On Thu, Dec 14, 2023 at 11:23 PM William Stein <wst...@gmail.com> wrote:
> I just did some new benchmarks comparing lualatex to xelatex on some > basic formulas like what I think is likely > to be relevant. This just involved running laulatex and xelatex many > times on the same tex file full of equations > and looking at the time. AFAIK there is a way to speeding things up with luatex on files with the same preamble. Namely, one can generate a custom format, speeding up things like font loading. (might be relevant for speeding up view()). But it's above my pay grade, we need a real TeX hacker here. Maybe Doris (in CC) could help us. Best, Dima > I used a ramdisk to avoid disk speed > mattering. The times were very stable: > > lualatex -- 1 second > xelatex -- 1.2 seconds > pdflatex -- 0.43 seconds > > So I totally retract my concern about lualatex being slower than > xelatex. In fact it's about the same or faster > in this benchmark. That said, pdflatex is noticeably faster than either. > > -- William > > On Thu, Dec 14, 2023 at 2:01 PM Dima Pasechnik <dimp...@gmail.com> wrote: > > > > On Thu, Dec 14, 2023 at 9:30 PM William Stein <wst...@gmail.com> wrote: > > > > > > Hi, > > > > > > Last time I benchmarked things, luatex was sometimes the slowest of > > > the three major latex compilers, sometimes by a factor of 2. > > > I realize benchmarking could be impacted by things like how exactly > > > Lua was built on the target platform, as part of latex. In any case, > > > for > > > this application, performance could matter. > > > > Let me say few words on preference for lualatex. Once I heard about > > Kwankyu's PR, I immediately suggested that should be luatex, not > > xetex. After trying xetex he decided to switch it to luatex. > > > > 1) lua(lat)tex is the official TexLive's successor of pdf(la)tex. It > > has proper unicode support, > > one can use unicode maths, etc. luatex is packaged better within TeXLive > > too. > > > > 2) xelatex has been in maintenance mode for some year by now, and > > it's unicode support is a bit funny. (in pdflatex it's next to nonexistent) > > > > 3) For building sagemath PDF docs it doesn't matter much, one doesn't > > do it often, it's not instant. > > Using unicode in docstrings makes them more readable and faster (less > > MathJax). > > > > OTOH Sage's view() might need the speed, but you can change what > > engine you use there > > irrespective of what's used for building the docs. > > > > Dima > > > > > > I had nothing to do with them making xelatex the default, and I think > > > Lua and Luatex are beautiful pieces of software. > > > > > > -- William > > > > > > On Thu, Dec 14, 2023 at 1:22 PM Dima Pasechnik <dimp...@gmail.com> wrote: > > > > > > > > Hi Doris, > > > > It just has been turned to use lualatex > > > > > > > > HTH > > > > Dima > > > > > > > > On Thu, Dec 14, 2023 at 8:06 PM 'Doris Behrendt' via sage-devel > > > > <sage-devel@googlegroups.com> wrote: > > > > > > > > > > Hi all, > > > > > > > > > > why xelatex and not lualatex? > > > > > > > > > > Doris > > > > > > > > > > > On 14. Dec 2023, at 10:18, Kwankyu Lee <ekwan...@gmail.com> wrote: > > > > > > > > > > > > Hi, > > > > > > > > > > > > The PR > > > > > > > > > > > > https://github.com/sagemath/sage/pull/36861 > > > > > > > > > > > > aims at switching from pdflatex to xelatex the latex engine to > > > > > > render objects in pdf using the command like "view(objects)" as > > > > > > well as in building the sage documentation. > > > > > > > > > > > > As xelatex natively supports Unicode, we removed lots of commands > > > > > > like > > > > > > > > > > > > \DeclareUnicodeCharacter{03B1}{\ensuremath{\alpha}} > > > > > > > > > > > > which were needed to support Unicode in pdflatex. > > > > > > > > > > > > We worry that some Unicode functionality was lost in this process, > > > > > > and so we invite Unicode users to test with the PR. > > > > > > > > > > > > Here is the Binder link to where you can play around with > > > > > > xelatex-enabled sage: > > > > > > > > > > > > https://mybinder.org/v2/gh/kwankyu/sage/use-xelatex-by-default-binder > > > > > > > > > > > > or you should build sage with the PR. > > > > > > > > > > > > If you find some Unicode-related defect not existent in sage 10.2, > > > > > > please report it to the PR. > > > > > > > > > > > > Thanks. > > > > > > > > > > > > -- > > > > > > You received this message because you are subscribed to the Google > > > > > > Groups "sage-devel" group. > > > > > > To unsubscribe from this group and stop receiving emails from it, > > > > > > send an email to sage-devel+unsubscr...@googlegroups.com. > > > > > > To view this discussion on the web visit > > > > > > https://groups.google.com/d/msgid/sage-devel/ac185d29-e64b-42a3-a4b9-d25e2b31474cn%40googlegroups.com. > > > > > > > > > > -- > > > > > You received this message because you are subscribed to the Google > > > > > Groups "sage-devel" group. > > > > > To unsubscribe from this group and stop receiving emails from it, > > > > > send an email to sage-devel+unsubscr...@googlegroups.com. > > > > > To view this discussion on the web visit > > > > > https://groups.google.com/d/msgid/sage-devel/93C23CA8-99D8-4051-B01B-DBF0CDEF8FD1%40me.com. > > > > > > > > -- > > > > You received this message because you are subscribed to the Google > > > > Groups "sage-devel" group. > > > > To unsubscribe from this group and stop receiving emails from it, send > > > > an email to sage-devel+unsubscr...@googlegroups.com. > > > > To view this discussion on the web visit > > > > https://groups.google.com/d/msgid/sage-devel/CAAWYfq343js%3DCYzBMt9z8PmX5AkPd2qfkpRqUaB%3DUSL21b%3D5tw%40mail.gmail.com. > > > > > > > > > > > > -- > > > William (http://wstein.org) > > > > > > -- > > > You received this message because you are subscribed to the Google Groups > > > "sage-devel" group. > > > To unsubscribe from this group and stop receiving emails from it, send an > > > email to sage-devel+unsubscr...@googlegroups.com. > > > To view this discussion on the web visit > > > https://groups.google.com/d/msgid/sage-devel/CACLE5GDwsFq2CJzQ9xFO8F14N-7zbZuTVQ1-hFuKo9mLt_Ekpw%40mail.gmail.com. > > > > -- > > You received this message because you are subscribed to the Google Groups > > "sage-devel" group. > > To unsubscribe from this group and stop receiving emails from it, send an > > email to sage-devel+unsubscr...@googlegroups.com. > > To view this discussion on the web visit > > https://groups.google.com/d/msgid/sage-devel/CAAWYfq2882pWK9YWE-ZiVzmR0xXqXM6zX64_Y7ko96ko%3Dh0ftg%40mail.gmail.com. > > > > -- > William (http://wstein.org) > > -- > You received this message because you are subscribed to the Google Groups > "sage-devel" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to sage-devel+unsubscr...@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/sage-devel/CACLE5GCVeJW_vyAXHA8UiejPUFDS4%3D_MoJuNfZ7Ro__Y%3D2YWRA%40mail.gmail.com. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/CAAWYfq39hGSAq73zjfLoRj50ntMzjgQwepyFmWpjKf%3DgM2u%3D%3Dg%40mail.gmail.com.
