Hello Jean,
On 02/19/2010 02:52 PM, Jean Delvare wrote:
> Hi Artem,
>
> On Thu, 18 Feb 2010 16:25:48 +0500, Artem S. Tashkinov wrote:
>> echo adt7473 0x2e> /sys/class/i2c-adapter/i2c-3/new_device produced the
>> desired results:
>>
>> adt7473-i2c-3-2e
>> Adapter: NVIDIA i2c adapter
>> in1: +3.00 V (min = +0.00 V, max = +2.99 V)
>> +3.3V: +3.33 V (min = +0.00 V, max = +4.39 V)
>> fan1: 889 RPM (min = 0 RPM)
>> fan2: 0 RPM (min = 0 RPM)
>> fan3: 0 RPM (min = 164 RPM) ALARM
>> temp1: +39.2°C (low = +65.0°C, high = +85.0°C) ALARM
>> (crit = +100.0°C, hyst = +98.0°C)
>> Board Temp: +33.8°C (low = +20.0°C, high = +60.0°C)
>> (crit = +100.0°C, hyst = +96.0°C)
>> temp3: +39.2°C (low = +80.0°C, high = +105.0°C) ALARM
>> (crit = +136.0°C, hyst = +132.0°C)
>>
>> The output needs to be configured (via /etc/sensors.d/chipname.conf),
>> but that's not a problem.
>
> Looks good indeed. I'd guess in1, fan2 and fan3 aren't used so you can
> add an "ignore" statement for them. The rest looks good, except the low
> limits of temp1 and temp3.
I'm attaching my adt7473.conf if anyone needs it - I'm not sure if it's
suitable for lm-sensors wiki - it's you to decide. Now this sensor
output looks this way:
adt7473-i2c-3-2e
Adapter: NVIDIA i2c adapter
+3.3V: +3.32 V (min = +2.96 V, max = +3.61 V)
GPU Fan: 876 RPM (min = 0 RPM)
GPU Core: +48.2°C (low = +20.5°C, high = +92.5°C)
(crit = +107.5°C, hyst = +105.5°C)
GPU Ambient: +35.2°C (low = +20.0°C, high = +60.0°C)
(crit = +100.0°C, hyst = +96.0°C)
>
>> (...)
>> I apologize - it's my double blunder - HWMon shows Winbond W83627DHG
>> chip. So I tried w83627dhg driver and it perfectly works on my system
>> when loaded this way (as you suggested in a previous letter):
>>
>> modprobe w83627ehf force_id=0xa020
>>
>> w83627dhg-isa-0290
>> Adapter: ISA adapter
>> Vcore: +0.82 V (min = +0.00 V, max = +1.74 V)
>> in1: +1.85 V (min = +0.69 V, max = +1.91 V)
>> AVCC: +3.39 V (min = +2.00 V, max = +0.56 V) ALARM
>> VCC: +3.39 V (min = +2.86 V, max = +1.76 V) ALARM
>> in4: +0.05 V (min = +1.53 V, max = +1.94 V) ALARM
>> in5: +1.71 V (min = +1.32 V, max = +1.86 V)
>> in6: +0.05 V (min = +1.66 V, max = +1.90 V) ALARM
>> 3VSB: +3.49 V (min = +1.50 V, max = +2.99 V) ALARM
>> Vbat: +3.34 V (min = +3.71 V, max = +3.74 V) ALARM
>> fan1: 0 RPM (min = 217 RPM, div = 64) ALARM
>> fan2: 1004 RPM (min = 397 RPM, div = 32)
>> fan3: 0 RPM (min = 357 RPM, div = 64) ALARM
>> fan4: 0 RPM (min = 84 RPM, div = 128) ALARM
>> fan5: 0 RPM (min = 703 RPM, div = 16) ALARM
>> temp1: +33.0°C (high = -13.0°C, hyst = -46.0°C) ALARM sensor =
>> thermistor
>> temp2: +36.0°C (high = +80.0°C, hyst = +75.0°C) sensor =
thermistor
>> temp3: +127.5°C (high = +80.0°C, hyst = +75.0°C) ALARM sensor =
>> thermistor
>> cpu0_vid: +0.000 V
>
> From the above output, it's hard to claim it works "perfectly". Some
> values look good, yes, but more careful inspection would be needed to
> confirm that all inputs and limits work they way they are supposed to.
In the meantime I cannot find the right formula for +12V which is shown
by Windows utilities, so probably I need to wait for a dedicated nuvoton
driver - alas, it only exists as a patch and I've no idea how to compile
and test it. The next strange thing is that in this output only VCore,
in1, in4, in6 and 3VSB change their values, all other sensors remain the
same no matter what my load is. Temp2 temperature which is theoretically
a CPU temperature changes much slower than coretemp indicates.
>
>> However this means two things, w83627ehf driver needs to be updated to
>> work with my PC (I can send any system information if it's required)
>
> Correct.... We must add support for your chip.
>
>> and sensors-detect script should also be modified to correctly
>> identify my sensors (it missed coretemp driver altogether).
>
> No, sensors-detect does the right thing at the moment. It properly
> identifies your Super-I/O chip (which is _not_ a W83627DHG so we just
> can't claim it is). As for the coretemp driver, the driver currently
> doesn't support your CPU so we can't point the user to it. If we did,
> we would receive complaints like "sensors-detect told me to load driver
> coretemp but it doesn't work." So we have to do changes in the right
> order, that is, first add support to the coretemp driver, and then
> change sensors-detect to point to it.
>
> In the meantime, please attach the contents of /proc/cpuinfo for the
> records.
I've filed a bug report on this topic:
http://bugzilla.kernel.org/show_bug.cgi?id=15348
Anyway, here's the information you requested:
$ cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 37
model name : Intel(R) Core(TM) i5 CPU 650 @ 3.20GHz
stepping : 2
cpu MHz : 1200.000
cache size : 4096 KB
physical id : 0
siblings : 4
core id : 0
cpu cores : 2
apicid : 0
initial apicid : 0
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 11
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx
rdtscp lm constant_tsc arch_perfmon pebs bts xtopology nonstop_tsc
aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3
cx16 xtpr pdcm sse4_1 sse4_2 popcnt aes lahf_lm ida arat tpr_shadow vnmi
flexpriority ept vpid
bogomips : 6626.04
clflush size : 64
cache_alignment : 64
address sizes : 36 bits physical, 48 bits virtual
power management:
>
>> coretemp driver started working after I applied the attached patch,
>> however the readings are wrong, so probably this driver is not yet
>> capable of monitoring Intel Core i5 650 CPU (temperatures are all wrong,
>> there are only two physical CPU cores):
>>
>> coretemp-isa-0000
>> Adapter: ISA adapter
>> Core 0: +23.0°C (high = +84.0°C, crit = +100.0°C)
>>
>> coretemp-isa-0001
>> Adapter: ISA adapter
>> Core 1: +26.0°C (high = +84.0°C, crit = +100.0°C)
>>
>> coretemp-isa-0002
>> Adapter: ISA adapter
>> Core 2: +23.0°C (high = +84.0°C, crit = +100.0°C)
>>
>> coretemp-isa-0003
>> Adapter: ISA adapter
>> Core 3: +26.0°C (high = +84.0°C, crit = +100.0°C)
>
> This is the reason why I am waiting for the Intel folks to give us the
> details. Apparently your Windows application is reporting temperatures
> 10 degrees higher, so I would guess that it assumes a critical
> temperature limit of 110°C for this CPU model (current temperature is
> reported relative to the critical limit.) But I wouldn't really assume
> they know it for sure either, given how hard it was to get the
> information for the other CPU models.
I have found out that TJMax for this CPU is 105C.
>
> As for the core count, the coretemp driver will show one device for
> each separate CPU entry. If you have 4 entries in /proc/cpuinfo, you
> get 4 coretemp devices. If this is a dual core CPU then I have no idea
> why you have 4 entries in /proc/cpuinfo.
This CPU has only two physical cores and it has HT (hyperthreading)
enabled, thus it shows itself as four cores for an operating system.
However I have a firm belief that HT "cores" should be omitted in
coretemp sensors output. I tried disabling the third and the fourth
cores using this configuration
chip "coretemp-*"
ignore temp3
ignore temp4
and even
ignore core3
ignore core4
but it doesn't seem to work.
>
>> Thank you very much, Jean, for your invaluable help.
>
> You're welcome.
>
# lm_sensors configuration for adt7473 chip on NVIDIA GeForce 8800GT GPU by Artem S. Tashkinov
# Requirements:
# * loaded NVIDIA binary module;
# * echo adt7473 0x2e > /sys/class/i2c-adapter/i2c-3/new_device (adjust for your environment)
chip "adt7473-*"
# Voltages
ignore in1 # Always shows 3.0V here, let's omit it
# Temperatures
label temp1 "GPU Core"
label temp2 "GPU Ambient"
ignore temp3 # Show the same temperature as temp1 sensor
compute temp1 @+7.5, @-7.5
compute temp3 @+7.5, @-7.5
set temp1_min 20
set temp3_min 20
# Fans
label fan1 "GPU Fan"
ignore fan2 # My 8800GT GPU has the only fan
ignore fan3 # Likewise
_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
