Date: Mon, 26 Sep 2022 08:20:15 -0600 From: Warner Losh <imp@bsdimp.com> To: Lev Serebryakov <lev@freebsd.org> Cc: FreeBSD Current <freebsd-current@freebsd.org> Subject: Re: Did clang 14 lose some intrinsics support? Message-ID: <CANCZdfqnAFYCvg8k88peh2bKB%2B9G9GX1BCQvhUoggYcJL0Z0Aw@mail.gmail.com> In-Reply-To: <8cd08b95-396a-b8a3-e33a-b0482a7f23e5@FreeBSD.org> References: <YzCL/%2BUEhLy7kHaL@lorvorc.mips.inka.de> <zgen-6zwv-wny@FreeBSD.org> <1A903FD8-D904-4B91-ABC4-2F704F0E2CF4@FreeBSD.org> <YzDKQsfciGd%2BJBQ5@lorvorc.mips.inka.de> <430A830E-3473-4EF4-9605-039F8254999C@FreeBSD.org> <8cd08b95-396a-b8a3-e33a-b0482a7f23e5@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--000000000000a5fec005e9953a5b Content-Type: text/plain; charset="UTF-8" On Mon, Sep 26, 2022, 7:54 AM Lev Serebryakov <lev@freebsd.org> wrote: > On 26.09.2022 13:03, Dimitry Andric wrote: > > > Sure, but if you are compiling without -mavx, why would you want the AVX > > intrinsics? You cannot use AVX intrinsics anyway, if AVX is not enabled. > Because autovectorization (generation of SSE or AVX instructions by > compiler itself, without intrinsics) can pessimize code. > > Sometimes it is valuable to know exactly where AVX is used. I don't > have examples on hands, but I've seen situations, when autovectorized code > was much slower than scalar code. > The detection method that dim@ outline will work just fine for the autodetect script. It just replaces the internal, charging undocumented names for standard ones. How you later compile individual bits of code is orthogonal. Warner > --000000000000a5fec005e9953a5b Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable <div dir=3D"auto"><div><br><br><div class=3D"gmail_quote"><div dir=3D"ltr" = class=3D"gmail_attr">On Mon, Sep 26, 2022, 7:54 AM Lev Serebryakov <<a h= ref=3D"mailto:lev@freebsd.org">lev@freebsd.org</a>> wrote:<br></div><blo= ckquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #c= cc solid;padding-left:1ex">On 26.09.2022 13:03, Dimitry Andric wrote:<br> <br> > Sure, but if you are compiling without -mavx, why would you want the A= VX<br> > intrinsics? You cannot use AVX intrinsics anyway, if AVX is not enable= d.<br> =C2=A0 =C2=A0Because autovectorization (generation of SSE or AVX instructio= ns by compiler itself, without intrinsics) can pessimize code.<br> <br> =C2=A0 =C2=A0Sometimes it is valuable to know exactly where AVX is used. I = don't have examples on hands, but I've seen situations, when autove= ctorized code was much slower than scalar code.<br></blockquote></div></div= ><div dir=3D"auto"><br></div><div dir=3D"auto">The detection method that di= m@ outline will work just fine for the autodetect script. It just replaces = the internal, charging undocumented names for standard ones.</div><div dir= =3D"auto"><br></div><div dir=3D"auto">How you later compile individual bits= of code is orthogonal.=C2=A0</div><div dir=3D"auto"><br></div><div dir=3D"= auto">Warner</div><div dir=3D"auto"><div class=3D"gmail_quote"><blockquote = class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid= ;padding-left:1ex"> </blockquote></div></div></div> --000000000000a5fec005e9953a5b--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANCZdfqnAFYCvg8k88peh2bKB%2B9G9GX1BCQvhUoggYcJL0Z0Aw>