Date: Tue, 01 Sep 2015 11:03:31 -0400 From: Lowell Gilbert <freebsd-questions-local@be-well.ilk.org> To: Dave <dave@dgmm.net> Cc: freebsd-questions@freebsd.org Subject: Re: system temperatures Message-ID: <44oahmdvwc.fsf@lowell-desk.lan> In-Reply-To: <1580689.71VicvVbGW@amd.asgard.uk> (dave@dgmm.net's message of "Mon, 31 Aug 2015 22:39 %2B0100") References: <55E324C4.4010500@hiwaay.net> <1568095.rmZeycZDhO@amd.asgard.uk> <55E4B2FA.9090501@hiwaay.net> <1580689.71VicvVbGW@amd.asgard.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
Dave <dave@dgmm.net> writes: > On Monday 31 August 2015 15:08:36 William A. Mahaffey III wrote: >> On 08/31/15 13:42, Dave wrote: >> > On Sunday 30 August 2015 10:49:34 you wrote: >> >> It is convenient on my various linux boxen to use lm_sensors to retrieve >> >> apparently accurate temps. for various system components (CPU's), as >> >> well as data on fan speeds, etc. Under FreeBSD (9.3R-p21), sysctl >> >> provides some of this info, but apparently inaccurately. It would be >> >> sweet to fix that minor issue, seems like it would be a bit more than >> >> just sweet for remote servers, etc. Do I file this as a problem report >> >> of some sort ? If so, how do I go about doing that ? TIA & have a >> >> nice weekend. >> > It might be worth doing this: >> > >> > sysctl -a | grep temp >> > dev.cpu.0.temperature: 42.0C >> > dev.cpu.1.temperature: 42.0C >> > dev.cpu.2.temperature: 42.0C >> > dev.cpu.3.temperature: 42.0C >> > dev.cpu.4.temperature: 42.0C >> > dev.cpu.5.temperature: 42.0C >> > dev.amdtemp.0.%desc: AMD CPU On-Die Thermal Sensors >> > dev.amdtemp.0.%driver: amdtemp >> > dev.amdtemp.0.%parent: hostb4 >> > dev.amdtemp.0.sensor0.core0: 42.0C >> > >> > ...just to see what else might show up in relation to temperatures >> > in the various sysctl variables. >> > >> > both healthd and xmbmon in their default states report a CPU temp of ~143c on my system. >> > 42c is ~107F so that's not the problem. >> >> >> Hmmmm .... OK, your command, followed by my usual incantation: >> >> >> [root@kabini1, /etc, 3:00:06pm] 352 % sysctl -a | grep temp >> amdtemp0: <AMD CPU On-Die Thermal Sensors> on hostb5 >> net.inet6.ip6.use_tempaddr: 0 >> net.inet6.ip6.temppltime: 86400 >> net.inet6.ip6.tempvltime: 604800 >> net.inet6.ip6.prefer_tempaddr: 0 >> hw.usb.template: 0 >> dev.cpu.0.temperature: 11.8C >> dev.cpu.1.temperature: 11.8C >> dev.cpu.2.temperature: 11.8C >> dev.cpu.3.temperature: 11.8C >> dev.amdtemp.0.%desc: AMD CPU On-Die Thermal Sensors >> dev.amdtemp.0.%driver: amdtemp >> dev.amdtemp.0.%parent: hostb5 >> dev.amdtemp.0.sensor0.core0: 11.8C >> [root@kabini1, /etc, 3:00:08pm] 353 % sysctl -A | egrep >> '(temperature|usage)' >> dev.cpu.0.temperature: 11.7C >> dev.cpu.0.cx_usage: 100.00% 0.00% last 880us >> dev.cpu.1.temperature: 11.7C >> dev.cpu.1.cx_usage: 100.00% last 10us >> dev.cpu.2.temperature: 11.7C >> dev.cpu.2.cx_usage: 100.00% last 11us >> dev.cpu.3.temperature: 11.7C >> dev.cpu.3.cx_usage: 100.00% last 10us >> [root@kabini1, /etc, 3:00:17pm] 354 % >> >> >> I'd say it's off, room temp is around 76F, or about 26-ish C, & the CPU >> is generating heat, so it can't be cooler (Thermodynamics). My CPU is an >> AMD Sempron 3850, jaguar-kabini, quad core, fairly new, April 9 2014 to >> be precise, exactly the day I bought it :-). I guess it's either >> something mbd or too-new CPU, but that is rank speculation. Your numbers >> look at least plausible, so I'd say you are good to go, just trying to >> get myself there as well. Those CPU's are apparently fairly widely used >> in laptops & low-end desktops, so while new-ish, they are certainly not >> rare birds. I like the output of your command a bit better than mine, I >> think I will adopt it :-) .... Any more info needed, don't hesitate. >> >> >> [root@kabini1, /etc, 3:03:44pm] 354 % uname -a >> FreeBSD kabini1.local 9.3-RELEASE-p24 FreeBSD 9.3-RELEASE-p24 #0: Sat >> Aug 22 01:54:44 UTC 2015 >> root@amd64-builder.daemonology.net:/usr/obj/usr/src/sys/GENERIC amd64 >> [root@kabini1, /etc, 3:03:47pm] 355 % >> >> > I'm pretty much at the limit of what I can suggest, but thought at > least checking the sysctl stuff might be useful since in my experience > pretty much everything else which claims to be able to read system and/or > CPU temperature tends to fail with massive over-readings. > > I'll be watching this thread with interest to see if anything useful for > me turns up, At some point, AMD claimed that they weren't going to try accurate temperature measurements any more, and just focus on having good enough information for power control. I don't really have the time to look up the details in their manuals, though -- this information is not covered in the data sheet where I expected to find it.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?44oahmdvwc.fsf>