Path: csiph.com!eternal-september.org!feeder.eternal-september.org!nntp.eternal-september.org!.POSTED!not-for-mail From: Kragen Javier Sitaker Newsgroups: comp.lang.forth Subject: Re: First result in riscv optimiser Date: Fri, 11 Sep 2026 07:50:13 -0300 Organization: Primarily biological and memetic Lines: 28 Message-ID: <877bksdpui.fsf@debian> References: <87ecf96ljg.fsf@debian> <87y0dcgvxf.fsf@debian> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Injection-Date: Fri, 11 Sep 2026 10:50:18 +0000 (UTC) Injection-Info: dont-email.me; logging-data="2966685"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1+ba0/BzXYgitNM6kt+6h5f"; posting-host="5e8b773666f2baa1b3e9fe7abeba1530" User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/28.2 (gnu/linux) Cancel-Lock: sha1:+4byqaFY8gLrJ7P3LuSA/qw1dxQ= sha1:1Ei3TA8ZCxrEEELmN6hDeQJ1ygo= sha256:c7MfGgXAfVre8S3BxeePeJeNWlSldnwrxJSrxV4b7l4= sha1:FlZBGzLEJK/io9FWkjcOBteWHX0= sha256:BCoPzLC94DROPpGviTnnYuQa4fg7N6xvzaknni/urVQ= Xref: csiph.com comp.lang.forth:135658 albert@spenarnc.xs4all.nl writes: > If a word is optimised it is totally inlined. As well as being a powerful optimization in its own right, I am guessing that that opens up a lot of opportunities for the stack-manipulation- annihilation optimization you were talking about in your previous post. > [...] > > I got to the point that the original (unadulterated to favor > a particular compiler) Byte benchmark performed in the league > of mpe Forth. That’s very impressive! > I have 106 peephole patterns, some quite complicated. How confident are you that all 106 are correct? > [...] > I decided to give up on i86 and concentrate on RISCV. This kind of thing seems like it could be a significant advantage for RISC-V, if people find that the cost-benefit ratio for writing compilers for it is better than for 8086, i386, or amd64 (not sure which ISA you meant by “i86”). Kragen