Jump to content

CPU temperature in iStat


  • Please log in to reply
207 replies to this topic

#121
oldnapalm

oldnapalm

    InsanelyMac V.I.P.

  • Moderators
  • 6,858 posts
  • Gender:Male
  • Location:Brazil

Is there a solution to the static temp readings? I get these with Oldnapalm or Netkas version of kext.

I have added Sbus and EC to DSDT, they both show up in Ioregexplorer although I haven't added all for support of these devices bits yet, causes a lot of compile errors if I try to. (as in Masterchief DSDT)

Still 38 degrees always shown for both cores, I would really like to know what temp my CPU is at to check the speedstep is working right.


Use VoodooMonitor to check speedstep, it shows temp, voltage, frequency and multiplier.

Run VoodooMonitor, open a terminal and run the command "yes", the values should go up (Ctrl+C to interrupt).

#122
scwhar

scwhar

    Scwhar

  • Members
  • PipPipPip
  • 112 posts
Thanks, I have tried this though, I mean speedstep is definitely working, it steps with yes, I've seen most of the 10 p-states but temperatures are also static in voodoomonitor.

Just would like to know if it is really keeping the CPU temps in check as I have heard sometimes CPU runs hot with applecpumngmnt, I have appleLPC loaded too so I think its probably OK, but it would be nice to know for sure.

#123
oldnapalm

oldnapalm

    InsanelyMac V.I.P.

  • Moderators
  • 6,858 posts
  • Gender:Male
  • Location:Brazil
Did you try with Bresink's Temperature Monitor?

#124
scwhar

scwhar

    Scwhar

  • Members
  • PipPipPip
  • 112 posts
Yes tried that same as others, 38c static, and 9728c static for CPU diodes, I think the HD smart temps are dynamic at least. Had a nose round see if it might be connected to a bios settting, but it doesn't seem to be.

#125
bcc9

bcc9

    InsanelyMac Legend

  • Coders
  • 1,281 posts
  • Gender:Male

Hi bcc9, thanks for the tip.
Please send this info to netkas if you have other way than writing in this thread. I don't have his e-mail, registration is disabled in his forum, and his PM box is full. I also have a question but don't know how to contact him (put it in first post).
Thanks again.

Yes, there seems to be no good place to have open technical discussions on hackintosh topics.

Looks like my _CRT idea wouldn't work for all platforms (may not be set or may be set wrong on some platforms).
So I'd probably recommend a static table for the common cpus, and _CRT as a backup check if the cpu type is not known and _CRT is present & seems sane.

#126
scwhar

scwhar

    Scwhar

  • Members
  • PipPipPip
  • 112 posts
OldNapalm - had another go at your suggestion, yes was only giving me 70-90% cpu usage so I ran geekbench at the same time for 100% load (doesn't help the rating) - and lo and behold the temps in voodoomonitor and istates do change - up to 45c for a moment on one core.

Like you said the temps drop after a few moments in istat menus but the readings in vm (which matched) drop instantly - makes me very suspect that these readings are to be trusted at all on this computer.

EDIT: alright been lookingf into it, just cos its about all there is left to look into on this computer now, turns out that realtemp (used this with Prime95 test), speedfan and some other prog on windows all give the same sort of readings, 38c unless really stressed, maxing 45 and dropping really quick when load removed (drop gradually if load reduced gradually, P95 test). Asus probe is the only one with different results, diff reading CPU not each core, but this is 6-8c lower and drops slower. So I still don't quite trust the reading but they obviously are what CPU reports, so thanks Oldnapalm, good work.

#127
oldnapalm

oldnapalm

    InsanelyMac V.I.P.

  • Moderators
  • 6,858 posts
  • Gender:Male
  • Location:Brazil

turns out that realtemp (used this with Prime95 test), speedfan and some other prog on windows all give the same sort of readings, 38c unless really stressed, maxing 45 and dropping really quick when load removed (drop gradually if load reduced gradually, P95 test). Asus probe is the only one with different results, diff reading CPU not each core, but this is 6-8c lower and drops slower.


Did you try Mark-i? You should have the same result as with Asus probe, it gets the temperature from a sensor on the mobo, not from the CPU itself.

#128
scwhar

scwhar

    Scwhar

  • Members
  • PipPipPip
  • 112 posts

Did you try Mark-i? You should have the same result as with Asus probe, it gets the temperature from a sensor on the mobo, not from the CPU itself.


No didn't try it, maybe next time I'm in windows I will as that reading seems a bit more believable.. just don't believe that cores always at 38 never lower hardly ever higher.. but it seems the cores reading would work to let you know if they're getting way too hot so not really important, I'm not fussed about knowing exactly what temps so long as its safe.

Read somewhere that its possible to get this mobo (and other) sensor reading out of DSDT, might be a nice project but beyond me for now and presumably vendor or chip specific so wouldn't help too many people..

Thanks again

#129
mitch_de

mitch_de

    InsanelyMacaholic

  • Retired
  • 2,902 posts
  • Gender:Male
  • Location:Stuttgart / Germany
Get that be also an way to read out GPU temp diode ?
Has anyone access to his GPU temp (even without that fakesmc)
I had in the past GPU temp shown of my old NV 6600GT (Hardwaremonitor) running 10.4.11, but never got that back since 10.5.x (lost of dsdt changes , new OS X version,...)

#130
Master Chief

Master Chief

    Just Chief

  • Members
  • PipPipPipPipPipPipPipPip
  • 1,176 posts
  • Gender:Male
@oldnapalm,

Who gave you this code snippet:
#define bit(n)		  (1UL << (n))
#define bitmask(h,l)	((bit(h) | (bit(h) - 1)) & ~(bit(l) - 1))
#define bitfield(x,h,l) (((x) & bitmask(h, l)) >> l)
Now look at this. Credits? I mean knowing you... you may want to correct this :P

#131
bcc9

bcc9

    InsanelyMac Legend

  • Coders
  • 1,281 posts
  • Gender:Male

Who gave you this code snippet:

#define bit(n)		  (1UL << (n))
	 #define bitmask(h,l)	((bit(h) | (bit(h) - 1)) & ~(bit(l) - 1))
	 #define bitfield(x,h,l) (((x) & bitmask(h, l)) >> l)
Now look at this. Credits? I mean knowing you... you may want to correct this :D

...which in turn is lifted from osfmk/i386/cpuid.c in the OSX kernel source. With an OS that fails to make a good SDK to access the cpu state, it's understandable that you have to do a little cut&paste since there isn't a proper library to reference. No person to credit here since it's from the OS. I'd still reference Apple's OSF copyright if I was the one cutting&pasting. Amit should have done that instead of just writing "Reuses code from the Mac OS X kernel" I think.

Voodoohda, voodoobattery, are guilty of lifting a bunch of code from OSX as well without reference. If everyone provided references the code wouldn't seem like voodoo, huh? Seriously, it would be nice if people sighted the apple code they cribbed from, it'd make the code more maintainable. (People would know where you found out how to do things).

#132
oldnapalm

oldnapalm

    InsanelyMac V.I.P.

  • Moderators
  • 6,858 posts
  • Gender:Male
  • Location:Brazil
That code snippet was added by Mojodojo (referred as from Markswell's utilities).

Thanks Master Chief and bcc9 for pointing out. Added a reference in the first post.

#133
Master Chief

Master Chief

    Just Chief

  • Members
  • PipPipPipPipPipPipPipPip
  • 1,176 posts
  • Gender:Male

...which in turn is lifted from osfmk/i386/cpuid.c in the OSX kernel source.

It should have been clear that I used the link as a reference to the original source code because it reads: "Reuses code from the Mac OS X kernel." and since I know that oldnapalm is clever enough to Google for it. He checked it – after my little note. Most definitely.

With an OS that fails to make a good SDK to access the cpu state, it's understandable that you have to do a little cut&paste since there isn't a proper library to reference. No person to credit here since it's from the OS. I'd still reference Apple's OSF copyright if I was the one cutting&pasting. Amit should have done that instead of just writing "Reuses code from the Mac OS X kernel" I think.

He should have used the Apple Software License Header yes. Why else do they have a license?

Voodoohda, voodoobattery, are guilty of lifting a bunch of code from OSX as well without reference. If everyone provided references the code wouldn't seem like voodoo, huh? Seriously, it would be nice if people sighted the apple code they cribbed from, it'd make the code more maintainable. (People would know where you found out how to do things).

I couldn't agree more. Making it a perfect target for "learning by example".

And oldnapalm knows me, that I wasn't blaming him for the error, but merely pointing him into the right direction :lol:

#134
Zakaria

Zakaria

    InsanelyMac Protégé

  • Members
  • Pip
  • 30 posts
Guys so I just have to replace the fakesmc.kext? I did that and my temperature is stuck at 66 degrees celcius. Im beggining to worry because this is not a gd sign.

Im running E6300 @ 3.0ghz on 965P DS3 rev 1.0

Im using a coolermaster hyper 212 plus heatsink.

Temperatures on Temperature monitor is usually 40-50. In istats it shows 66 there all the way.

Did I do anything wrong or is this the actual reading coz as of now Im running under aircon too! O.O

#135
oldnapalm

oldnapalm

    InsanelyMac V.I.P.

  • Moderators
  • 6,858 posts
  • Gender:Male
  • Location:Brazil

Guys so I just have to replace the fakesmc.kext? I did that and my temperature is stuck at 66 degrees celcius. Im beggining to worry because this is not a gd sign.

Im running E6300 @ 3.0ghz on 965P DS3 rev 1.0

Im using a coolermaster hyper 212 plus heatsink.

Temperatures on Temperature monitor is usually 40-50. In istats it shows 66 there all the way.

Did I do anything wrong or is this the actual reading coz as of now Im running under aircon too! O.O


This patch was added to the latest version of FakeSMC, so if you use it you don't need to replace it with this kext.
http://netkas.org/?cat=15

Did you set the Tjmax value in Info.plist? (It's the "specified upper limit" in Temperature Monitor).

About temp being stuck, please read some recent posts about it
http://www.insanelym...p...t&p=1341033

#136
knobsi

knobsi

    InsanelyMac Protégé

  • Members
  • PipPip
  • 58 posts
is it possible to use it in any way with a pentium d 920 processor?

#137
jbolted

jbolted

    InsanelyMac Protégé

  • Members
  • PipPip
  • 71 posts
  • Gender:Male
  • Location:New York
  • Interests:Computer building, hacking, reading, fishing, movies. Learning things I can't figure out.
I just put together a build. Snow Leopard, I installed the istat menus. On CPU temp I was getting 138 degrees on all 4 cores. Is this possibly a wrong reading or am i ready to burn out my new cpu or worse.

Any hep appreciated.

#138
oldnapalm

oldnapalm

    InsanelyMac V.I.P.

  • Moderators
  • 6,858 posts
  • Gender:Male
  • Location:Brazil

is it possible to use it in any way with a pentium d 920 processor?


Sorry, I don't have one to test, but AFAIK it only works with Core 2 Duo or later.

I just put together a build. Snow Leopard, I installed the istat menus. On CPU temp I was getting 138 degrees on all 4 cores. Is this possibly a wrong reading or am i ready to burn out my new cpu or worse.

Any hep appreciated.


138 degrees Celsius or Fahrenheit? You should also set the correct Tjmax for your CPU.

#139
Akutsito

Akutsito

    InsanelyMac Protégé

  • Members
  • PipPip
  • 53 posts
Works great ;), M1530 with T5450 and no major issues, changed the SMC ID to match what im supposed to mimic (Macbook pro 3.1), fixed the tjMax (85), and everything seems fine :).

#140
the_gunners

the_gunners

    InsanelyMac Protégé

  • Members
  • Pip
  • 12 posts
It's not working in my SL, still not showing CPU option only HD.

Do I need to edit something or just install the updated FakeSMC.kext ?





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users

© 2014 InsanelyMac  |   News  |   Forum  |   Downloads  |   OSx86 Wiki  |   Mac Netbook  |   PHP hosting by CatN  |   Designed by Ed Gain  |   Logo by irfan  |   Privacy Policy