MasterDuke17

> Why can we assume that the array is in a good state? Well, I believe we’re allowed to *act* as if it’s in a good state. Modifying an array...

(Caveat up front, I don't know anything about layout, print design, etc, so I'm not intending this as a response to the merits of this specific issue). It may be...

Does https://github.com/Raku/nqp/commit/e7d1bbe3df help?

A very rough overview of the timing over time https://gist.github.com/Whateverable/7b710bfdd4ec04296c9af6541a708ee9 (notice this is with a value 10x smaller). Here https://gist.github.com/Whateverable/4be57500c7f78aa78e08367cba9b9421 you can see the impact of the new-disp merge for...

Yeah, I think disabling spesh also disables the jit.Sent from my iPhoneOn Jun 29, 2024, at 2:11 PM, Elizabeth Mattijsen ***@***.***> wrote: I've been able to golf this to: my...

FWIW, on my Ryzen 3700x (8-core Zen 2), a normal spectest is ~200s and disabling the exprJIT drops that to ~195s. A normal compile of CORE.c takes ~63s and disabling...

t/spec/S32-io/IO-Socket-Async.t is a flapper.

This is still a problem, Zoffix's last example just crashed my KDE session.

> but `*.is-prime` is perhaps not a great workload for this benchmark anyway. I assume we are just calculating it way too fast, so the overhead of coordinating the workers...

Attaching gdb and running a backtrace and then `call MVM_dump_backtrace(tc)` gave this: ``` Attaching to process 2156078 [New LWP 2156079] [Thread debugging using libthread_db enabled] Using host libthread_db library "/usr/lib/libthread_db.so.1"....