Amd Dual Core Temp differ

G

getSIDEWAYS

n00b
Joined
Dec 12, 2006
Messages
46
well my AMD 5200+ is running great but i noticed that core# 1 runs almost 10c hotter than core # 2 at an idle, does this mean i didn't put my heat sink on right? im using everest to get this temps
 
Once again. Coretemp or Intel TAT are the only reliable programs for temperatures.

Avoid others like icecubes in Mexico.
 
getSIDEWAYS said:
well my AMD 5200+ is running great but i noticed that core# 1 runs almost 10c hotter than core # 2 at an idle, does this mean i didn't put my heat sink on right? im using everest to get this temps
Click to expand...

If it makes you feel any better my core temp readings are 38 and 22. A whole 16 degrees off. I know it's not the heatsink mounting. I've done it three diff times since I had the chip. This is like the third case I've seen of the temp differences on this forum. In another thread someone posted a quote about how some of the chips could have been shipped with bad and or inacurate temp probes. Still it could be a bad mounted IHS, If it's not what do we do?
 
right now core #0 is running at 29C to 30C and Core # 1 is running 35 to 36C using core temp
should i pull my heatsink??? and im using a Zalman CNPS9700 LED
 
Nice temps... those Idle?

Those are FINE. A reseat might help, but it's not like you're in a danger zone.
 
yeah at an idle, i may reset it just to see if theres a difference but i used artic sliver 5 and haven't had the computer running more than 5 days so ill see if it runs any lower after it sets all the way. thanks for the help
 
Using Core Temp here and it's telling me 41-43C on core #0 and 26-28C on core #1 at idle. That's a 13-17C difference. Abit EQ, which came with my motherboard but cannot distinguish the two separate cores, reads 36C. Although I admit to using stock cooling, I do not have any reason to think the heatsink is mounted incorrectly as I just built this last week and have a good memory of everything going smoothly with the installation. I am wondering if this is somewhat normal for dual core AMD chips, especially since it seems like others are seeing the same thing?
 
Sta!nless said:
someone posted a quote about how some of the chips could have been shipped with bad and or inacurate temp probes. Still it could be a bad mounted IHS, If it's not what do we do?
Click to expand...

^^^^^^^ Yeah see...so I'm wondering which core we should go by?
 
Are you using two instances of Prime95 to torture test, to make sure one Core is not doing all the work?

My two cores are only about 4-5C apart when using 2 X Prime95 to test load temps. At idle they are 1C apart, so I don't think it's a temperature probe inaccuracy.
 
Well for me as I type this and the chip is pretty much idle, my temps are 40C and 20c. I'm on water with no IHS, and room temps are 70F... If the colder core Was read in Fahrenheit, it'd be 64 degrees. Like I said room temps are 70.. So what's reporting what wrong? I guess either way I'm glad it doesn't load up past 48C.

With both cores loaded 100% each with prime, they get closer. Closer as in 10C apart, rather then 20.

I had this problem with the IHS on, and on air with a xp90. I mounted the xp90 twice. And mounted twice on water.. For my case you can't blame the mount.

Too bad I can't send the chip back and say it's defective because it doesn't report temps right.
 
Sta!nless said:
Well for me as I type this and the chip is pretty much idle, my temps are 40C and 20c. I'm on water with no IHS, and room temps are 70F... If the colder core Was read in Fahrenheit, it'd be 64 degrees. Like I said room temps are 70.. So what's reporting what wrong? I guess either way I'm glad it doesn't load up past 48C.
Click to expand...

There's no way a 1 inch square piece of metal and silicon can have a 20C differential, it must be a faulty reading.
 
Yes.. I just took a screenshot of coretemp. Check it out. I get about the same readings with everest too.
 
Sta!nless said:
Yes.. I just took a screenshot of coretemp. Check it out. I get about the same readings with everest too.
Click to expand...

18*C??? That's 64.4*F. I'm willing to bet that your ambient is higher and that's inaccurate.
 
I think CoreTemp is just crap - because when I turn on Prime95, my Coretemp readings shoot up 10C instantly. Turn it off, they shoot down 10C instantly.

No way that is real, and since the sensors are reading voltages, probably a voltage spike causes this when the CPU is suddenly loaded.

Both SpeedFan and Epox's USDM report CPU temps that i think are much more meaningful.
 
Varmint said:
I think CoreTemp is just crap - because when I turn on Prime95, my Coretemp readings shoot up 10C instantly. Turn it off, they shoot down 10C instantly.

No way that is real, and since the sensors are reading voltages, probably a voltage spike causes this when the CPU is suddenly loaded.
Click to expand...

I have noticed this too... I do understand that Coretemp is using the sensors inside the cores themselves instead of relying on motherboard sensors... but temperature swings so dramatic, makes me have doubts.

Another thing I noticed was that CORE 0 would display a NEGATIVE temp whenever I start up BF2.... Alt-tabbed out of BF2 after the map is loaded, and coretemp still shows a negative temp for Core 0. Issue disappears when BF2 is shut down... core 0 temp returns to normal. Very odd.

Sta!nless said:
Well for me as I type this and the chip is pretty much idle, my temps are 40C and 20c. I'm on water with no IHS, and room temps are 70F... If the colder core Was read in Fahrenheit, it'd be 64 degrees. Like I said room temps are 70.. So what's reporting what wrong? I guess either way I'm glad it doesn't load up past 48C.
Click to expand...

Wow... big temp difference for a no load situation. My temps are only apart by 1-2c when at idle, but under load, Core 1 starts to pull away and I end up with a 12-15c difference between cores.
I think the IHS will be taken off this weekend if that doesn't help the difference in temps between the cores... at least it'll lower temps a few degrees
 
I have a 3 degree difference at idle, and 6 at load.
I too was wondering about the way the temp rises so fast. When I shut off orthos it drops back to idle temps in .05 seconds lol.
 
tims880 said:
I have a 3 degree difference at idle, and 6 at load.
I too was wondering about the way the temp rises so fast. When I shut off orthos it drops back to idle temps in .05 seconds lol.
Click to expand...

I'm guessing the sensors read by CoreTemp are calculating temperature based on voltage across circuits rather than measuring it directly. That's why it shoots up suddenly when you increase the load. Or something like that.

In any case, I think the TCaseMax rating limit is for the metal case of the CPU (reported as 'CPU Temp' in Speedfan, the BIOS etc), not for the internal measurement that CoreTemp is supposedly reading.
 
NordicRX8 said:
Another thing I noticed was that CORE 0 would display a NEGATIVE temp whenever I start up BF2.... Alt-tabbed out of BF2 after the map is loaded, and coretemp still shows a negative temp for Core 0. Issue disappears when BF2 is shut down... core 0 temp returns to normal. Very odd.
Click to expand...

This must have been a glitch in my system, even though BF2 ran fine (FFOLKES unlocks678) on numerous occasions. After a long needed reboot, core temp is fine now.

NordicRX8 said:
Wow... big temp difference for a no load situation. My temps are only apart by 1-2c when at idle, but under load, Core 1 starts to pull away and I end up with a 12-15c difference between cores.
I think the IHS will be taken off this weekend if that doesn't help the difference in temps between the cores... at least it'll lower temps a few degrees
Click to expand...

Ran out of rental movies and crap on TV... so I thought I would just pop the lid off my 170 tonite. Got the IHS cut off, reinstalled the cpu, applied arctic silver, replaced storm block (also modified the mount while I was working on the computer), turned her back on, and boosted the FSB a tad (a mere 10MHz more caused a hard reboot in XP, :( I need to tweak this some more). Been running Orthos on it for last >50mins. Overall temps dropped by 7-8C, load temp difference between the cores is now only 1-2C! :D :D :D Load temps dropped by alot too.. CORE1 would hit 64C! (Tcase 59C), but now both cores hover around 45. This and my previous Opty, had really shitty IHS contact... good thing I'm willing to slice them up!!! 30 minutes of effort, $0.90 worth of parts and I'm very happy with the results! I'm also very glad these AMD IHSs aren't soldered on.
 
According to PC Probe II, both of my cores are running at 10C with my 9500 CNPS and about 5 fans.

However, my BIOS says 40C.

Whatever.
 
Xvash2 said:
According to PC Probe II, both of my cores are running at 10C with my 9500 CNPS and about 5 fans.

However, my BIOS says 40C.

Whatever.
Click to expand...

:eek: 10C is petty cool... I mean... like, you should wear a sweater cool. That doesn't sound right... see what core temp or speedfan (I think it supports dual cores) says.
 
Coretemp is off noticable at under 40c actuals on my system. The only time I can rely on it is when I'm doing a dual burn that is going 44c and higher. It's typical for coretemp to report below ambient temps when the closer you get to 20c actual on the cores (way undervolted cores, ie 0.9vcore downclocked @ 800mhz).
 
You must log in or register to reply here.
Back
Top