPL.LineEdit.ModalInterface, s::REPL.LineEdit.MIState)
@ REPL.LineEdit /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.7/REPL/src/LineEdit.jl:2602
[4] run_interface(terminal::REPL.Terminals.TextTerminal, m::REPL.LineEdit.ModalInterface, s::REPL.LineEdit.MIState)
@ REPL.LineEdit /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.7/REPL/src/LineEdit.jl:2481
[5] run_frontend(repl::REPL.LineEditREPL, backend::REPL.REPLBackendRef)
@ REPL /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.7/REPL/src/REPL.jl:1126
[6] (::REPL.var"#44#49"{REPL.LineEditREPL, REPL.REPLBackendRef})()
@ REPL ./task.jl:406
caused by: IOError: read: i/o error (EIO)
Stacktrace:
[1] wait_readnb(x::Base.TTY, nb::Int64)
@ Base ./stream.jl:408
[2] eof(s::Base.TTY)
@ Base ./stream.jl:106
[3] eof(io::REPL.Terminals.TTYTerminal)
@ Base ./io.jl:416
[4] match_input(k::Dict{Char, V} where V, s::Union{Nothing, REPL.LineEdit.MIState}, term::Union{REPL.Terminals.AbstractTerminal, IOBuffer}, cs::Vector{Char}, keymap::Dict{Char, V} where V) (repeats 4 times)
@ REPL.LineEdit /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.7/REPL/src/LineEdit.jl:1475
[5] prompt!(term::REPL.Terminals.TextTerminal, prompt::REPL.LineEdit.ModalInterface, s::REPL.LineEdit.MIState)
@ REPL.LineEdit /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.7/REPL/src/LineEdit.jl:2571
[6] run_interface(terminal::REPL.Terminals.TextTerminal, m::REPL.LineEdit.ModalInterface, s::REPL.LineEdit.MIState)
@ REPL.LineEdit /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.7/REPL/src/LineEdit.jl:2481
[7] run_frontend(repl::REPL.LineEditREPL, backend::REPL.REPLBackendRef)
@ REPL /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.7/REPL/src/REPL.jl:1126
[8] (::REPL.var"#44#49"{REPL.LineEditREPL, REPL.REPLBackendRef})()
then nothing for a while until I pressed ENTER (again?!) and got:
_ _ _(_)_ | Documentation: https://docs.julialang.org
(_) | (_) (_) |
_ _ _| |_ __ _ | Type "?" for help, "]?" for Pkg help.
| | | | | | |/ _` | |
| | |_| | | | (_| | | Version 1.7.0-DEV.714 (2021-03-16)
_/ |\__'_|_|_|\__'_| | Commit da96fef327 (105 days old master)
This is getting strange, running that julia version straight after no problem. It's either only random and/or my machine getting strange. Any idea what could cause it? Everything else seems to work fine. I do not recall if I was using that exact git commit of 1.7 before.
I was probably using all my three 1.7 downloads frequently before, not sure which most, and running all of them once went ok now:
$ ~/julia-1.7-
julia-1.7-a7848a28e5/ julia-1.7-DEV-da96fef327/ julia-1.7-DEV-f2ea26d1a1/
B.
Since I got "ERROR: TaskFailedException" on the same line, it it plausible it was from previous invocation, i.e. when running with sh? If not then not even sh needs be implicated. I guess that's the mostly likely reason, since I tried later to run that 1.7 version 20 times in a row without any problems.