Printable View
My temperatures in the BIOS:
30,29,28,32
Real Temp reading:
http://img228.imageshack.us/img228/8...0110131712.png
Btw,
respect to everyone,as Sandy Brigde is already out,monitoring software are still on X58 stage(even AIDA64,CPU-Z),only HWiNFO32 v3.66-1087 Beta follow up.:up:
stasio did some thorough testing and has discovered that AIDA64 is the reason why RealTemp 3.60 is having problems on the new Sandy Bridge CPUs. If you don't run AIDA64, including the side bar, RealTemp 3.60 works fine on Sandy Bridge.
AIDA64 is using the shared performance counters within Sandy Bridge CPUs in an unfriendly manner and is continuously stopping or resetting them when it is supposed to leave them running so as not to interfere with other monitoring software. I can't fix AIDA64 so my plan is to switch RealTemp to a new set of high performance counters within the CPU that are hopefully not being misused by AIDA64. I'll try to get a new version of RealTemp uploaded later today.
Edit: Here's an insanely huge 134 MB AVI video that shows what happens to RealTemp 3.60 as soon as AIDA64 starts running.
http://www.sendspace.com/file/5agwlj
Everest has had this problem since the beginning of the Core i7 era as far as I know and now AIDA64 is the new name for this problem. When monitoring programs stop or reset any of the shared timers within an Intel CPU, it can create havoc for any other monitoring program that is also trying to use these timers.
In the video you will see the multiplier dropouts that RealTemp reports as soon as AIDA64 is started. These happen every 5 seconds which is exactly how often AIDA64 is sampling these timers. As soon as AIDA64 is stopped, the multiplier dropouts start and RealTemp is 100% accurate again.
This forced me to switch RealTemp 3.65 to use a different set of monitoring timers. Hopefully no software is abusing these new timers that RealTemp is now using.
I love this temp program, and just got a q9300 upgrade for my Windows Home Server. The problem is I think the chip is running hot. I have 2 120 fans (one front -intake, one rear exhaust), plus the 120 on the psu. I also have a Zalman 9500LCD cooler running at a round 2200RPM. My server is in a cabinet with the door left open and the cabinet temp is around 67-70 F. I just put this in 2 days ago, and it seems to be quite high on the temps. I have a AMD quad 630 and it never goes over 50c under full load (same cpu cooler). I also have a Q8400 (same cooler) and it idles in the high 20's and 100% load maxes never over 60c. graph of the Q8400 and Q9300 below.
So when i saw this Q9300 this high(60's) i thought maybe i did something wrong with installing the cooler or something. So can you guys that understand the charts get back to me on what is going on with the high temps. I would really appreciate it. I am not OC this, just want the chip to last a long time in the server. Thanks in advance guys!
any update yet for use on the sandy bridges?
RealTemp 3.65
http://www.mediafire.com/?f4cybi6zf06d1cb
http://img266.imageshack.us/img266/1852/rt365.png
I switched RealTemp so it uses a different set of timers in the CPU for better compatibility with some of the other rogue monitoring programs out there and VID voltage monitoring has been added for the new Sandy Bridge CPUs.
Edit: Thanks PolRoger for the picture. :up:
The voltage that RealTemp 3.65 displays seems to be a VID voltage and is not actual core voltage so don't get too excited.
bryanchewy: Intel's 45nm temperature sensors used on their Core 2 CPUs can be completely inaccurate with huge amounts of error in their results. For that reason alone, you can't compare the data they display to any other CPU, even CPUs of the exact same model can have completely different amounts of error in their temperature sensors. It's often times more like random data. You especially can't compare them to an AMD CPU which positions its temperature sensors completely differently than Intel does.
As long as your CPU is running stable, there's no need to be too concerned about its core temperature. It's just a number and unfortunately not a very accurate number on the 45nm Core 2 based CPUs. I find it is consistent from day to day but because of the 101 issues, it's not comparable to anything.
Thanks uncle,speed OK now. :up:
Much thanks, uncle!
Thanks, 3.65 works great. I havent figured out the setting to show VID though.
Edit: found it.... doh.
I'm still not sure whether the voltage number stored in these new Sandy Bridge CPUs is like a Core 2 VID type voltage or if it is an actual voltage. I've seen some boards where this number looks like actual voltage but on other boards it is too low to be actual. Someone will have to get the multimeter out to find out what this new number means.
This is the voltage number that RealTemp and Core Temp are both reporting.
http://img341.imageshack.us/img341/5365/msr0x198.png
thanks for the update Uncle ... :)
thx for update!
Thanks for great software!
I tried 3.60 on Vista x64, 3.60 would not show the temps of My video cards(a 6200 and a GTX295), Yet 3.58 does, So I went back to 3.58.
zoom314: Did you try checking the Nvidia box in the Settings window? GPU monitoring is optional in the recent versions of RealTemp so maybe that's the problem. Nvidia and ATI temperature monitoring code is not the most efficient stuff so I thought making it optional would be the best solution. If it still doesn't work, let me know what happens.
Huh? settings window, Is that the Nvidia control panel? If It's not, please advise, As that is as clear as mud to Me, As I just installed Vista Business x64 on Monday.Quote:
Originally Posted by unclewebb
Ok, I think You meant the 3.60 settings window, the gpu button in 3.60 is disabled, So no dice. I'd provide a screen shot, But My normal video software is not installed under Vista x64 yet.
http://img26.imageshack.us/img26/6449/rt360.th.png
What happens when you put a check mark in the Nvidia box?
If the box goes gray and nothing happens that's a problem.
Oopsie, I didn't see that, My bad, That Nvidia fixed that, I must be more tired than I thought I was.Quote:
Originally Posted by unclewebb
Any updates to RealTempGT as part of the 3.65 bundle?
RealTempGT doesn't need Sandy Bridge support at the moment so I won't be working on it quite yet. I learned a few interesting things about monitoring power consumption in Sandy Bridge as well as correctly adjusting the Sandy Bridge power limits so that's going to come first.
Both Realtemp and Aida64 both freeze on me running Prime95 after about half an hour although Prime95 keeps chugging along. If I use OCCT large, Realtemp freezes immediately and stops responding. Am I the only one with this issue on a 2600K? I do not run Realtemp and Aida64 at the same time. Any idea why Realtemp stops responding?
I have to use the Asus beta BIOS 1053 to boot into Windows at anything above 4.7 and then I can run 5.0 stable at 1.425, but with all my temp programs freezing, I don't have any idea what the temps are. Before it freezes, my highest temp is about 65c. Another bad thing about the Asus Beta is that I cannot run my memory over 1600 MHz or it won't even boot. I can do 4.7+ with memory at 2133 or 5.0 with memory at 1600.
I have an As.s board. I can't use their software to monitor anything - it screws up the interfaces for everything but itself.
I had random freezing problems recently when checking out AIDA64 on an older Core i5-650 CPU. I had no freezes before installing AIDA64 and I've had no freezes since removing both versions that I tried.
Try completely removing AIDA64 including the side bar gadget. I can't think of anything that RealTemp is doing any differently than what it has been doing for a long time without any issues.
Daddyjaxx: Are you running RealTemp 3.65?
It is definitely Aida64. I just reinstalled Windows because I was having all kinds of flakiness from switching to the P67 chipset from an X58 backup. On a fresh install at stock, the computer froze on me. When I restarted, the Aida64 side gadget was stuck where it was when it froze. Uninstalled it and no more freezes. I am using 3.65. I'm going to let Prime run now and see if I get the same wierdness. I don't think I will because I used 3.60 on a fresh install with the same CPU and motherboard before I put it in the case and restored the backup.
RealTemp 3.67
http://www.mediafire.com/?n99nq4kn95u6i6a
http://img257.imageshack.us/img257/8229/rt367.png
-added Sandy Bridge VID / power consumption monitoring.
-added Sandy Bridge power limit adjustments if they are not locked by the bios.
-added ability to save the log file to a network computer.
-added Disable Turbo (IDA) option for Core 2 mobile CPUs.
-updated i7 Turbo GT program for Sandy Bridge CPUs.
-also updated the RTCore.dll RivaTuner plugin.
Thanks to burebista and somebody for all their help testing.
Daddyjaxx: Thanks for confirming that AIDA64 is also causing you some random lock up problems. Without AIDA64 installed, the computer I was using seemed stable. I've heard this issue might be memory voltage related but I lost interest trying to debug this without doing any thorough testing.
Too busy working on my own projects. :)
Thanks Uncle ! Works good in RT.