Date: Fri, 19 Aug 2022 23:39:34 -0700 From: Kevin Oberman <rkoberman@gmail.com> To: Alexander Motin <mav@freebsd.org> Cc: current <current@freebsd.org> Subject: Re: Status of Alder Lake support Message-ID: <CAN6yY1s%2BG3mMDvto3kUjC-spLfNXF2aU1oq6iLxCqCna8%2BpC=w@mail.gmail.com> In-Reply-To: <2ba3e191-64c1-b854-3eb3-080345b56839@FreeBSD.org> References: <CAN6yY1t75UEVBK5cnvwmEopCOyKmcOEX3bRJkRnX37kPa2gkgg@mail.gmail.com> <2ba3e191-64c1-b854-3eb3-080345b56839@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--000000000000fefa8f05e6a67aed Content-Type: text/plain; charset="UTF-8" Thanks for the info, Alexander. I really appreciate it. I'll find out for myself next week when my SSD gets here, though I might try booting from a thumb drive and see what happens. On Fri, Aug 19, 2022, 18:43 Alexander Motin <mav@freebsd.org> wrote: > Hi Kevin, > > On 19.08.2022 20:50, Kevin Oberman wrote: > > What is the current state of support for Alder Lake CPUs with a mix of > > "performance" and "Efficiency" cores. I just received my first system > > with such a processor and will be installing FreeBSD as soon as my SSD > > arrives. I have no idea what issues I might run into. (Will it even > work?) > > Generally they work. I have one in my lab since they appeared, and its > biggest problem is UEFI console screwed by ASUS, which I partially > fixed. At some BIOS versions they also broke PXE booting, but then > fixed it in later update. > > The FreeBSD scheduler still has no idea about P and E cores, just trying > to balance load equally based on cache topology (which is not symmetric > there), but it is not too bad, since cores performance is not so > dramatically different. > > hwpmc(4) also does not differentiate P and E cores, and since they have > different counters, it means that only few universal architectural > counters are usable, which are sufficient for basic profiling though. > > There were some stability issues reported, but were solved by either > FreeBSD or BIOS update, so still not really diagnosed AFAIK. > > -- > Alexander Motin > --000000000000fefa8f05e6a67aed Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable <div dir=3D"auto">Thanks for the info, Alexander. I really appreciate it. I= 'll find out for myself next week when my SSD gets here, though I might= try booting from a thumb drive and see what happens.</div><br><div class= =3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_attr">On Fri, Aug 19, 2022= , 18:43 Alexander Motin <<a href=3D"mailto:mav@freebsd.org">mav@freebsd.= org</a>> wrote:<br></div><blockquote class=3D"gmail_quote" style=3D"marg= in:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi Kevin,<br> <br> On 19.08.2022 20:50, Kevin Oberman wrote:<br> > What is the current state of support for Alder Lake CPUs with a mix of= <br> > "performance" and "Efficiency"=C2=A0 cores. I just= received my first system <br> > with such a processor and will be installing FreeBSD as soon as my SSD= <br> > arrives. I have no idea what issues I might run into. (Will it even wo= rk?)<br> <br> Generally they work.=C2=A0 I have one in my lab since they appeared, and it= s <br> biggest problem is UEFI console screwed by ASUS, which I partially <br> fixed.=C2=A0 At some BIOS versions they also broke PXE booting, but then <b= r> fixed it in later update.<br> <br> The FreeBSD scheduler still has no idea about P and E cores, just trying <b= r> to balance load equally based on cache topology (which is not symmetric <br= > there), but it is not too bad, since cores performance is not so <br> dramatically different.<br> <br> hwpmc(4) also does not differentiate P and E cores, and since they have <br= > different counters, it means that only few universal architectural <br> counters are usable, which are sufficient for basic profiling though.<br> <br> There were some stability issues reported, but were solved by either <br> FreeBSD or BIOS update, so still not really diagnosed AFAIK.<br> <br> -- <br> Alexander Motin<br> </blockquote></div> --000000000000fefa8f05e6a67aed--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAN6yY1s%2BG3mMDvto3kUjC-spLfNXF2aU1oq6iLxCqCna8%2BpC=w>