Something smells rotten in the state of luametatex...
Hi all,
I have a document here for one of my classes: xml processed via context,
some layers, some two column paragraphs, a bibliography, nothing out of
the ordinary. With mkiv, it compiles cleanly:
mkiv lua stats > used engine: luatex version: 1.13, functionality
level: 7420, banner: this is luatex, version 1.13.0 (tex live 2021)
mkiv lua stats > used hash slots: 48846 of 65536 + 250000
mkiv lua stats > lua properties: engine: lua 5.3, used memory: 212 MB,
ctx: 210 MB, max: unknown MB, hash chars: min(64,40), symbol mask: utf (τεχ)
mkiv lua stats > runtime: 2.775 seconds, 119 processed pages, 119
shipped pages, 42.883 pages/second
system | total runtime: 5.996 seconds
When I compile the same document with luametatex, things slow down to a
crawl:
mkiv lua stats > used engine: luametatex version: 2.0921, functionality
level: 20210903, format id: 594, compiler: gcc
mkiv lua stats > tex properties: 740039 hash slots used of 2097152,
48372 control sequences, approximate memory usage: 36 MB
mkiv lua stats > lua properties: engine: lua 5.4, used memory: 100 MB,
ctx: 93 MB, max: 138 MB, symbol mask: utf (τεχ)
mkiv lua stats > runtime: 27.509 seconds, 119 processed pages, 119
shipped pages, 4.326 pages/second
system | total runtime: 139.458 seconds of 139.493 seconds
So, it takes more than 20 times as long... The CPU of my computer goes
up to 100 %, and the entire machine sometimes becomes unresponsive
during the run. Obviously, I can't provide a MWE, but this seems
bizarre. One thing I see on my terminal: there are hundreds of lines
like this
Overfull \hbox (0.8pt too wide) in paragraph at lines 18--18
6:17: [glue][glue]<l2r par>
On 10/8/2021 6:09 PM, Thomas A. Schmitz via ntg-context wrote:
Hi all,
I have a document here for one of my classes: xml processed via context, some layers, some two column paragraphs, a bibliography, nothing out of the ordinary. With mkiv, it compiles cleanly:
mkiv lua stats > used engine: luatex version: 1.13, functionality level: 7420, banner: this is luatex, version 1.13.0 (tex live 2021) mkiv lua stats > used hash slots: 48846 of 65536 + 250000 mkiv lua stats > lua properties: engine: lua 5.3, used memory: 212 MB, ctx: 210 MB, max: unknown MB, hash chars: min(64,40), symbol mask: utf (τεχ) mkiv lua stats > runtime: 2.775 seconds, 119 processed pages, 119 shipped pages, 42.883 pages/second system | total runtime: 5.996 seconds
When I compile the same document with luametatex, things slow down to a crawl:
mkiv lua stats > used engine: luametatex version: 2.0921, functionality level: 20210903, format id: 594, compiler: gcc mkiv lua stats > tex properties: 740039 hash slots used of 2097152, 48372 control sequences, approximate memory usage: 36 MB mkiv lua stats > lua properties: engine: lua 5.4, used memory: 100 MB, ctx: 93 MB, max: 138 MB, symbol mask: utf (τεχ) mkiv lua stats > runtime: 27.509 seconds, 119 processed pages, 119 shipped pages, 4.326 pages/second system | total runtime: 139.458 seconds of 139.493 seconds
So, it takes more than 20 times as long... The CPU of my computer goes up to 100 %, and the entire machine sometimes becomes unresponsive during the run. Obviously, I can't provide a MWE, but this seems bizarre. One thing I see on my terminal: there are hundreds of lines like this
Overfull \hbox (0.8pt too wide) in paragraph at lines 18--18 6:17: [glue][glue]<l2r par>
[glue][glue] It's almost like some debug setting has been triggered. But that may be a red herring. Anyway, I just wanted to report this; maybe some bigger problem (with the binary or the format?) is lurking somewhere.
no, that's just a more verbose 'overfull box' message ... does adding \dontcomplain help? (After all you can live with .8pt which smells like 2 * .4pt which is the default rule width so maybe some frame stricks out) Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------
participants (2)
-
Hans Hagen
-
Thomas A. Schmitz