Date: Sun, 17 Mar 2019 11:37:26 +0000 From: bugzilla-noreply@freebsd.org To: toolchain@FreeBSD.org Subject: [Bug 236582] Enable LLVM openmp on i386 Message-ID: <bug-236582-29464-ALIlREtMlh@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-236582-29464@https.bugs.freebsd.org/bugzilla/> References: <bug-236582-29464@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D236582 Dimitry Andric <dim@FreeBSD.org> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |emaste@freebsd.org --- Comment #2 from Dimitry Andric <dim@FreeBSD.org> --- Unfortunately it does not compile for our old default i486 CPU: --- kmp_csupport.pico --- /usr/src/contrib/openmp/runtime/src/kmp_csupport.cpp:565:7: error: '_mm_set= csr' needs target feature sse __kmp_load_mxcsr(&serial_team->t.t_mxcsr); ^ /usr/src/contrib/openmp/runtime/src/kmp.h:3669:29: note: expanded from macro '__kmp_load_mxcsr' #define __kmp_load_mxcsr(p) _mm_setcsr(*(p)) ^ 1 error generated. So we can't unconditionally apply this to the i386 arch. Personally I'd be = in favor of raising the requirements for i386 to having at least SSE2, but tha= t is going to get lots of pushback from people that like to run old hardware. :) Maybe there is some way to detect the actual CPU or CPUTYPE setting at build time, and trigger on this. Note also that the patch from ports r447281 is maybe not entirely correct, = in the sense that it enables code also for arm, aarch and mips. I have no ide= a at all whether those are supported in any way. --=20 You are receiving this mail because: You are the assignee for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-236582-29464-ALIlREtMlh>