Welcome to the main channel on the development of MoarVM, a virtual machine for NQP and Rakudo (moarvm.org). This channel is being logged for historical purposes.
Set by lizmat on 24 May 2021.
Kaiepi what do spesh problems tend to look like? 12:09
nine Kaiepi: usually, error messages that don't make any sense at all. In the worst case segfaults. 13:16
Kaiepi there's this bizarre timing issue that prevents me from benchmarking more than one code block in a script 14:11
it goes away with MVM_SPESH_DISABLE=1, but not MVM_JIT_DISABLE=1
past the first code block that ever gets iterated, there's a jump in how long it takes for the code object to invoke, even when written identically, but as different code objects 14:12
it happens even when each code block is iterated over in parallel 14:13
cloning that first code object and iterating with that is fine, though 14:14
nine Well that's the point of spesh: to take data collected at runtime and create versions of the code that process similar inputs faster. Of course that does make benchmarking harder. 16:16
Geth MoarVM: MasterDuke17++ created pull request #1733:
Bump mimalloc to v2.0.9		 19:45
Geth MoarVM: d54305637c | (Daniel Green)++ | 3rdparty/mimalloc
Bump mimalloc to v2.0.9
 20:47
MoarVM: a1e2a57a5a | MasterDuke17++ (committed using GitHub Web editor) | 3rdparty/mimalloc
Merge pull request #1733 from MasterDuke17/bump_mimalloc_to_v2.0.9

Bump mimalloc to v2.0.9

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!