• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

RealTemp General Discussion

slipstream

New Member
Joined
Jul 14, 2011
Messages
19 (0.00/day)
I wasn't gaming at the time, i think my GPU averages ~48(idle) with optimus turned on then why the log show the temps at 15, 49...? though the 3.60 logs their correct value even when GPU is idle

I'm not getting any error message, the 3.69 launches in the background, is not visible in the taskbar
but shows up in the task manager and still logs the temps in the log file

Played Bioshock for around 40 mins, 20:39 to 21:19 , the 3.60 reports correct values of the GPU along with CPU temps and Load%:
http://www.mediafire.com/?wxe14tul5111byb

But 3.69 had some aces up its sleeve:
http://www.mediafire.com/?ticonyr1p5xvv69

PS: I use throttle-stop to disable turbo boost when gaming cause the cooling on my notebook sucks
at the time i had set CPU multiplier to x18
 

unclewebb

ThrottleStop & RealTemp Author
Joined
Jun 1, 2008
Messages
7,273 (1.26/day)
When I had a look at your log file on the previous page, when the GPU wasn't disabled, it reported temperatures between 47C and 53C. When it was disabled it didn't report anything so the Load % is shifted over and showing up in the GPU column. That's why I'm going to move the GPU temperatures to the far right side of the log file so it will be easier to see when there is GPU data to report and when there isn't any data being reported because the GPU is asleep.

In the second log file it looks like you have two different instances of RealTemp running and you have them outputting data to the same log file. GPU monitoring does not seem to be turned on. Kill all the RealTemp versions in the Task Manager, delete the INI file in case there are any problems and try running a single instance of RealTemp 3.69 again and see what happens.

In Windows 7 try turning on show all icons in the system tray. Depending on the colors chosen, the RealTemp temperature numbers can get hidden because Windows generally uses a dark background for the System Tray but then uses a white background color when icons are hidden. This can make the icons impossible to see or find. If you show all icons then at least you should be able to find the RealTemp temperature icons.

If you are running ThrottleStop then why not use it for logging purposes? No need to run RealTemp too.
 

slipstream

New Member
Joined
Jul 14, 2011
Messages
19 (0.00/day)
What i meant was that 3.60 version logs GPU temps even on idle, you say that when it is disabled there was nothing to report (in the 3.69 log file) hence the shifting of LOAD% but shouldn't it report temps when GPU is active like when i played bioshock still it reports the temps inconsistently.

I have customized the taskbar to show the GPU temp only and hide the rest still i was not seeing any icons hidden or visible, deleting the INI file did the trick now i am able to run 3.69 version

How can different versions output to same file? aren't the .exe contained in different folders
Also i prefer realtemp as throttle stop logs data at an interval of 1 second, realtemp logs can be customized and i can easily calculate CPU, GPU temp averages using average function in excel.
 
Last edited:

unclewebb

ThrottleStop & RealTemp Author
Joined
Jun 1, 2008
Messages
7,273 (1.26/day)
What probably happened in your case is that you started up RealTemp 3.69 and then you couldn't see that it was already running so you started a second instance of RealTemp 3.69. The same version was running twice and both instances were writing to the exact same log file in the same directory. The time data in that log file that is going backwards and forward is usually a good indication of this problem.

I haven't done any Optimus testing. RealTemp gets its GPU information from the Nvidia driver. When the Nvidia GPU is not active, I don't know if querying the temperature from it will cause it to wake up or not. By your results, it looks like the driver doesn't wake up the GPU so it ignores temperature requests. It's also possible that sampling interval that you set in RealTemp can affect this. A shorter sampling interval like 1 second might keep the GPU awake while a longer sampling interval like 20 seconds might result in no regular data being reported. I'm just guessing at what might be going on here. You'll have to do some testing to find out more. Try running a single instance of RealTemp 3.60 in one folder and run a single instance of RealTemp 3.69 in a different folder. Use the same logging interval for both of them and see what happens in the log file.

If you right click on an empty area of RealTemp, there is an option called Set Log File Location that lets you select where you want to write your log file data to.

I'll look into adding an adjustable time interval to the ThrottleStop logging option in the future.

When I set Windows 7 to show all icons, I never have a problem. RealTemp seems to have problems with icons disappearing or being hard to find as soon as you use the Windows 7 hide some icons feature. The only way to maybe fix this is to write the temperature number onto a solid background color. I kind of like the temperature numbers on a transparent background the way they are now so I don't plan to change this.

Edit: I might have to change RealTemp so it checks for an Nvidia GPU more often but doing that could also defeat Optimus saving energy so that might not be a good idea.
 
Last edited:

slipstream

New Member
Joined
Jul 14, 2011
Messages
19 (0.00/day)
Yes i did had multiple instances of 3.69 running, guess that explains the GPU temp log screw up during normal load as well as when i was gaming.

As of now 3.69 is correctly logging both CPU and GPU temps, CPU ~44, and GPU ~47(on normal load), going by these numbers i think Realtemp does wakes up the driver which in turn returns the correct temp at that moment
http://www.mediafire.com/?z6f3pq6if5wdakh
One last question
Is the calculation of averages by me using the logged temps in the excel file valid? i've observed that sometimes CPU, GPU temps peak at ~86 but their temp averages at ~78, i ask this as i've had different readings of the GPU average using the average function in excel log and using GPU-Z, GPU-Z always reports average GPU temp at least 4-5 ° C less as compared when i calculate average in excel.

Almost forgot to add that one could gain respectable amount of knowledge reading this thread and the Realtemp documentation.
 
Last edited:

unclewebb

ThrottleStop & RealTemp Author
Joined
Jun 1, 2008
Messages
7,273 (1.26/day)
I think GPU-Z reads the GPU temperature information directly from a monitoring chip while RealTemp is reading it from the driver so that could cause some slight differences.

CPU and GPU temperatures can change instantaneously. If you are collecting data using RealTemp with a sampling interval of 20 seconds then that is also going to lead to differences in the average compared to sampling at a different interval.

You could try running a GPU-Z and RealTemp log at the exact same time and sampling at the exact same interval. In theory, the data should be very similar and comparable so working out an average over the same interval should not be showing a 4C or 5C difference. I'm sure the GPU-Z temperature averaging works fine and would produce comparable results as long as the interval was the same. I didn't write the driver or GPU-Z but if GPU-Z is sampling the sensor directly then I'd be inclined to trust that data.

Another test you could do is use Battery Bar or ThrottleStop with GPU monitoring turned off and have a look at Battery Power Consumption with RealTemp logging the GPU compared to RealTemp running but not logging the GPU. If power consumption goes up significantly then that would likely be a sign that RealTemp is constantly waking up the GPU which wouldn't be a good idea on battery power. ThrottleStop will log battery power consumption in mW. It gets this data directly from Windows but Windows only updates it about every 10 seconds. It takes a while to get some repeatable data when testing so you need some patience.
 

slipstream

New Member
Joined
Jul 14, 2011
Messages
19 (0.00/day)
I'll try and run those tests, thank you for your time, you've been a real help.
 

donkrx

New Member
Joined
Jul 29, 2011
Messages
8 (0.00/day)
6 cores! what do people need 6 cores for? to calculate the answer to life universe and everything :) don't know much but i assumed Intel would not release a 6 core CPU until next year, guess Moore's law still holds.

My brother said he needs 2 hex cores (not sure if he meant 1 with hyperthreading, or actually 2 with hyperthreading).... he's a 'computer scientist'. Not sure exactly what he does, but I think the use is mostly for compiling code, and for that, you always need faster. It's not a greed thing.

Also, FEA analysis. Typically things get really complicated with multiphysics analysis, and for that, you again need the CPU ultra fast. Solutions can take days, months, a year if you wanted. We always need more accurate answers and less simplifications being made, so there's always a real world demand.

For gaming... that's different. Please don't buy hex cores for gaming, if anyone's reading this lol.

My testing found that these sensors work like the graph in the documentation shows. Intel came along and released some information at one of their IDF forums that showed the error as a totally linear relationship from idle to TJMax. As a small individual programmer, I decided to go along with their version of the truth but in hindsight, I should have ignored this "new information" and never changed my original formula. Most of the information that was released at the first IDF conference turned out to be a farce and Intel released an updated version of the truth a month later. I was naive thinking that Intel was going to finally come clean and release some engineering data about these sensors at IDF to help out developers but it never happened.

After that, I gave up on doing calibrations and I haven't looked at this INI option for a long time. I'll check if it still exists or if I got rid of this option from lack of feedback.

I found Prime95 Small FFTs was the best application for testing purposes. LinX / Linpack testing was able to create more heat but Small FFTs puts a very equal load on each core.

Interesting, thanks. My gut says that the original findings are still true despite what intel said there. If you ever test any more chips put some info on the site, that would be good to know if anything's really changed over time.

And yeah Prime small FFT is definitely my favorite. High temps, equal loading, but not crazy insane temps that are unrealistic for most people.
 

slipstream

New Member
Joined
Jul 14, 2011
Messages
19 (0.00/day)
@donkrx
my point was that normal users or even hardcore gamers don't need six cores

people who employ six cores have state of the art cooling and aren't gonna rely on realtemp alone to log/regulate the temps(not a derogatory remark, in fact it's the best i've used so far)

the university from where i'm pursuing my masters(in computer application) has a state of art simulation lab for weather forecasting, protein analysis etc, among other things their server room is insane most sensors there are taken right out of a sci-fi movie.
 

Mussels

Freshwater Moderator
Staff member
Joined
Oct 6, 2004
Messages
58,413 (8.21/day)
Location
Oystralia
System Name Rainbow Sparkles (Power efficient, <350W gaming load)
Processor Ryzen R7 5800x3D (Undervolted, 4.45GHz all core)
Motherboard Asus x570-F (BIOS Modded)
Cooling Alphacool Apex UV - Alphacool Eisblock XPX Aurora + EK Quantum ARGB 3090 w/ active backplate
Memory 2x32GB DDR4 3600 Corsair Vengeance RGB @3866 C18-22-22-22-42 TRFC704 (1.4V Hynix MJR - SoC 1.15V)
Video Card(s) Galax RTX 3090 SG 24GB: Underclocked to 1700Mhz 0.750v (375W down to 250W))
Storage 2TB WD SN850 NVME + 1TB Sasmsung 970 Pro NVME + 1TB Intel 6000P NVME USB 3.2
Display(s) Phillips 32 32M1N5800A (4k144), LG 32" (4K60) | Gigabyte G32QC (2k165) | Phillips 328m6fjrmb (2K144)
Case Fractal Design R6
Audio Device(s) Logitech G560 | Corsair Void pro RGB |Blue Yeti mic
Power Supply Fractal Ion+ 2 860W (Platinum) (This thing is God-tier. Silent and TINY)
Mouse Logitech G Pro wireless + Steelseries Prisma XL
Keyboard Razer Huntsman TE ( Sexy white keycaps)
VR HMD Oculus Rift S + Quest 2
Software Windows 11 pro x64 (Yes, it's genuinely a good OS) OpenRGB - ditch the branded bloatware!
Benchmark Scores Nyooom.
@donkrx
my point was that normal users or even hardcore gamers don't need six cores

people who employ six cores have state of the art cooling and aren't gonna rely on realtemp alone to log/regulate the temps(not a derogatory remark, in fact it's the best i've used so far)

the university from where i'm pursuing my masters(in computer application) has a state of art simulation lab for weather forecasting, protein analysis etc, among other things their server room is insane most sensors there are taken right out of a sci-fi movie.

i'm not a hardcore gamer, and need 6 cores.

you have it completely backwards: its non gamers who need the more cores.

games are the slowest things to update for more CPU threads, whereas non gaming apps like encoding video and audio can use all that we have today, and still need more.

you do realise that both intel and AMD have 6 core CPU's out that work on their stock coolers? slabs of metal with a fan? they dont need any state of the art or fancy cooling.
 

unclewebb

ThrottleStop & RealTemp Author
Joined
Jun 1, 2008
Messages
7,273 (1.26/day)
RealTemp 3.69.1
http://www.mediafire.com/?4uixpjtezznuzkd

-fixed number of log file headings when using Core i Dual Core CPUs.
-moved GPU column to far right end of the log file.
-improved formatting and added more white space to the ThrottleStopLog.txt file.




slipstream: Can you give this version a try and let me know how it works on your CPU?
 

slipstream

New Member
Joined
Jul 14, 2011
Messages
19 (0.00/day)
@ Unclewebb
new version works perfectly, thank you again.

@Mussels
From my experience i've seen that you can run any current game on a core i5 and a high end graphics card i'd be more than happy to retract my statement if you could provide me with some literature on this
also if you ever come to my part of the world i'd be more than happy again to show you around the simulation lab in my college.
 
Joined
Aug 23, 2011
Messages
217 (0.05/day)
System Name HP Pavilion - 15-cw1068wm
Processor AMD Ryzen 5 3500 U
Motherboard HP
Cooling standard laptop cooling
Memory 16 gb DDR4
Video Card(s) Radeon™ Vega 8 Graphics
Storage 125 gb internal M.2 SSD and 1 TB internal laptop drive.
Display(s) Built in
Case Laptop
Audio Device(s) built in B&O audio.
Power Supply Laptop 45w
Mouse track pad , and SteelSeries mouse.
Keyboard built in.
Software Windows 10 pro
i went to this program today from core temp as core temp had a new stupid installer. i installed this it works good but what it mean to calibrate all temps look just like they did with core temp. i'm not a technical inclined do i need to calibrate it?
 

unclewebb

ThrottleStop & RealTemp Author
Joined
Jun 1, 2008
Messages
7,273 (1.26/day)
Don't waste your time calibrating. Intel's temperature sensors are designed for thermal throttling and thermal shut down control only. They are not 100% accurate temperature monitoring devices so what software reports is not 100% accurate either. Use RealTemp or Core Temp as a general guide line but don't compare your temperatures to your friend's temperatures because neither is 100% accurate. It's just a number.
 
Joined
Aug 23, 2011
Messages
217 (0.05/day)
System Name HP Pavilion - 15-cw1068wm
Processor AMD Ryzen 5 3500 U
Motherboard HP
Cooling standard laptop cooling
Memory 16 gb DDR4
Video Card(s) Radeon™ Vega 8 Graphics
Storage 125 gb internal M.2 SSD and 1 TB internal laptop drive.
Display(s) Built in
Case Laptop
Audio Device(s) built in B&O audio.
Power Supply Laptop 45w
Mouse track pad , and SteelSeries mouse.
Keyboard built in.
Software Windows 10 pro
ok thanks i like this program better then core temp thanks for making it :). makes it much simpler that i don't have to install anything. i have one more question i set it in task scheduler to open up on windows 7 upon boot. but it ask always for user permission even though i went into property's and set it to always run as admin. any whey around that?.
 
Joined
Aug 23, 2011
Messages
217 (0.05/day)
System Name HP Pavilion - 15-cw1068wm
Processor AMD Ryzen 5 3500 U
Motherboard HP
Cooling standard laptop cooling
Memory 16 gb DDR4
Video Card(s) Radeon™ Vega 8 Graphics
Storage 125 gb internal M.2 SSD and 1 TB internal laptop drive.
Display(s) Built in
Case Laptop
Audio Device(s) built in B&O audio.
Power Supply Laptop 45w
Mouse track pad , and SteelSeries mouse.
Keyboard built in.
Software Windows 10 pro
oh thank you it worked :)
 

pipsie

New Member
Joined
Aug 26, 2011
Messages
3 (0.00/day)
I used realtemp today because i didn’t trust coretemp anymore. (and it was indeed wrong).
And i did the stress test. And i doesn’t look good :(

http://www.uploadarchief.net/files/download/pctemp.png
I'm wondering how close you may get to the Tjmax? (i think the 100°C that intel gives as Tjmax is still in the save zone because most electronics can go to 110-120°C?)
But nevertheless 95°C is probably still too hot.

Ow and it’s a laptop no desktop :)
 

unclewebb

ThrottleStop & RealTemp Author
Joined
Jun 1, 2008
Messages
7,273 (1.26/day)
RealTemp 3.69.1
http://www.mediafire.com/?4uixpjtezznuzkd

You should be using the 4 core version of RealTemp which is called RealTemp.exe
The GT version is for the 6 core CPUs. The above download has more features for the newer CPUs like you have. Core Temp also released a new version recently which should correctly support your CPU but I haven't fully tested it yet.

The RealTemp sensor test was designed more for the desktop CPUs. They tend to have a lot better cooling than laptops do. Your CPU did not reach the thermal throttling point during that test so that's a good thing. When it reaches TJMax, Intel CPUs are designed to slow down and reduce the core voltage to help reduce heat. Some laptop manufacturers set a very conservative shutdown temperature at 100C. The Intel thermal shutdown temperature isn't until 130C but you will never see that temperature if your laptop manufacturer decides to shut your laptop off if it reaches 100C. The well designed Intel thermal throttling method gave users some time to still use their laptop when it was getting hot and you also had time to save anything important if there was a problem like your heatsink was loose. Most laptop manufacturers have killed that feature.

Enjoy your laptop. Sandy Bridge laptops run hot because many manufacturers are using inadequate cooling solutions. I have seen a lot worse than yours. One user running WIndows 7 (BootCamp) on his Apple laptop was only able to run his CPU at 50% of its rated load because of overheating issues. Other laptops like the Acer 3830TG have severe throttling problems when trying to use the CPU and GPU at the same time. Every major manufacturer has one throttling problem or another and severe throttling when running on battery power is common. Some laptops are running at less than one third of their rated speed when on battery power which is false advertising in my opinion. For laptops the other program I wrote is called ThrottleStop. That program is very useful at monitoring for and correcting a variety of throttling issues.

ThrottleStop 3.30
http://www.techinferno.com/downloads/
 
Joined
Aug 23, 2011
Messages
217 (0.05/day)
System Name HP Pavilion - 15-cw1068wm
Processor AMD Ryzen 5 3500 U
Motherboard HP
Cooling standard laptop cooling
Memory 16 gb DDR4
Video Card(s) Radeon™ Vega 8 Graphics
Storage 125 gb internal M.2 SSD and 1 TB internal laptop drive.
Display(s) Built in
Case Laptop
Audio Device(s) built in B&O audio.
Power Supply Laptop 45w
Mouse track pad , and SteelSeries mouse.
Keyboard built in.
Software Windows 10 pro
hi just d-loaded prime and did the sensor test my cpu has always run hot i thought but i figured i would post this here to see if it is ok. also how would you remove prime if you don't want it just delete it?



Uploaded with ImageShack.us
 

pipsie

New Member
Joined
Aug 26, 2011
Messages
3 (0.00/day)
Your CPU did not reach the thermal throttling point during that test so that's a good thing.
http://www.techinferno.com/downloads/
Yea i thought that too but you can’t imagine how hot the air is coming out of the side.
It’s not normal in fact I’m going to measure it once.
When it’s just running the OS and google chrome (CPU 6%) the air coming out 49.5°C
I'm searching for a temperature sensor that goes to 100°C so i can measure the heat at 70% load

Some laptop manufacturers set a very conservative shutdown temperature at 100C.
http://www.techinferno.com/downloads/
Yea already so a few blue screen alike with text temperature on it

Sandy Bridge laptops run hot because many manufacturers are using inadequate cooling solutions.
http://www.techinferno.com/downloads/
To be clear i have the earliest i7 processor so no sandy bridge if i'm correct

Other laptops like the Acer 3830TG have severe throttling problems when trying to use the CPU and GPU at the same time.
http://www.techinferno.com/downloads/
Yea that’s also one of my fears the GPU is around 60°C (normal because it's a chipset so the heat sink is the same of the CPU/CPU)

For laptops the other program I wrote is called ThrottleStop. That program is very useful at monitoring for and correcting a variety of throttling issues.
http://www.techinferno.com/downloads/
If i read the info on the site correct this programming is for overclocking?
And measuring the temperature and boost of the CPU?
I think i wouldn’t be wise to do this because the PC is already reaching its Tjmax at 100% load.

[edit]
forgot the test result with the correct software :)
 
Last edited:

unclewebb

ThrottleStop & RealTemp Author
Joined
Jun 1, 2008
Messages
7,273 (1.26/day)
TonyB: Nothing wrong with peak core temperatures in the low 60C range. That's normal. Most users are overclocking their CPUs and running them much hotter than this. If Prime95 came with an installer then go into the Control Panel - Add Remove Programs or Programs & Features and uninstall it. If Prime95 didn't come with an installer then delete the folder it is in and it will be gone from your system.

pipsie: I didn't enlarge your original picture this morning so I thought you had a 2720QM. The new Sandy Bridge CPUs seem to report higher core temperatures during normal use. Manufacturers are letting them run hot. The 720QM is built on the older 45nm technology and seems to put out more overall heat. Makes for a nice lap warmer during the winter.

ThrottleStop can be used to significantly overclock the Extreme CPUs like the 920XM and 940XM. For your CPU, it's more of a tool to test for throttling and to correct for this problem if you have it. When the Dell XPS-1645 first came out with an inadequate 90W adapter, it had lots of throttling problems where the CPU would be left running at less than one quarter of its rated speed. Dell did an adapter recall program which helped but it didn't 100% correct the problem. When combined with an ATI 4670 GPU, it was just way too much heat to dissipate in a laptop.

The latest trick that Dell is using on their top of the line Alienware M18x gaming laptop is to use clock modulation throttling on 7 of the 8 threads.



Clock modulation throttling slows down a CPU internally. ThrottleStop lets you see how all 8 threads are performing so it gives you the most accurate look at how your Intel CPU is really performing.
 
Joined
Aug 23, 2011
Messages
217 (0.05/day)
System Name HP Pavilion - 15-cw1068wm
Processor AMD Ryzen 5 3500 U
Motherboard HP
Cooling standard laptop cooling
Memory 16 gb DDR4
Video Card(s) Radeon™ Vega 8 Graphics
Storage 125 gb internal M.2 SSD and 1 TB internal laptop drive.
Display(s) Built in
Case Laptop
Audio Device(s) built in B&O audio.
Power Supply Laptop 45w
Mouse track pad , and SteelSeries mouse.
Keyboard built in.
Software Windows 10 pro
awesome thanks :)
 

n3tmaster

New Member
Joined
Aug 26, 2011
Messages
5 (0.00/day)
Hi uncleweb, congratulation for RealTemp..I think it is great but I haven't understood few its settings.

How does "Power Limits" work? (on my Sand Bridge with Z68 and 2500k I check Lock Power Limit and Lock Current Limit in order to prevent changes on CPU multiplier..isn't it?)

Moreover I haven't understood the Advenced Calibration Settings...

Thank u in advance for answer and I apologize for my poor english

Best regards
 
Top