Printable View
Now I understand what you're saying Bludd. I've always assumed most users of RealTemp have at least one temperature icon displayed in the SystemTray area but it would be easy enough to create some more minimize / close options and a picture icon for those that don't.
Thanks. Don't feel you have to hurry, it's not a huge deal, but some improvement that can be made in future versions time allowing. :)
Min/Hide to Systray/Taskbar Feature Work Great :)
And Honestly, I wouldn't change a thing.
Congrats Unclewebb for a superb program.
Hi unclewebb :) Thanks for updates and thx for new relase :)
some tests on my rig .. just installed RT 3.0 onto my (multi boot) Asus RE /(b.ver1003) // qx9650/ 2x2gb (a-data) ..aircooled rig
run it on separate OS's -
1) Xp pro (32bit) /
2) Vista HomePrem. sp1 (32bit) and
3) W7 beta(64bit) 7000.0.081212-1400_client_en-us_Ultimate-GB1CULXFRE_EN_DVD.iso
so far seems all OK ..hmm i didn't got clear the problem ( upper posted) about ''close minimize'' ok what ever ..
anyway no problem for me, so far all is ok here .. (ok :) for me personally )
just ''Ph.shopped'' together screens capped from diff. Os 's on same hW /settings
all looks OK :)
only some annoyances .. in W7 beta -- ''RT icons in systry ..(well.. i remember ..it was posted/discussed earlier .. )
these icons (order) just do not save (or os cannot recall?) their set order .. on next boot,
or other words (no prob. can drag them into right order).. no big deal .. but this seems OS(w7) problem not RT ..
anyway .. on others os's like xp -vista - RT 'systry icons' get aligned as they should be in logical order of cores (apic id - 0123).. on every boot up
..
http://img407.imageshack.us/img407/5...1200iv8.th.png
http://img407.imageshack.us/my.php?i...pyw1200iv8.png
-------
also .. just noticed that Mhz. calc utility (program) get somewhat ? wrong readins here
specially on vista 32 bit and w7 64bit .. - at same time on XP32bit .. it reads the (fsb, multi.) setings as was set in bios ( currently have no options to test xp64bit and vista 64bit ) sorry if this false alert.. but honest... haven't seen this earlier
.. only changes in my hw is RAM - replaced current 2gb kit ->to 4gbkit aka (current was 2gb(2x1gb) to-> 2x2gb (4gb)
just have tested QX9650 @3600 / 9x400 // strap333 / DDR3-1600 with my current (4GB kit ) A-data 2x2gb ram kit ..
first thought .. could be problem on 32bit os with 4GB .. but later on 64bit w7 - same ''misreading'' appears
so far, at the moment cannot test other ram kits (my def cellshock (2x1gb) 2gb ddr3-14400 kit)
unfortunately one of the CellShock's modules have ''passed away ''
so only options currently is this cheap 3x2GB A-data (DDR3-1600G) pc 12800 (nehalem)kit ..
but using only 2 modules now
I've tested RT3.00 on XP, Vista and Windows 7 32 & 64 bit and overall everything is pretty good. So far the only major bug is in Windows Server 2008. I'm in the process of upgrading my Visual Studio version and SDK but I don't know if that will fix the Windows Server 2008 missing cores bug. I have a few other options so as long as darkzone is patient, I should be able to come up with a solution for him.
The only RealTemp Windows 7 bug I found is what you mentioned. The System Tray icons do not get lined up in the correct order. The exact same code works 100% in XP and Vista. Presently when software inserts a new icon into the Windows 7 system tray area it can get inserted anywhere and the next one that gets inserted can get inserted either to the left or the right of the previous one so there is no way to get them properly lined up. The final order is random. Most software only inserts one icon into the system tray so maybe someone at Microsoft was hoping that no one would notice. I haven't noticed how Everest v5.00 is working in Windows 7 which inserts a lot more icons than RealTemp does. I'm not going to worry about this issue until Windows 7 final comes out. I sent a bug report to Microsoft but it's hard to say if they'll make any changes. It's probably not very important to them so it might be saved for Win7 - SP2 or SP3 or might be fixed up and sold as a new and improved feature in Win8. :) If you want this fixed up, more Windows 7 beta testers need to complain.
The Minimize on Close issue is going to be a non-issue for most users. Bludd would like a few more options when he is using the TaskBar option combined with the Minimize on Close option and that's been put on the things to do someday, maybe list.
It looks like my old MHzCalc tool needs an update. The good news is that RealTemp 3.00 appears to be working 100%. 3603.93, 3603.89 and 3603.92 in three different OS versions looks correct. Those are the exact same numbers that the King, CPU-Z, is reporting only RealTemp is a little more consistent. ;)
I found out that after XP, Windows changed the clock generator it uses to calculate time intervals. It looks like in my MHzCalc testing tool, my fix for the bugs of XP don't work very well after Microsoft got around to fixing their clock generator code. Do you use the MHzCalc tool? If so then maybe I'll update it with the code RealTemp is presently using to calculate MHz.
New try with new build, nothing is changed, same issue of two missing cores :shrug: But, if the older version still working, the right question is what make the newer version to not run correctly? I think the operating system is OK by functionality, I didn't have any problems with it in these 65 days from last shutdown (when I've mounted Musashi on 9800GT), as you can see the uptime in upper right corner.
darkzone: No worries. It probably doesn't seem like it but I'm making progress. I'm finding out all the things that aren't causing this problem.
The Thomas Edison approach :up:Quote:
Originally Posted by unclewebb
Ive been using the 3.00 for a few days now and there are no problems yet on my end, thank you very much for the updated version.
Your welcome Demo. The issues are pretty minor so far. Hopefully in a day or two I'll figure out what's causing darkzone's problem in Windows Server 2008.
sorry but tried tis program and i thik this may wrong its imposible to me to be 7 degrees, Cº with and ambient temperature of 19º, (i'm watercooled but i think it is not as it shows)
danielniel: Thank you for providing me with lots of information so I can help you diagnose your problem. Have you read the documentation yet about calibrating and how inaccurate these sensors are at idle? That's always a good place to start.
http://www.techpowerup.com/realtemp/docs.php
Tell me a few more details about what processor you're using, etc., etc.
The program is only as good (or bad) as your sensors, just like all the others :) Provide some information as unclewebb said.Quote:
Originally Posted by danielniel
Its been a while since I posted, glad to see you are still refining and tweaking this program. Hats off to you Unclewebb for all the hard work you have done. I just upgraded from RT 2.7 to 3.0 last night, recalibrated and used your new cooldown test. I have been reading all the recent post in this thread and didnt realize that you can have a different tJmax on each core. I have my tjmax set at 95 on all 4 cores, which is what I thought it was on the q6600. I believe that I have done my calibration correctly, but under load cores 0 and 1 read 4-5 C higher than cores 2 and 3. Is this a case of adjusting tJmax for 2 of the cores? And which two cores would it be?
[IMG]http://img34.picoodle.com/img/img34/...nm_de1031e.jpg[/IMG]
Here is load pic
http://img32.picoodle.com/img/img32/...dm_708b884.jpg
UncleWebb, are Atom processors supported? I just ordered Aspire One and i'm wondering if i could use RealTemp on it to monitor CPU temperature.
Hi RejZoR. Long time no see around here. :)
The Atom is supported. There are a couple of different Atom models including some with hyper threading but I believe that RealTemp can handle all of them correctly now.
There is one bug / incompatibility that I know of but I'm not sure how many or what model of Atom that this bug effects. The Load meter doesn't work at all and only reports zero percent, all the time. The Load meter depends on some internal timers in the Core processors so my best guess is that maybe the Atom doesn't have these timers available. Let me know what you find or if there are any other Atom owners can you post a screen shot of this. The Intel documentation says this should work but one Atom user had issues.
Nevrsadie: Originally RealTemp was based on my IR temperature measurements of the heat spreader (IHS) which covers the cores. rge's testing has shown me that the actual core temperature is likely about 5C higher than the IHS temperature. That's why the default TJMax has been bumped up to 100C for the Q6600 G0.
Intel admits that TJMax is not a perfectly fixed number. They have what they call a TJ Target value and then actual TJ Max is somewhere around there or slightly higher. My theory is that there seems to be a lot of processors where TJMax is not consistent from core to core. It's very easy to see this on the 45nm Quad chips where 4 cores can be all over the place. With the 65nm Quads, I have seen a lot of screen shots where at higher temperatures running Prime95 Small FFTs, core 2 and core 3 will report exactly 5C less than core 0 and core 1. The only explanation I can come up with is that actual TJMax is 5C higher on those two cores.
That's how I have my Q6600 G0 set up. TJMax = 100, 100, 105, 105 and then I use idle calibration factors to equalize the temperatures at idle. Doing this gives me 4 cores that track each other very well from idle to TJMax.
When you have 4 cores covered by the same heatsink, all running the exact same consistent load, I think the core temperatures should be pretty consistent across all 4 cores.
It's possible that my theory is out to lunch. Maybe in the Quad design at full Prime load, core 2 and core 3 are getting less voltage and creating less heat because of internal vdroop or something like that. Maybe the IHS makes better contact with core 2 and core 3 so they can transfer heat better and run cooler. :shrug: You might be able to explain a Q6600 with those two theories but there seems to be 45nm Quads that are more like TJMax = 100, 105, 100, 105 or even 110 which can't be explained by voltage or IHS issues.
All I can say is keep an open mind and give my way a try and see how it looks from idle to full load. Now that Intel has confirmed that TJMax is not fixed, I don't feel so bad using different TJMax values on different cores.
The Coolest has been adding lots of useful features to Core Temp lately including Low/High temp reporting, MiniMode, the ability to offset individual cores and some thermal shutdown options.
http://www.xtremesystems.org/FORUMS/...93#post3636493
I'm a little biased about what MiniMode I like better but it's worth a look. Competition is good for both programs. With RealTemp you can grab on to the edge of the window frame when in MiniMode and drag it up or down to have a look in there. Double right click to check out Anchor Mode and hold down the shift key when you double right click to set your own Anchor position.
http://img21.imageshack.us/img21/4534/minimodena8.png
I tried 100 100 105 105 and it gave me high temps at idle. My room temp is 22.5 and idle temps where about 12C higher.
http://img37.picoodle.com/img/img37/...5m_3b88416.jpg
I then used 95 95 100 100 and it worked out almost perfect. I had to adjust core 2 and thats it.
http://img32.picoodle.com/img/img32/...0m_8405108.jpg
I went ahead and tried your suggestion of 100 100 105 105 and calibrated from there.
The weird thing is after calibrating, then overclocking, my full load temps are still off by 5C between cores 0,1 and 2,3.
http://img18.picoodle.com/img/img18/...dm_bbfabe6.jpg
Yeah, it has been a while :) I'm kinda busy lately, but i'm also lazy and i usually circle around forums. I just moved like 12 of those that i rarely visit to another bookmarks folder.
I think i still have over 20 of them left in the "active" folder. I'll check it out when i get my Aspire One to see if RealTemp works fine with it.
Nevrsadie: Did you have your case open when you ran it at low voltage. The reason I ask is because your minimum temperatures are showing up as 30, 30, 27, 30 in your screen shot with TJMax set to 100, 100, 105, 105. The 30C numbers in a 22.5C room seems about right to me. I'd use my TJMax values and then you'd only have to bump up core 2 a little bit with some idle calibration correction.
My Quad is similar that when I run Prime95 for a few minutes, there is still a difference in temps. Not as big a difference as what you're seeing. As I run Prime95 small FFTs longer, core 2 and core 3 start to catch up. It will forever be a mystery what is really going on here. I'll go run some Prime and post a screen shot or two so you'll have something to compare to.
Case open, matter of fact I just used some canned air to spring clean my HSF and case fans.
Here's how my Q6600 looks at idle using split TJMax and some idle calibrations factors to keep things balanced at idle.
http://img24.imageshack.us/img24/3845/rtidlese7.png
After a couple of minutes of Prime with my Tuniq fan on low, there is a 3C difference between the two sets of cores.
I'm using 1.384 volts to create some heat and keep it stable. My Q6600 is no prize winner.
http://img15.imageshack.us/img15/7001/rtload1oa5.png
I turned up the CPU fan to its highest setting and a couple of minutes later, all 4 cores were closer together.
http://img8.imageshack.us/img8/4941/rtload2gm4.png
A half hour later, they were about equal across all 4 cores.
http://img4.imageshack.us/img4/2180/rtload3hm5.png
I turned the fan back down and a few minutes later, there was a difference again.
http://img4.imageshack.us/img4/9486/rtload4li5.png
Maybe this proves that some of the difference is caused by cooling efficiency and heatsink design. I'm not smart enough to understand heat pipe design.
Previously I turned my computer upside down to see if the weight of a large air cooler was causing some of the difference but that didn't seem to change anything. I've also laid my computer flat on the ground, case open and there is still some temperature data that is hard to explain. The only thing that covers all of what I've seen is s.h.i.t happens. Sorry about all the white space. Vista Paint sucks when you're used to using Photoshop.
Unclewebb, if you're using Alt+Print Screen, click Image -> Crop after pasting into Paint. One of the few Paint changes from XP to Vista, but one of the most useful. :)
I used ALT+PRINT SCREEN but didn't notice that with Paint I also need to do a crop until after I uploaded them all so I said, it's good enough. With Photoshop you just Paste and save and you're done. Paint in Windows 7 is finally looking a little more modern.
I'm on a spare hard drive at the moment trying to come up with a fix for darkzone. I re-wrote a chunk of RealTemp and have switched to a newer compiler. It's almost ready for him to do another test. If 3.00 is working fine for you then there probably won't be much reason to download the next beta when it comes out.
Install Paint.net if you want a good and free Paint replacement.
What is Paint? IrfanView FTW. :D