Printable View
I have tried with the folder located in the root directory or on another drive, it doesn't seem to matter where it is. UAC is enabled and I have the program set to always run as administrator and listed as an exception in DEP. Please let me know if I can provide anything else.Quote:
Originally Posted by unclewebb
I'll do some Windows 7 testing later tonight to see if I can duplicate your problem. Have you tried testing with the RealTemp folder located on your Desktop or in some other sub folder of your account? Do you have Administrator privileges? I'll create a basic account to see if I can figure this out.
Some anti-virus programs like Trend Micro PC-cillin will also try to block the installation of the WinRing0 open source driver that RealTemp depends on to read the sensors.
I appreciate your looking into this. I had the exact same problems under Vista. I have an admin account and have tried having the folder in various directories and located on various drives but it just will not autostart for me.
@ uncle: just a tip, with the latest bete isnt' possible to view the min/max VID correct? can you put all the values in the next? right now i can view only the min VID (with the C1E enabled) but i need to know (also) the max VID without put my CPU @ fullload :)
i think this is a glitch like coretemp that sometimes tell us the uncorrect VID :)
edit: what about the movements? how i can understand these values?
KURTZ: In the latest beta there is the option to show the current VID on the main screen but I promise to add Min / Max VID back to the settings window. :)
I rarely hear anyone talking about Min / Max VID. Most users only compare Max VID.
During a sensor movement test, in theory, the sensors should move similar amounts. This test puts an equal load on each core so in theory the temperature of each core should change a similar amount. It was originally a quick way to check if a sensor is sticking. On a Quad if 3 sensors move 10 and one sensor only moves 2 then the one that moved the least amount is likely getting stuck at idle. The 45nm Core 2 sensors were bad for this. Do a full CPU Cool Down Test at whatever your normal overclocked MHz is and post the results and I'll try to explain to you what I see. Include any details about your setup.
century child: I just started up Win7 and I'm playing with different UAC and Administrator settings. I'll see what I can figure out. What CPU are you using and what MHz? I've seen some problems where a slower Dual Core was so over loaded when starting up that it was interfering with RealTemp from starting up. I might have to do some under clocking to simulate this.
@ uncle: i know the general theory (about the sensors), but i don't understand the difference from 15 and 0 movements ...
i try to explain here, for example:
0 movs: seems to be dead (the sensor obviously)
from 1 to 2: the sensors aren't reliable
from 3 to 15: they are reliable, BUT if i've a clocked system my sensors go around 15 (instead if i've a default cpu my sensors go around 3)
what do you think about? TIA
I have Administrator privileges on my account and enabled UAC at its Default setting.
When using UAC, the StartUp folder method seems to cause some issues so I decided to add RealTemp to the TaskScheduler.
Open up Task Scheduler and start by clicking on Create Basic Task... on the right hand side.
http://img18.imageshack.us/img18/3923/pic6m.png
Enter a name for the task, RealTemp, and then click on Next.
http://img210.imageshack.us/img210/9994/pic0.png
For Trigger, choose the option, When I log on.
http://img268.imageshack.us/img268/7531/pic1xkv.png
For Action select, Start a program and then click on Next.
http://img12.imageshack.us/img12/2452/pic2kcr.png
I dragged the RealTempBeta folder into the Program Files(x86) folder.
Click on Browse and tell it where it can find RealTemp.exe
http://img12.imageshack.us/img12/4752/pic3s.png
Click Next to go to the Summary window.
Before clicking Finish select, Open the Properties dialog for this task when I click Finish
http://img268.imageshack.us/img268/6409/pic4pxn.png
Now select, Run with highest privileges and click on OK and you're done.
http://img338.imageshack.us/img338/6465/pic5.png
Now reboot and see if it works.
I know this is a big hassle but when running UAC, it's the correct way to add items to your Start Up routine that need Administrator privileges. You will also have to do the above if UAC is enabled to get Core Temp to start correctly. Temperature monitoring programs need to run at this level to access the temperature sensors which are a protected resource. Maybe we can hang some of the blame on Intel as well since reading the core temperature shouldn't be rocket science or a dangerous activity needing Ring 0 access.
If you ever have problems where RealTemp seems to be running in the Task Manager but doesn't appear on your screen, Kill the RealTemp.exe process and then go into the RealTemp.ini file and delete the WindowXY= key. RealTemp stores the last position it was at in that key so if it gets screwed up there might be a problem. I noticed one issue when testing that if the RealTempBeta folder is located directly in the C: directory then this window position info gets ignored and RealTemp was starting up for me in the top left corner. I think that's another Windows 7 UAC induced feature. :D
Edit: The above method also works in a limited account but you will need to enter an Administrator password before RealTemp will be allowed to run at Start Up.
Edit #2: Richard was having a lot of problems with this method. RealTemp would show up in the Task Manager but he couldn't access it anywhere. He finally figured out that for the last step it is critical that you have "Run only when user is logged on" is checked.
If you follow the above method exactly then it should work with or without UAC enabled and whether you are in an Admin or Standard account.
KURTZ: The initial sensor test puts an equal load on each core for about 5 seconds and measures how much the core temperature changes during that 5 second period. If you are overclocking with lots of core voltage then your temperature is going to go up much higher on each core than if you are at default settings or under clocked. At 1600 MHz and 1.10 volts your sensors will barely move during this test because your CPU isn't creating a lot of heat at full load. A well overclocked Core i7 with the OEM cooler is going to create a small inferno so you will see huge numbers during the sensor movement test.
If you swap your cooler to a better one, the amount the temperature changes during this test should decrease showing you that your new cooler is working correctly. If the sensor movement increases after swapping a cooler then you either installed a crappy cooler or didn't do a good job installing it.
http://www.pctunerup.com/up/results/...13_CPULoad.JPG
http://www.pctunerup.com/up/results/...Load3.0MHz.JPG
thanks Uncle, what do you think about these shots?
I seem to be having a problem with Real Temp. It fails to read my TRUE processor speed. For more details :
http://i266.photobucket.com/albums/i...Truth/th_1.png
http://i266.photobucket.com/albums/i...Truth/th_2.png
http://www.xtremesystems.org/forums/...90#post3971290
Wishmaker: Something changed in Windows 7 which caused an issue for the MHz in RealTemp 3.00.
Try updating to RealTemp 3.30 and show a comparison of that and CPU-Z 1.52.
http://www.fileden.com/files/2008/3/...alTempBeta.zip
KURTZ: It looks like you've got some seriously borked sensors or RealTemp has a problem. Try downloading the latest version or RealTemp from the above link.
After you boot up and are at idle does it say HOT HOT in the Thermal Status area?
Run a SuperPI mod bench at 3000 MHz so I can have a look at your results.
1M is OK, 4M or 8M if you're bored. I just want to see if your CPU is taking a performance hit or if just the sensors are screwy.
I followed your instructions for adding RealTemp to the task scheduler and this time it worked. I must have overlooked something when adding it before. Thank you very much for your help.
i'm waiting for those shot Uncle :), however what about the 'clock modulation'? can i clock my cpu with this tool?Quote:
Originally Posted by unclewebb
also: why the core 1 has 4 movs and grabs the T correctly for all the % and the core 2 has 5 movs and fails @ 1.0% ... i think that also the core 1 must have a trouble ... or not?
check here:
http://img39.imageshack.us/img39/1690/testdqw.png
KURTZ: These sensors are far from perfect but your Q9550 has about the best sensors I've ever seen on a 45nm Quad. All 4 of your sensors work from 37C to 100C and beyond so be happy. :)
If air cooled you would have to run your computer in a colder environment to find the sticking point of your sensors. All of these sensors get stuck at some point if you run them cool enough. That's just the way they are.
Where's your previous E8400? RealTemp 3.30 RC12 had a big bug in it but I haven't heard of any issues or had any issues with the previous RC11. I'm in the middle of a RealTemp overhaul to clean up the code but I decided to enjoy the summer instead. Hopefully in the next few weeks I will be motivated to work on RealTemp again. I only work hard when someone pushes on the Donate button and that hasn't happened for a while. :rofl:
Edit: Clock Modulation will slow your CPU down internally. The more clock modulation you use, the more clock pulses are ignored by the CPU. That's why the temps drop when using this while fully loaded. Your CPU isn't working as hard internally even though it will continue to report the same MHz. This was invented by Intel during the Pentium 4 era for heat control. It's handy when testing or to keep a laptop from burning through your lap. It can also be used to extend battery life while watching a movie.
century child: I'm glad I could help. I think I'll go post a link to that tutorial in the first post so other users can find it. Core Temp has an option to add itself to the registry but when using UAC, I couldn't get it to start up that way. As far as I can tell, when using UAC, you have to use the Task Scheduler method.
:D i'm waiting the screen shot from my friend ... :)
however check here ... :D
http://img8.imageshack.us/img8/7606/...bitmapk.th.jpg
really interesting ...
Quote:
Originally Posted by unclewebb
Thank you. That version works. Much appreciated. :up:
KURTZ: That sensor test is meaningless. Your have to run 2 threads of Prime95 Small FFTs to test your sensors. Running one thread isn't good enough.
It looks like the slope of the temperature curves for each core are completely different. That's not unusual. They likely have two different TJMax values as well. Intel says that there is some error in how they set TJMax but wasn't too specific beyond that. With a proper CPU Cool Down Test, I could probably make some sense out of his sensors. At least they are not sticking at idle. When your friends are testing, ask them to turn off as much background crap as possible so the CPU can idle down. Don't run any other monitoring programs when doing a cool down test.
Wishmaker: Good to hear.
I have been using the RealTemp plugin for Rivatuner and it works
great now and here it shows RealTemp Q6600 temperature in
Google Earth 5 and I'm now looking for some slick tools to use
that is even better than the "CPU usage history" as the frame
around it is sometimes distractive and would be better if it could
be easier controlled.
Haven't had the Rivatuner "history" window on the desktop
before, and it is useful and shows relevant infomation but has
some buttons etc that are a bit distracting.
Google Earth with Rivatuner server ON and the cloud animation
was running for a while which made the four cores work some
but then I paused it to get a nice picture of that hurricane
http://img20.imageshack.us/img20/826/wpneoge1q95.jpg
Google Earth with Rivatuner server OFF
http://img20.imageshack.us/img20/197...e2noosdq85.jpg
nview Transparency in turned on when Google Earth is closed
which shows on some windows
http://img20.imageshack.us/img20/5916/wpneoge3trq95.jpg
It is very nice to have the possibility to have the four cores
from RealTemp overlayed in all sorts of programs, games etc.
It has worked stable, no chrashes, in Prime95, games like
Grid, Grysis and the old UT2004.
I installed Google Earth 5 two weeks ago but it did a BSOD
and restarted after a few days and once again this week,
could be a bug or something not stable in my computer but
the games should have crashed if there are stability issues.
I did up the Volt a bit on the DDR2 memory some months
ago and got the games stable but I still got an issue
sometimes when I restart (warmboot), some applications
just have a Restart Computer option, and then when the
computer warmboots it beeps three times (memory issues)
and starts with lower FSB/MEM and other settings in bios
like 266/(533=2*266) instead of 400/(800=2*400) but
the if I coldboot it is never a problem.
Hi guys, I'm new here and I've been using realtemp for over a year for my q9450 rig and it's been working great so far! I have a question concerning laptops though. I have a Dell studio 1537 laptop with a T5800 peryn processor. I fired up realtemp on it and with my TJ max set at 100c, the 2 cores idle at 80C. This can't be normal right? I've never opened up the laptop or anything like that either. Should I be concerned or is realtemp the best program to be using for core 2 duo laptops.
Most 45nm mobile processors have a TJMax = 105C but I had a look at the specs for a T5800 and it shows 85C.
http://processorfinder.intel.com/det...x?sSpec=SLB6E#
Intel turns out CPUs faster than I can keep up with them. Their documentation of TJMax has plenty of inconsistencies so I have no idea what the truth is for many of their CPUs. I kind of gave up on getting the whole truth from them. Due to lack of cooling, mobile CPUs are impossible to test and calibrate like you can do for the desktop processors. If the Intel documentation says 85C then go into the Settings window and set TJMax=85C.
That will lower your reported temperatures but it still doesn't leave you with a lot of head room before thermal throttling will kick in when Distance to TJMax approaches zero. Throttling usually starts to occur when the Distance to TJMax number counts down to about 2 or 3 for the 45nm Core 2 desktop CPUs. Turn off as much background junk as possible and enable all of your power saving options / C-States and reduce the Minimum processor state in the Power Options. Some of these on a mobile chip can reduce idle temps by 5C or so at idle and maybe more. Maybe some company needs to design a thermal blanket for ones lap while using a laptop. Computing in the summer time can be a miserable experience. :)
-X-hellfire: Interesting to see RealTemp's temps in the eye of the storm. :up:
@ Uncle: http://www.hwupgrade.it/forum/showpo...8&postcount=35 these are the 8400's tests that you want :)
KURTZ: Those tests show that his E8400 is running like any other E8400 at those speeds.
Here's a couple of screen shots of my E8400 at 3000 MHz and at 3200 MHz.
I'm using DDR2-1000 memory speed so maybe that is why my times are slightly better but the times are very similar so his CPU is not thermal throttling during normal use.
http://img223.imageshack.us/img223/4...4003000mhz.png
http://img32.imageshack.us/img32/754...4003200mhz.png
The Thermal Status information that RealTemp reports comes from model specific register (MSR) 0x19C within the processor for each core. Intel calls this the IA32_THERM_STATUS register and bit[0] of this register is called the Thermal Status bit. When it is set, the CPU is reporting that thermal throttling is in progress.
The last 4 bits in this register contains various information about thermal throttling including a log of any thermal throttling incidents in bit[1]. On his E8400, all 4 bits might be set so the last digit in this MSR for him could be the letter F. It should be the number zero (0). In my tool you have to enter 0x19C in the MSR Number box and then click on the Read MSR button to display the Thermal Status data.
http://img524.imageshack.us/img524/2582/msr0x19c.png
He could use my MSR Tool or any other similar tool that lets him read the value of this MSR.
http://www.fileden.com/files/2008/3/3/1794507/MSR.zip
http://release.crystaldew.info/OlsMsrEditor
You can also use CPU-Z. Just click on the About tab and then click on Save Report and it will output the value of this MSR into its report for each core. That would be handy to send to Intel if there is a problem because they tend to trust CPU-Z.
My opinion is that this CPU temperature sensor is broken and Intel should be willing to exchange his CPU for a CPU that is not broken. Ask him to post a screen shot of MSR 0x19C and I will interpret it for him. This MSR also contains the temperature data in the higher bits but it looks like that is stuck / broken too. :(
Yeah, nice to see RealTemp were all the actions is :cool:Quote:
Originally Posted by unclewebb
...in Google Earth which is an awesome program with all kinds of tools/addons that can present all kinds of data and it can be stressful for many parts of the computer, cpu (multicores) , gpu , memory (it handles a allot of memory) , harddisk, ISP (broadband need to be faster and faster, my ADSL 24/3 mbit/s is crippled by my old D-Link DI-804 firewall that 5/5 mbit/s
Did some test with nVidia PhysX fluid-demo today and it is great to see the temperature OSD in realtime and it is interesting to see what happens when you run in hardware vs software mode as in GPU PhysX vs CPU. The performance on my setup is something like PhysX 22 fps and CPU 5 fps. The four cpu cores are however not used very much compared to Prime95 but is is probably very har to optimize the code to squeeze everything out of the cpu cores.... but they looked like they were stalling alot, waiting while some other threads finished its tasks, anyway the RealTemp plugin in Rivatuner showed this beahavior
check here :)Quote:
Originally Posted by unclewebb
http://www.hwupgrade.it/forum/showpo...0&postcount=44
Hi Kevin,
any work done on this little puppy? The numbers look pretty small to me. Is the temperature monitoring mechanism on LGA1156 the same with the one used on LGA1366, or ar there any subtle changes (TjMax, register adress, etc)???
http://lab501.ro/wp-content/uploads/...lTemp-i860.jpg