Real Temp - New temp program for Intel Core processors

[unclewebb]

Time to plug in my Win 7 drive for some testing.

[DJSUB]

uncle what's the tjmax for i7 920 D0?

thanks!

[unclewebb]

Using Windows 7 (x64) Build 7000, I downloaded RealTemp 3.20 RC3, unzipped it into a new folder and started it up and it popped up in the middle of the screen like it should.

DJSUB: Could you try downloading RC3 again? Either that or go into the RealTemp.ini file and delete the entry for WindowXY. If this value gets screwed up it's possible that RealTemp might disappear somewhere when you open it so it will appear in the Task Bar area and in the System Tray but you won't be able to see it on your screen. This is somewhere between an obscure feature and a RealTemp bug. Let me know if you get this figured out. I don't have Build 7068 to do any testing with.

As for TJMax, that information is stored within the CPU in Core i7 processors and RealTemp or Core Temp should be able to read that value correctly. In the RealTemp Settings Window you should be able to click on Defaults and it will set TJMax to the appropriate default value. All of the C0 CPUs that I've seen are set to TJMax = 100C by Intel but I've seen a couple of D0 CPUs that were TJMax = 92C or 93C. I think those were both ES engineering sample processors though. So far, the retail CPU screen shots I've seen for the new D0, have all been TJMax = 100C. I added some code to RealTemp a while ago so if you switch CPUs, it should automatically check to see if TJMax has changed.

When comparing temperature programs I like to use Prime 95 Small FFTs. It puts a consistent load on the cores so all programs should report pretty much the same thing. If the APIC ID is not sequential in the Settings window then it's possible that RealTemp will report core temperatures in different positions than the competition does but that's a feature, not a bug.

[Draxx]

What version number is displayed in the About... box of RealTemp?

Version 3.20 RC3 is available here if you don't have it already:
http://www.fileden.com/files/2008/3/...alTempBeta.zip

If RC3 has this issue then I'll try sending you some test versions to see if I can find out exactly what the issue is.
I thought this bug was fixed but with XS down for a couple of weeks, maybe I just didn't hear about it.

Yes it's RC3 Uncle.

[Pyr0]

Hi, I have a similar problem to i43
I'm running Windows 7 7077 x64. RealTemp (3.20 RC3) opens correctly and displays temps. But when exiting, i get a popup "RealTemp has stopped working"

[i43]

Hey Unclewebb !
keep your new RT builds coming currently have such equipment, 3 separate machines i have access and ready to test new versions
-1- Quad qx9650 - ( multi boot ) 1) Xp-Pro 32bit (sp3) / 2)Vista Home Premium sp1 32bit / 3)W7-b7000 64bit
-2- Dual E8400 - 1) Xp-Pro 32bit(sp3)
-3- (Dual) T5500 - 1) Vista Home Premium Sp1 32bit /laptop/(fujitsu simens Amilo)


+edit adding /
booted quad from W7 64bit (b7000)
RT 3.20 RC3 - all seems working ok but on close[x] getting same error as PyrO

//direct link: http://img26.imageshack.us/my.php?im...320rc3test.jpg //

[unclewebb]

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

I think I found the bug that was causing a crash when closing RealTemp.

If you had problems before then try RC4 and see if your problems go away.

[i43]

RealTemp 3.20 RC4 - seems like all OK ! working
anyway ..so far did not notice any problems on these ...

-1- Quad qx9650 - ( multi boot )
---A) Xp-Pro 32bit (sp3) / (ATI 4870x2 ..ccc9.4
---B)Vista Home Premium sp1 32bit / (ATI 4870x2 ccc9.4)
---C)W7-b7000 64bit / (ATI 4870x2 - default driver
-2- Dual E8400 - 1) Xp-Pro 32bit(sp3) nVidia 9600GT Fw 182.50 xp
-3- (Dual) T5500 - 1) Vista Home Premium Sp1 32bit /laptop/(fujitsu simens Amilo) - (Mobile intel(R) 945 Chipset Family)

-----
CY

[unclewebb]

Thanks for your help testing this i43. I went through RealTemp trying to make it more efficient by freeing up any unused resources. I hate memory leaks.

When using an ATI video card, RealTemp RC3 was freeing up the same block of memory twice, once during initialization and again after clicking on the exit button. In programming, freeing up the same block of memory twice is definitely not a good thing. RC4 should be all better now.

I added a new System Tray menu item so it moves RealTemp up to the upper right corner and out of the way. That's where I like it so it doesn't get in the way when it's not minimized.

[Pyr0]

exiting works without error now. thank you unclewebb

tested on the sys in sig, using win7 7100 x64

[edit] liking the new option to put the window in the top corner too

[D3mon_Hunt3r]

Just to congratulate you on some kickass software.
And tremendous work keeping it bug free.

[DJSUB]

thanks for the 3.20 RC4 uncle!
it works fine now with 7068 windows 7

keep up the good work uncle!

[gymenii]

No errors at all with win7100 x64.

ty sir.

[unclewebb]

Thanks everyone for bringing this issue to my attention. It was a pretty minor bug and easy to fix but I didn't notice it sooner because I've been using an Nvidia card lately. Version 3.20 looks like it's ready to be released next week.

No errors at all with win7100 x64.

Do you know if Windows 7 Build 7100 still changes the order of the icons in the system tray? In Vista and XP, RealTemp keeps the System Tray icons ordered so it is core 0, core 1, core 2, core 3, GPU but in Windows 7 the order is random. Unfortunately there is nothing I can do to fix this.

[gymenii]

Do you know if Windows 7 Build 7100 still changes the order of the icons in the system tray? In Vista and XP, RealTemp keeps the System Tray icons ordered so it is core 0, core 1, core 2, core 3, GPU but in Windows 7 the order is random. Unfortunately there is nothing I can do to fix this.

It's reversed, Core1,0. Not that it matters much since I only have 2 cores, but it might be annoying for quads.

[Draxx]

All works now Uncle

[unclewebb]

Thanks Draxx.

It's reversed, Core1,0.

That's a Windows 7 issue. In XP and Vista, when software inserts an icon into the system tray, it always gets inserted at the end of the line to the left of the previous icon. It's easy enough for RealTemp to insert icons in the reverse order so that they show up in the correct numerical order.

With Windows 7 they changed that. Now when you insert an icon it can randomly end up on the right side or on the left side of the previous icon so there's no easy way to predict or control the final order. Kind of a pain with Quads. Maybe Microsoft will come up with a fix for that in Win 7 SP1 but I doubt it. By then we'll all be using 8 core CPUs and then we'll have a real mess of unorganized icons to look at.

[WaterFlex]

3.20 works fine

[RyanH]

Could someone explain to me why Realtemp is showing a lower cpu multiplier then what I have? I have 8x set in BIOS and Cpu-z and Everest show that but Realtemp shows 7x I have both C1E and EIST disabled. It doesn't change either when under load w\prime95 small ffts. I'm sorry if this has been explained before I couldn't find the answer in this thread after a bit of searching. I'm running Vista SP1 with a C2Q 9450 and the latest versions of these programs.

[unclewebb]

RyanH: Can you post a screen shot of CPU-Z and RealTemp while Prime95 Small FFTs is running?

You can upload it to www.imageshack.us and then post a link here.

At idle RealTemp is sometimes different based on settings in your Power Options but at full load they should show the same thing. If they don't then obviously there's a problem that I need to look into.
http://www.fileden.com/files/2008/3/...alTempBeta.zip

Can you also check your multi with this tool:
http://www.fileden.com/files/2008/3/3/1794507/Turbo.zip

It's designed for Core i7 and is still a work in progress for Core 2 processors but should give me a clue as to what's going on.

[RyanH]

Heres the info uncle thanks for the help. I'd also like some help on calibrating my TJMax as you can see my sensor test too. Should I change my idle calibration or raise my TJMax? I've read you believe that the TJMax values could be different on the 2 separate dies on Core2 Quad MCMs but I'm wondering if they could be different on each individual core as well what's your take on that? My core 0/1 are very close across the load spectrum but core 2 and 3 are Also is it possible for TJMax or TJTarget to not be a factor of 5 like say 101 or 103 or something? Has Intel ever said anything on the matter? I don't know why the turbo program is showing EIST enabled I just double checked my BIOS still shows EIST and C1E disabled. I'm set to High Performance Power Plan but I haven't changed any of the advanced power settings.

[unclewebb]

The Turbo program reads the EIST (Enhanced Intel Speedstep) bit directly from your CPU and on your board, it's reporting that EIST is still enabled. I know there are a few boards and bios versions that do not turn EIST off correctly.

Could you try double checking EIST by using RMClock 2.30?

Version 2.35 seems to have a bug and doesn't close properly so I prefer:

RMClock 2.30
http://cpu.rightmark.org/download/rm...0_bin_upd1.exe



It can also report your C1E status and will let you toggle that. EIST can only be displayed. When I turn EIST off in my bios, RMClock shows that it really did get turned off.

I don't have any ideas yet why RealTemp is only showing a multiplier of 7.0 while running Prime. I'm definitely interested to find out why though so I'll try to update my Turbo tool by tomorrow to properly support Core 2 CPUs so we can gather some more data.

Can you try using RealTemp 3.20? There's a link to it in my previous post.
I don't think this will fix it but then at least I know we're using the same version.

You've obviously read some of this thread so I guess you've read my opinion about 45nm Core CPUs and their sensors. The bottom line is that TJMax is not a fixed number. The information Intel released last year about TJMax was pretty much useless because they didn't include details like their calibration procedure or how much error is in their TJMax number. Calling it 'X' on their graphs only made me think that the amount of variation is a number larger than their lawyers would let them admit to. The Core i7 sensors are better so most users have given up on this issue.

My best guess at the truth based on what Intel stated last year in their IDF presentations and based on my testing and 101 user screen shots is that TJMax on 45nm Quad CPUs may have a range of 10C. If 100C is their target, I'm not sure if TJMax falls in the range of 90C to 100C or 95C to 105C or maybe 100C to 110C. I believe that this amount of variation can be from one CPU to the next of the same model number and it can also be from one core to the next on the same 45nm Quad CPU. The variation from core to core doesn't have to be an exact multiple of 5.

These sensors have multiple issues and weren't designed or intended to be used to report accurate core temperatures. They were only supposed to be used to control thermal throttling and for that purpose they work fine.

If I find where your 8.0 multi has gone to then I'll try to figure out if your core temps are reasonably accurate. Stay tuned for an updated tool or two.

[RyanH]

Ok heres an update with the new beta. For some reason I couldn't get the RMclock app to run I get an error message as you can see on the left side. I have it installed in my Program Files (x86) folder as are all my other programs and I'm running in administrator mode so I don't see what the problem is. Am I suppossed to copy any of the .sys or .dll files anywhere or are these programs all self contained I'm talking about Realtemp and iTurbo as well.

[unclewebb]

With RealTemp and i7 Turbo, you can drag the folder where ever you want and it should work.

For RMClock, I'm using Vista 32 bit and didn't have a problem. I think the folder it creates in Program Files can be dragged to your desktop. You could try that in case it's some sort of permission issue.
Maybe I'll try it in Windows 7 x64 tomorrow to see if it works.

I did some testing with my Q6600 and here's what I got.

At idle with EIST and C1E enabled and with the Minimum processor state set to 50%, this CPU idles down and uses a multiplier of 6.0 like Intel intended.

The multiplier can be read from a model specific register (MSR) or it can be calculated using high performance timers within the CPU. The calculated method has a slight amount of error at idle on some CPUs but at full load during testing, it locks on one number and doesn't move. If you are using a Core i7 and the calculated multiplier starts to sag at full load then that is an excellent sign that Turbo mode is starting to cycle on and off.




For the second picture I started Prime95 Small FFTs. The calculated and multiplier read from the MSR were very steady at 9.000.




For the third picture I set the Minimum processor state to 100% and disabled C1E.
EIST was still enabled but on this board at idle, that's enough to get the multiplier steady at 9.0. There is a slight amount of variation in the calculated multi but not much.




RyanH: I updated the Turbo tool tonight so it should work OK with Core 2 processors. I didn't see any issues when testing my Q6600.

http://www.fileden.com/files/2008/3/3/1794507/Turbo.zip

Maybe try running only a single thread of P95 Small FFTs and run the Turbo tool and see what it says. Hopefully another user with a Q9450 will post a screen shot or two for comparison purposes. RealTemp could be out to lunch but this is the first I've heard of this problem and it seems to work fine on my Quad so I'm not sure what's going on, yet. Have you updated your bios and does your bios support your 45nm Quad CPU?

[burebista]

For some reason I couldn't get the RMclock app to run I get an error message as you can see on the left side.

Replace your RTCore64.sys in RMClock install folder with this.