Real Temp - New temp program for Intel Core processors

[unclewebb]

Uncle,I saw Real Temp 3.50 and i5-660,didn't match CPU speed and multi..

Most of the screen shots are taken at idle. The actual multiplier at idle can be influenced by how the Power Options -> Minimum processor state is set as well as C1E and SpeedStep. I don't know how his computer is set up so it's hard to comment.

In the second post with 4 pictures, CPU-Z and RealTemp are in agreement on 3 out of the 4 pictures. For the first picture, RealTemp is reading a lower multiplier at idle which might be a sign that he has C1E enabled.

In the first post with RealTemp reporting a 25 multiplier, I'm not too sure. RealTemp uses internal timers which can accurately determine the multiplier but unfortunately they are not a protected resource so any monitoring program can change or reset them. I haven't kept track of what program is using or abusing these timers so I'm not sure if maybe one of the other monitoring programs he was using might be using these timers inappropriately.

When RealTemp is running by itself, I haven't seen it get the multiplier wrong on too many Core i7/i5/i3 CPUs.

As for Core 2 CPUs like M4rk, that's another story.

I was helping JohnZS with his QX CPU a couple of months ago and accidentally released a "test" version designed for him into the main beta download area. That one off version got hosted everywhere so I said F it. I got side tracked with other projects at that time so left it and since there have been very few complaints, I haven't been motivated enough to fix it. Between that bug and the ATI code blowing up randomly on some computers while working fine on others, I just kind of lost interest in project RealTemp.

M4rk: Try downloading the last official version, RealTemp 3.40, and see if that gets your multiplier right. If so I'll send you an updated version tomorrow.

[andressergio]

Uncle,I saw Real Temp 3.50 and i5-660,didn't match CPU speed and multi.
1.http://www.xtremesystems.org/forums/...&postcount=101
2.http://www.xtremesystems.org/forums/...2&postcount=21
Any particular reason?Thanks.

Not only on I5 Stasio on my Xeon X3380 the same...also Realtemp, Coretemp shows 5C more than Everest, Speedfan, HWinfo and so others

you can see the temp issue here

[unclewebb]

Here's the formula that all software uses:

Reported Temperature = TJMax - Digital Thermal Sensor reading

All software reads the same data from the same on chip digital thermal sensors. The disagreement is in TJMax. RealTemp and Core Temp assume that value is 100C and other software assumes that value is 95C so RT and CT will always report a temperature 5C higher than the other programs.

Which value is correct? The truth is that there is no documented correct value. All software is guessing.

The only documentation ever published by Intel calls this value TJ target and they've said that this is not a fixed value. It can and does vary from one CPU to the next with the same model number and in my testing it also varies from core to core on the same CPU. Your core 1 reading much lower than the other 3 is a good indication of that.

These sensors are designed for thermal throttling and thermal shut down control and are accurate enough for those purposes. For any other purpose like accurate temperature reporting, the 45nm Core 2 Duo/Quad sensors are crap with a long list of known issues. 100% accurate temperatures from these undocumented sensors from idle to TJMax is simply not possible.

If you want RealTemp and Core Temp to low ball your temperature readings like the other programs are doing then set TJMax to 95 instead of 100.

[andressergio]

Here's the formula that all software uses:

Reported Temperature = TJMax - Digital Thermal Sensor reading

All software reads the same data from the same on chip digital thermal sensors. The disagreement is in TJMax. RealTemp and Core Temp assume that value is 100C and other software assumes that value is 95C so RT and CT will always report a temperature 5C higher than the other programs.

Which value is correct? The truth is that there is no documented correct value. All software is guessing.

The only documentation ever published by Intel calls this value TJ target and they've said that this is not a fixed value. It can and does vary from one CPU to the next with the same model number and in my testing it also varies from core to core on the same CPU. Your core 1 reading much lower than the other 3 is a good indication of that.

These sensors are designed for thermal throttling and thermal shut down control and are accurate enough for those purposes. For any other purpose like accurate temperature reporting, the 45nm Core 2 Duo/Quad sensors are crap with a long list of known issues. 100% accurate temperatures from these undocumented sensors from idle to TJMax is simply not possible.

If you want RealTemp and Core Temp to low ball your temperature readings like the other programs are doing then set TJMax to 95 instead of 100.

Thanks Uncle

[M4rk]

M4rk: Try downloading the last official version, RealTemp 3.40, and see if that gets your multiplier right. If so I'll send you an updated version tomorrow.

Yep, that works. Thanks

[WARDOZER9]

Eh, I pref the cleanliness of CoreTemp as I get the exact same temp readings from both utils. The extra info in RealTemp is nice but I just don't care for the extra info for my under-volted E7600.

[Siobhan]

unclewebb, just for your info, Real Temp 3.50 won't run if I don't put the NoATIGPU=1 into the RealTemp.ini like you suggested. I have a Sapphire Radeon 4850X2 on Win7 64-bit with the 10.1 drivers, and I am running in CrossFire. Let me know if you need more info.

[PolRoger]

Uncle Webb,

FYI... I came across this bug in RealTemp 3.50. This is on my wife's older 975X chipset mobo running Windows XP Pro. It does not report the fsb/multi like the other two versions.

[radaja]

quick question, is 105c the correct tjmax for the 32nm i3-530?

[burebista]

It's written in MSR (no more guess) so it should be correct.

[radaja]

ok thanks.

[jasjeet]

Hi Uncle, ive been using Realtemp for over a year now. Ive upgraded to the latest beta and am getting this issue:
Realtemp crash on open
Event ID 1000
Faulting application name: RealTemp.exe, version: 3.5.0.0, time stamp: 0x4b25596b
Faulting module name: ntdll.dll, version: 6.1.7600.16385, time stamp: 0x4a5bdb3b
Exception code: 0xc0000005
Fault offset: 0x0002de64
Faulting process id: 0x12c0
Faulting application start time: 0x01caab602ffe1283
Faulting application path: D:\Jasjeet\Downloads\OC\RealTemp_350\RealTemp.exe
Faulting module path: C:\Windows\SysWOW64\ntdll.dll
Report Id: 6dc56147-1753-11df-8efd-00221517d003

Windows 7 x64, Q6600 3.2ghz, 4gb Ram, ATi 4850 (maybe issue getting the GPU temp)
Prime95 for 2hrs passed today to confirm CPU stable
Memtest 200% pass today

Older versions run fine. Though i cant rule out an OS file system problem, but its only Realtemp causing an issue. If i log off, then log on, it works OK. But if i reboot/shutdown then on boot up it will not launch even with Run as Admin. I have to log off, log on for it to work. Some kind of Ram writing/permission issue i think.

If this is a problem i hope you can fix it =)

Thanks

[randomizer]

I have been getting crashes lately too with 3.50 RC6. It worked fine for months and all of a sudden I'm getting this:



It works fine if I run any compatibility mode. Oddly enough it started crashing about the same time as my .NET runtime started going crazy and crashing any .NET program that tries to use Windows Forms (even in compatibility mode). But RT doesn't use .NET...

[burebista]

Count me in too.



I only change my GTX260 with an HD5850 and a couple of days ago Seven x64 made some updates.

[jasjeet]

Compatibility mode to Vista worked, thanks for that solution, atleast for now =)

[jcmarfilph]

Any support for Caspian Processor and Mobility Radeon 4200?

Thanks

[unclewebb]

RealTemp 3.56
http://www.fileden.com/files/2008/3/...alTempBeta.zip

Finally got around to fixing a few bugs like the half multiplier bug in 3.50RC6.

The ATI GPU monitoring code is now optional since it seemed to be causing too many problems for too many users. It's always worked fine for me.

Here's how my Sapphire Vapor-X 5770 card looks.



You will need to add GPU=2 to the RealTemp.ini file to see if this code works for your ATI GPU. The default is GPU=1 for Nvidia cards. GPU=0 can be set if you don't have either type of card or if this code causes any problems.

GPU=3 can be used if you have one of each type of card such as if you are using an ATI card plus an Nvidia card for Physx. Click on the GPU button twice to see if it can read both of your cards.

I don't yet have access to a second ATI GPU so CrossFire support is still in the future.

I just updated Windows 7 x64 with all of the latest and greatest recommended patches from Microsoft but I haven't been able to duplicate any crashes that users have reported. I'm hoping that separating the ATI code might help isolate some of these issues. Thanks randomizer for finding a quick work around to this issue.

I haven't tried installing this optional update yet. It looked dangerous.



Sorry jcmarfilph, there's no planned AMD support.

[randomizer]

It appears as though that's the first one Windows installed for me (in October). So it can't be related, since the issue only started recently. I'll have to look at other updates to see if they may be involved.

[randomizer]

If you're having crashes, I recommend updating. It fixed the problem I was having, even though I didn't have it initially.

[Ol'Baditude]

I don't yet have access to a second ATI GPU so CrossFire support is still in the future.

I have a 4670 or two 4890's for short term loan if that would help you out. Lemme know and I'll take a drive out there.

[unclewebb]

Ol'Baditude: Thanks for the offer. I'll send you a PM to discuss further.

[Wishmaker]

Added the line and these are my results:

GPU1: 48
GPU2: 48 -don't know why RT sees that as GPU3.

Which is not correct. One of my GPU's is hotter and Afterburner shows the correct temp as 51. Bottom right corner.

[burebista]

If you click on that GPU3 what do you see?

[donmarkoni]

Added the line and these are my results:

GPU1: 48
GPU2: 48 -don't know why RT sees that as GPU3.

Which is not correct. One of my GPU's is hotter and Afterburner shows the correct temp as 51. Bottom right corner.

Those two temps of 48 degrees are minimum and maximum, not gpu1 and gpu2. Click bottom right button which says gpu3 (same as RT) and you'll see other gpu's temp.
EDIT: @burebista: you were quicker!

[Wishmaker]

The temps are identical. This is not my first BBQ lads and Afterburner does not agree with the temps .


***removed attachments because it was making a nasty layout .