Q9550 idle so HOT... Help

[danimon]

i got my core2quad Q9550 2months ago with stock cooler,and run it@ stock speed..it idle @ 54-50C each core,and when running 3d rendering application,it bumps quickly to 84-82C @ peak..

that forced me to spend some money to buy vendetta2+xigmatek crossbow+OCZ freeze for what i tought would solve that problem..

but after installing them, the core temp still relatively hot..idle @ 46-43C..but the CPU temp just @ cool 30C
then for max temp i use prime95 n run small FFT test about 15minutes..
i got 59-59-59-58 @ core and 48C for CPU.. i just set all voltage to auto wich is about 1.14V max says the HW monitor..

once i OCed it to 3,4GHz,core temp went to 49-45C @ idle and 69-67C max(running prime95) with CPU temp@ 59C.it auto voltage recorded 1,22V @peak..

i'm guessing that those temp stil too high for Q9550 coz many people out there say that their 9550 idle temp never go beyond 40C @idle..so please help me to cool things down..

note that my room ambience temp is about 25C, n my case is a smaller one(maybe mid-tower).

 

[Redmist77]

Your Q9550 results are very similar to mine.

I'm currently running at 8.5x425 (3.6GHz) @ 1.23V and my idle temps are about 45C (according to 'Core Temp') and under load with Prime95, Core Temp reports about 85C for all cores. (room temp is about 26C)

My cooler is a Thermalright Ultra120 with a low airflow.

Don't believe most of the temperatures people report on the net - it's something that people seem to lie about. A lot of people who report them seem to live in Finland or Canada too...

 

[ravenware]

Feel the air around the CPU to see if it actually feels hot, maybe your getting a false reading.

 

[Inktfish]

Try using other programs like Realtemp and Speedfan to view if Coretemp gives you correct readings. Speedfan reports to me 20 celsius less than the Core and Realtemp monitors.

 

[danimon]

thanks redmist..u make me feel better
never thought about that before or ignoring that fact(pepole who live in finland,canada,or iceland).

@ravenware: actually i touch that vendetta2 when running @ peak temp..that's as cool as my skin tho..i suspect too much thermal paste when installing the cooler..note that i use that two small line method..what do you think guys?

@inkfish,i use 3 different temp program and they show identical result.my mobo(biostar i45) has own onboard "CPU meter" but the temp it shows similar to "ACPI THRM" rather than "core" temp. it always report much lower temp like 15-20C less than the core temp..

before i upgrade the cooler, i thought maybe i could gain more temp decrease coz from what i learned,vendetta2 is one of best bang cooler n OCZ freezer also great thermal compund.but for a 5-7C decrease @ idle( at around 50C point), it really frustating me..maybe i just need some report from u guys using same processor(or it's brother Q9x50) n same CPU cooler as mine or other 12cm stuff like HDT1283,TRUE..if so, dont forget to notice your ambient temp..

thanks so much for attention b4..

 

[Redmist77]

One thing that I've noticed is that my north bridge heatsink is incredibly hot. It's os hot that I can't even really touch it. It has a L-shaped heatpipe going from the north bridge to the heatsink near the CPU but it obviously doesn't seem to help much.

My CPU heatsink feels 'warm' under load but not 'hot'

 

[jsc]

I upgraded my Q6600 to a C1 Q9550 last Christmas vacation. (The Q6600 will go into my presently unused 680i motherboard.) The first thing that I noticed was that the heatsink that came with the Q9550 was half the size of the one that came with the Q6600. I could safely run the Q6600 at 3.0 GHz with the stock HSF. The Q9550 HSF is clearly no bigger than it has to be to run at stock frequencies.

Right now, the Q9550 is OC'd to 3.6 GHz. (425 X 8.5).

System specs: GA-EP35-DS3P, Q9550, OC'd to 3.6 GHz. Vcore at 1.275 volts, drooping to 1.25 volts. TRUE/S-Flex SFF-21F oriented vertically (warm air wants to rise, why fight it) in an Antec 900 case with the fans set to LOW. Ambient temp around 22 C.

Idle temps:
CoreTemp Ver 0.98.1: 42 C - 51 C.
RealTemp Ver. 3.00 : 35 C - 45 C.

Load temps - 4 instances of Prime95:
CoreTemp Ver 0.98.1: 60 C - 63 C.
RealTemp Ver. 3.00 : 55 C - 58 C.

 

[danimon]

@ remist: mine too.the northbridge heatsink was soooo HOT..btw the outer/top plate of vendetta2 never really warm even i stress the CPU @ 3,6GHz.that's why i thought it was the thermal paste messing the temp..but still not so sure about that..

@jsc:are u sure set the fan to LOW?that core temp just too good to be true.i mean that's pretty cooler than 85C redmist's TRUE. is it only the case fan or the TRUE fan too?
but your score seems to be more identical to mine, which mine was a little higher thanks to higher ambient temp..@3,6Ghz i got 44-48C idle and 66-68C @small FFT load in core temp.btw could you mention your thermal paste and the way u apply it?maybe could help me decide wether reapply it or not..
thanks b4 for posting ur result..

now it seems that 45C still reasonable idle temp for Q9550..this is what i'm looking to be clarified..

 

[jsc]

Yes, the case fans are set to LOW. Then Antec is pretty noisy with the fans set to MEDIUM. With the fans set to HIGH, it sounds like a vacuum cleaner setting on the desk.

CPU fan? I don't know. BIOS reports the HSF fan running at 1500 rpm. I cannot tell if it goes to high(er) speed. I usually have two computers running. The Q9550 is the gaming and multimedia machine. The other is an old AMD Barton core box that I use for a basic workhorse machine - internet, office apps, and so forth. It is a little louder than the new box.

I have had my TRUE for more than two years. Mine had a somewhat sloppy machining job on the base. It had a small, off-center hump on the base. I hand lapped the base down to bare copper finishing with 1500 grit. I teamed it with a Scythe S-Flex SFF-21F fan. The "F" model seemed to represent a good compromise between noise and moving air.

I had been using Arctic Silver, but this time I shifted to Ceramique when I upgraded from the Q6600 to the Q9550. As long as it's pretty good, I don't think the thermal paste you use makes much difference.

Application - I tried the "grain of rice" bit once back when. I clamped the HSF down, then pulled it up. I didn't cover the top of the heat spreader. What I do now is divide the top of the CPU mentally in quarters, then put a drop about 2/3's the size of a grain of rice in the center of each quarter and clamp the HSF down. I seem to get really good coverage that way.

My Q6600 at 3.6 GHz. with the same setup idled at 35 C - 40 C and ran 61 C - 65 C under load.

By the way, if anyone is wondering, you can upgrade a processor without removing the motherboard if you are using a TRUE. The support plate will stay in place when you remove the heatsink.

 

[Parashrew]

i have a q9550 and think i am having some temp issues. would someone please let me know what you think?

My cpu is at 37c and the 4 cores are as follows; 37 54 31 46...

is it just me or does it seem strange that cores 2 and 4 are much greater than the others and core 3 is so much lower??? i am dumbfounded, pleas help

thank you