Date: Thu, 30 Aug 2018 01:56:23 -0700 From: Mark Millard <marklmi@yahoo.com> To: Emmanuel Vadot <manu@bidouilliste.com> Cc: Daniel Braniss <danny@cs.huji.ac.il>, Mark Millard via freebsd-arm <freebsd-arm@freebsd.org> Subject: Re: FYI: A Pine64+ 2GB thermal result during a -j4 buildworld's libllvm build on head -r338341 Message-ID: <BED40358-8EF7-434B-9081-D4DCB2118C3A@yahoo.com> In-Reply-To: <984F9F42-E8A0-4E01-9391-AADFA87A01D6@yahoo.com> References: <4FB22C52-44DF-4868-B30F-5E07DE9F2198@yahoo.com> <20180829135057.742947ba512467bd82671cf0@bidouilliste.com> <C72097A5-931B-4420-8336-EB9B1366165C@cs.huji.ac.il> <20180829151201.e3acdd8f65a60e861987489f@bidouilliste.com> <5B6E0726-2ACA-4051-8692-8D8A901BBF91@yahoo.com> <984F9F42-E8A0-4E01-9391-AADFA87A01D6@yahoo.com>
next in thread | previous in thread | raw e-mail | index | archive | help
[Adding figures for the Pine64+ 2GB being idle.] On 2018-Aug-29, at 11:48 AM, Mark Millard <marklmi at yahoo.com> wrote: > [The build got past doing llvm library related builds and so is doing > less intense buildworld activity.] >=20 > On 2018-Aug-29, at 8:58 AM, Mark Millard <marklmi at yahoo.com> wrote: >=20 >> On 2018-Aug-29, at 6:12 AM, Emmanuel Vadot <manu at bidouilliste.com> = wrote: >>=20 >>> On Wed, 29 Aug 2018 15:41:32 +0300 >>> Daniel Braniss <danny@cs.huji.ac.il> wrote: >>>=20 >>>>=20 >>>>=20 >>>>> On 29 Aug 2018, at 14:50, Emmanuel Vadot <manu@bidouilliste.com> = wrote: >>>>>=20 >>>>> On Wed, 29 Aug 2018 04:07:28 -0700 >>>>> Mark Millard via freebsd-arm <freebsd-arm@freebsd.org = <mailto:freebsd-arm@freebsd.org>> wrote: >>>>>=20 >>>>>> This head -r338341 based Pine64+ 2GB has a case, heatsinks, >>>>>> and a fan. >>>>>>=20 >>>>>> # sysctl -a | grep therm >>>>>> dev.aw_thermal.0.gpu2: 28C >>>>>> dev.aw_thermal.0.gpu1: 29C >>>>>> dev.aw_thermal.0.cpu: 31C >>>>>> dev.aw_thermal.0.%parent: simplebus0 >>>>>> dev.aw_thermal.0.%pnpinfo: name=3Dthermal_sensor@1c25000 = compat=3Dallwinner,sun50i-a64-ths >>>>>> dev.aw_thermal.0.%location:=20 >>>>>> dev.aw_thermal.0.%driver: aw_thermal >>>>>> dev.aw_thermal.0.%desc: Allwinner Thermal Sensor Controller >>>>>> dev.aw_thermal.%parent:=20 >>>>>>=20 >>>>>=20 >>>>> That seems very very low. >>>> mine (Friendlyarm/allwinner/a64)is too cold to believe: >>>>=20 >>>> dev.aw_thermal.0.gpu2: 17C >>>> dev.aw_thermal.0.gpu1: 17C >>>> dev.aw_thermal.0.cpu: 15C >>>> dev.aw_thermal.0.%parent: simplebus0 >>>> dev.aw_thermal.0.%pnpinfo: name=3Dthermal_sensor@1c25000 = compat=3Dallwinner,sun50i-a64-ths >>>> dev.aw_thermal.0.%location:=20 >>>> dev.aw_thermal.0.%driver: aw_thermal >>>> dev.aw_thermal.0.%desc: Allwinner Thermal Sensor Controller >>>> dev.aw_thermal.%parent: >>>=20 >>> Indeed it does. Do you have the big ass heatsink + fan on it ? >>> = https://www.friendlyarm.com/index.php?route=3Dproduct/product&path=3D82&pr= oduct_id=3D133 >>>=20 >>> On my pine64 (Not LTS so A64), when it's idle the three sensors are >>> ~22C. My IR thermometer reports a max temp of 33C when I 'scan' the >>> chip. >>>=20 >>> Could you sysctl dev.aw_sid.0.ths-calib ? >>=20 >> The buildworld is doing libclang compiles now: >>=20 >> # sysctl dev.aw_sid.0.ths-calib dev.aw_thermal >> dev.aw_sid.0.ths-calib: 820788078307 >> dev.aw_thermal.0.gpu2: 27C >> dev.aw_thermal.0.gpu1: 28C >> dev.aw_thermal.0.cpu: 30C >> dev.aw_thermal.0.%parent: simplebus0 >> dev.aw_thermal.0.%pnpinfo: name=3Dthermal_sensor@1c25000 = compat=3Dallwinner,sun50i-a64-ths >> dev.aw_thermal.0.%location:=20 >> dev.aw_thermal.0.%driver: aw_thermal >> dev.aw_thermal.0.%desc: Allwinner Thermal Sensor Controller >> dev.aw_thermal.%parent:=20 >>=20 >>> I'll try to test on my NanoPi 64 with the heatsink. >>>=20 >>>> and it?s summer here. >>>>=20 >>>> danny >>>>=20 >>>>>=20 >>>>> With 4 cpuburn-a53 process running I easily get to ~60C for cpu = and >>>>> ~55C for the gpus sensor. >>>>> https://people.freebsd.org/~manu/cpuburn-a53 = <https://people.freebsd.org/~manu/cpuburn-a53> >>=20 >> It will be a notable before the buildworld completes >> (if it finishes normally). >>=20 >>>>>> This was during a -j4 buildworld's: >>>>>>=20 >>>>>> Building = /usr/obj/cortexA53_clang/arm64.aarch64/usr/src/arm64.aarch64/lib/clang/lib= llvm/CodeGen/MachineBranchProbabilityInfo.o >>>>>> Building = /usr/obj/cortexA53_clang/arm64.aarch64/usr/src/arm64.aarch64/lib/clang/lib= llvm/CodeGen/MachineCSE.o >>>>>> Building = /usr/obj/cortexA53_clang/arm64.aarch64/usr/src/arm64.aarch64/lib/clang/lib= llvm/CodeGen/MachineCombiner.o >>>>>> Building = /usr/obj/cortexA53_clang/arm64.aarch64/usr/src/arm64.aarch64/lib/clang/lib= llvm/CodeGen/MachineCopyPropagation.o >>>>>>=20 >>>>>> No HDMI is plugged in. Ethernet is in use. The serial console >>>>>> is in use. A USB powered hub with a USB drive is plugged in >>>>>> to provide the swap partition. The UFS root file system is on >>>>>> a microsdhc card in the microsd slot and it has TRIM enabled. >>>>>> Nothing else other than the power cord is connected. >>>>>>=20 >>>>>> (The microsd card use with TRIM was requested by someone. >>>>>> I normally use a root filesystem on a USB device because >>>>>> booting from an e.MCC on an sdcard adapter that is put in >>>>>> the sdcard slot does not work these days and I normally >>>>>> avoid microsd cards for much beyond loading the kernel.) >>>=20 >>=20 >=20 > During a less intense build stage it shows the likes > of: >=20 > # sysctl dev.aw_sid.0.ths-calib dev.aw_thermal > dev.aw_sid.0.ths-calib: 820788078307 > dev.aw_thermal.0.gpu2: 24C > dev.aw_thermal.0.gpu1: 24C > dev.aw_thermal.0.cpu: 26C > dev.aw_thermal.0.%parent: simplebus0 > dev.aw_thermal.0.%pnpinfo: name=3Dthermal_sensor@1c25000 = compat=3Dallwinner,sun50i-a64-ths > dev.aw_thermal.0.%location:=20 > dev.aw_thermal.0.%driver: aw_thermal > dev.aw_thermal.0.%desc: Allwinner Thermal Sensor Controller > dev.aw_thermal.%parent:=20 >=20 > It is summer and the room temperature is above > 26 degC, more like 28.3 degC. >=20 > It does look like the figures are systematically low in > my context. >=20 > It will be some time before I'll get a figure for idle. The Pine64+ 2GB having been basically idle for hours (and still being basically idle): # sysctl dev.aw_sid.0.ths-calib dev.aw_thermal = = dev.aw_sid.0.ths-calib: = 820788078307 dev.aw_thermal.0.gpu2: 19C dev.aw_thermal.0.gpu1: 19C dev.aw_thermal.0.cpu: 18C dev.aw_thermal.0.%parent: simplebus0 dev.aw_thermal.0.%pnpinfo: name=3Dthermal_sensor@1c25000 = compat=3Dallwinner,sun50i-a64-ths dev.aw_thermal.0.%location:=20 dev.aw_thermal.0.%driver: aw_thermal dev.aw_thermal.0.%desc: Allwinner Thermal Sensor Controller dev.aw_thermal.%parent:=20 So: Well below room temperature and similar to Daniel B.'s results. It does suggest a delta of about 13 degC between idle and the -j4 libllvm example cpu figures --and about 10 degC for the gpus. =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?BED40358-8EF7-434B-9081-D4DCB2118C3A>