![](https://secure.gravatar.com/avatar/46562c7d27388490d62e8d7a5bc960dd.jpg?s=120&d=mm&r=g)
Hi, When making sure that the metafun manual processes with lmtx (where we have sometimes different implementations of macros) I also checked performance. When we started with this manual, it was still MKII time so we either had direct calls or inbetween calls. Especially multiple runs could take a while (many minutes, iir some 15 depending on how it was run). That was no fun. In mkiv with luatex and the built in mp library this dramatically went down to 18.1 seconds for one run and 14.2 seconds for luajittex, for 428 pages. This not bad considering that a lots of features are used: thousands of mp images, a bunch of external images, many fonts, plenty of buffers, positioning, color, hyperlinks, backgrounds etc. This document uses plenty of lua so that is why we gain a lot with luajittex. In lmtx where we dropped jit and have a less performing backend we expect to need more runtime but in the meantime we're quite okay. Not only the engine is faster, but also some low level macros benefit a bit from new features. I don't know how much lua speedup plays a role but probably not that much. For 448 pages (some 20 more than we had before) we now need 16.5 seconds and 15.7 when we use the compact font mode. So we're halfway between luatex and luajittex now. More than 3 seconds of the runtime is for metapost. (I use the machine as before.) Hans PS. If all the stories are true is should be 1 second on a modern apple at which point differences in measurements become noise but I cannot check that. Some day I might check it on a rpi 4. ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------
![](https://secure.gravatar.com/avatar/5faea8b4fa0adb143383811dc94f0d22.jpg?s=120&d=mm&r=g)
Hi Hans, Thanks for the new upload. For your information, I would like to report that the metafun.tex file typesets almost corretly (actually I often typeset it for testing a new upload…) but I get a PDF with 400 pages instead the 448 you are mentionning… The issues are typically some figures not showing up (for instance figures 13.1 upto 13.13 in chapter 13). If you think this may be useful for you, I can send you off list the PDF I get on MacOS 10.14.6 and today's release of LMTX. Best regards: Otared
On 20 Feb 2021, at 17:27, Hans Hagen
wrote: Hi,
When making sure that the metafun manual processes with lmtx (where we have sometimes different implementations of macros) I also checked performance.
When we started with this manual, it was still MKII time so we either had direct calls or inbetween calls. Especially multiple runs could take a while (many minutes, iir some 15 depending on how it was run). That was no fun.
In mkiv with luatex and the built in mp library this dramatically went down to 18.1 seconds for one run and 14.2 seconds for luajittex, for 428 pages.
This not bad considering that a lots of features are used: thousands of mp images, a bunch of external images, many fonts, plenty of buffers, positioning, color, hyperlinks, backgrounds etc. This document uses plenty of lua so that is why we gain a lot with luajittex.
In lmtx where we dropped jit and have a less performing backend we expect to need more runtime but in the meantime we're quite okay. Not only the engine is faster, but also some low level macros benefit a bit from new features. I don't know how much lua speedup plays a role but probably not that much. For 448 pages (some 20 more than we had before) we now need 16.5 seconds and 15.7 when we use the compact font mode. So we're halfway between luatex and luajittex now. More than 3 seconds of the runtime is for metapost. (I use the machine as before.)
Hans
PS. If all the stories are true is should be 1 second on a modern apple at which point differences in measurements become noise but I cannot check that. Some day I might check it on a rpi 4.
----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl ----------------------------------------------------------------- ___________________________________________________________________________________ If your question is of interest to others as well, please add an entry to the Wiki!
maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://context.aanhet.net archive : https://bitbucket.org/phg/context-mirror/commits/ wiki : http://contextgarden.net ___________________________________________________________________________________
![](https://secure.gravatar.com/avatar/46562c7d27388490d62e8d7a5bc960dd.jpg?s=120&d=mm&r=g)
Thanks for the new upload.
For your information, I would like to report that the metafun.tex file typesets almost corretly (actually I often typeset it for testing a new upload…) but I get a PDF with 400 pages instead the 448 you are mentionning… The issues are typically some figures not showing up (for instance figures 13.1 upto 13.13 in chapter 13).
If you think this may be useful for you, I can send you off list the PDF I get on MacOS 10.14.6 and today's release of LMTX. Did you check the log? Does some graphic fail? The missing pages can be
On 2/20/2021 6:43 PM, Otared Kavian wrote: the summary chapter (due to just one run). 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
-
Otared Kavian