Date: Fri, 25 Oct 2019 00:19:04 -0700 From: Mark Millard <marklmi@yahoo.com> To: Phil Scarr <phil.scarr@pm.me> Cc: freebsd-ppc@freebsd.org Subject: Re: FreeBSD 11.3 or 12.x on IBM 9115-505 Message-ID: <53C49809-0A92-4FA8-B7B8-2CB9B9803DE0@yahoo.com> In-Reply-To: <20191024232308.GA23833@lonesome.com> References: <B3B54E60-55FE-4B32-814C-EB1FDD1E8B9B@pm.me> <201d5187612f43a4bc3747f22e7a6712@bege.email> <A19BBF65-9783-4BF8-B279-D3787D774D92@pm.me> <20191024232308.GA23833@lonesome.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2019-Oct-24, at 16:23, Mark Linimon <linimon@lonesome.com> wrote: > On Thu, Oct 24, 2019 at 12:11:41PM +0000, Phil Scarr via freebsd-ppc = wrote: >> I'm disappointed that TeamLinux is so much further ahead than TeamBSD >> on this, but given the relative marketshare of each team, I can't say >> I'm surprised. :-( >=20 > It's more that there are not a lot of these machines are in private > hands, compared to the older Macs and newer IBMs and especially the > Taloses. My tests with Debian showed problems when I tried it several > years ago (just as a "does this box work" test). >=20 > AFAIK within FreeBSD we had a maximum of 3 people with IBM Power5s. >=20 I had a vague memory that C++'s investigation for its memory model included power/powerpc and power5/5+ ended up special. Looking around shows "ISO/IEC JTC1 SC22 WG21 N2745" was revised to say: "Finally, applications that need to use C/C++ sequentially consistent = atomic operations on Power 5/5+ must adhere to the rules set out = separately." with those rules being listed separately in: = http://www.rdrop.com/users/paulmck/scalability/paper/N2745rP5.2010.02.19a.= html Looks like power5/5+ was a bit of an odd ball compared to the others. =3D=3D=3D Mark Millard marklmi at yahoo.com ( dsl-only.net went away in early 2018-Mar)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?53C49809-0A92-4FA8-B7B8-2CB9B9803DE0>