High Temp question

PC eye said:
PC Wizard shows 26C for the cpu temp there. With the latest SpeedFan you now see cpu, chipset, ambient, hard drive unlike all previous version where the cpu and board temps weren't identified. In PC wizard you simply look down lower at the processor temperature. For a 3rd opnion or to simply compare the reading with a 3rd tool against what is seen in the bios the SIW tool displays readings when choosing the sensors item. http://www.gtopala.com/siw_on_pc_world.html
Click to expand...

Right, it does say 26*C, but according to SpeedFan, my CPU was 36*C and something else was 26*C. I just uninstalled/reinstalled SpeedFan and this is what it's giving me on it's first open:



...leading me to assume that PC Wizard is getting it's Processor and Power/Aux temps confused. Or it SpeedFan getting confused? Or am I a complete idiot and not seeing any sort of sense that this is making? :P
 
Also, with SIW, it only shows my Nvidia card and HDD, both of which show consistent temps with both other programs.
 
I just confirmed it. PC Wizard is labeled the wrong way. I did a Burn using SIW and according to PC Wizard, CPU temp stayed flat at 26*C while the Power/Aux shot up identically to what 'Temp 3' (formerly labeled CPU) was reading on SpeedFan.
 
The SIW tool is pretty reliable and one of the best tools for temps and other types of hardware and software information. You'll note that SpeedFan only shows one cpu temp while the SIW tool's sensor readings break it down into core #0 and core #1 for dual core cpus. The core #1 temp is also what is seen when going into the bios's hardware monitor.
 
I don't see any CPU temp reading using SIW. Only GPU and HDD are shown. I looked at each core listed also under the CPU part of Hardware and it gives no mention on any temps. odd.
 
If you look closely at the screen shot here you can see the CPU CORE #0 and CPU CORE #1 directly under the temperature heading. Note the highlighted item "sensors" right under motherboard in the hardware section in the left column to verify you selected the correct item there. Those are not GPU but cpu temps being shown for both cores here.

 
Don't have that in mine. Did you have to do anything special to get it? Mine shows just the GPU and HDD under the same area (Sensors).
 
You can't read video card temps without a sensor installed on the card itself. I think you are misreading C.entral P.rocessing U.nit as GPU when that's the two core temps being shown there. Are you running a single or dual core cpu there?
 
I don't know how true this is, but I read it somewhere.

After applying the thermal compound, you should give it a couple of weeks for it to cure to get the best temps.

Anyone know how true that statement is?
 
The breakin period is the thing you are referring to there. For Artic Silver the manufacturer points at a 200hrs. of total use to see the thermal paste become fully conductable for heat transfer. That was readily seen on a new build here when the first board quit after just 3 days of use. The AS-5 was adhered to the copper base of the Zalman 9700 while only a tiny bit was found on the cpu's heat shield. The heat the hsf saw had already seen a bonding begin.

For the first few days especially you run everything at the basics meaning no ocing or running large apps or heavy gaming where any large load is placed on the cpu. That prevents the cpu from overheating. Different thermal pastes or compound will see their own period where you have to look over their information for this. Some like Liquid Pro are more epoxy based and usually see their adhesion much faster for the one time application.
 
PC eye said:
You can't read video card temps without a sensor installed on the card itself. I think you are misreading C.entral P.rocessing U.nit as GPU when that's the two core temps being shown there. Are you running a single or dual core cpu there?
Click to expand...

I *promise* you that the info I'm giving is true. Refer to the above Screenshot in my above post. The fresh install of SpeedFan shows *no* CPU area. On my SIW screen (I was in the same exact area as yours referenced) I had no CPU Core 0 or CPU Core 1. It did, however, specifically list 'Nvidia' for my GPU and my HDD as the only two temps it could see. Every other sensor was a ghost to it.

I said screw it, wiped the entire drive from all the testing I was doing, and reinstalled XP. Sitting at 36*C at idle, and hope to see it drop back down to 30*C by tomorrow night or Friday.

I'll post the MoBo results tomorrow--way too tired to do it tonight after the ordeal that is setting up XP.
 
If you were seeing that on all three programs with PC Wizard not refreshing what it was seeing for a temp you probably had a bug in there. But keep the links for those tools to see how things are when the fresh copy is running good. It's not too uncommon for one or more of these to see problems due to being software. But keep an eye in the bios to see what is shown there.
 
Ok, refer to earlier in this Thread to see why I was testing this if you haven't already. My ATX-sized Abit was reading a much higher temp than my mATX-sized ECS, so I got to wondering how much MoBo size had to do with proper cooling.

Test subjects:
ECS 945GZT-M (mATX) Link
Abit IL9 Pro (ATX) Link
Intel D945GTPL (mATX) Link

Test system:
Pentium D940 3.2
Masscool HS/Fan 2700 RPM
Ultra 450 Watt PSU, 1 120mm fan blowing into case
1 80mm intake fan
1 80mm exhaust fan
eGeForce 7100GS (mounted with old chipset fan on top of HS blowing air away from card)
Windows XP SP2, completely updated
All test results obtained from SpeedFan. I did not think of cross-referencing with PC Wizard until after results were compiled, and did not wish to begin again :P

Isolated would-be variables:
SMART Fan disabled
All testing performed without OC'ing (if applicable based on Board)
All testing performed with EIST enabled (for my CPU, speed would cut back to 2.4 GH)
Board-specific Drivers CD fully loaded before test, then all drivers for previous Board completely removed.

Results (all temps recorded after being idle for 10 minutes):

ECS Temps (Min/Max)
CPU: 29/41 Average: 35
HDD: 34/35 Average: 35
GPU: 40/51 Average: 46
Board (Chipset?): 25/26 Average: 26

Abit Temps (Min/Max)
CPU: 38/40 Average: 39
HDD: 33/34 Average: 34
GPU: 46/48 Average: 47
Board (Chipset?): 27/29 Average: 28

Intel Temps (Min/Max)
CPU: 40/49 Average: 45
HDD: 34/36 Average: 35
GPU: 42/52 Average: 47
Board (Chipset?): 31/36 Average: 34

As you can see, the ECS posted lowest temps pretty much across the board. The Abit did post a slightly lower HDD, however. At this point, I'm not convinced that systems using ATX Boards just 'run hotter' than those using mATX, as the Intel posted the highest temps by far of all three tested. The only part I noticed not to be moderately (by more than 2*C) affected was the HDD, which only fluctuated by 1*C in all three tests. Also, the average temp of the GPU was only off by 1*C, however the temp range during testing fluctuated greatly.

After this test (and finding out that I could still use Ubuntu after previously thinking I couldn't due to the x4 PCI-E slot), I decided to keep the ECS. Had I to choose between getting the Abit, or Intel, the Abit would have been selected, as my experience with this Intel Board was a nightmare; I'll stick to their CPU's, but end the relationship there. The Intel BIOS was very poorly-laid out and offered very limited manipulation for a board of it's price ($85 USD). The Abit had an easy-to-use BIOS, but was somewhat limited in it's manipulation, though did offer more choices and an easier configuration over the Intel, making it worth the extra $5 ($90 USD). Oddly enough, the ECS (priced the lowest at $40 USD after a $10 rebate) offered the most versatile BIOS and included ways to OC, something the other two boards did not, making this the ideal choice for the average to slightly above average user. In fact, after playing very briefly with OCing after test results were obtained, an OC to 3.5 GH still showed lower temps than the Abit at 2.4 GH Idle.

Hope that helps. I know I learned a few things while doing this test. Smaller Boards do not always contribute to lower temps, and Larger boards do not always offer more useful features to the average user.

--Jay
 
Larger boards "can" offer more features while not always seeing them. Abit is a good old reliable make of board while the more popular Asus. Gigabyte, MSI boards will often see just one thing mainly more features on most while not all models over Abit. Don't forget that these other makes also see Intel as well nForce, VIA, SIS, ATI, and now AMD chipsets.

Intel cpus are also known for running warmer then the AMD models you would compare to. Intel addressed the heat issues with the P4 line with now lower temps seen on the newer Core 2s and quad core models available. Besides cpus even with the same socket type and even chipset don't forget one other thing.

What would that be? Board design. Even with different models of the same make power distribution and other things will see various temps between boards. Now compare a full sized with more features seeing more power drawn through it for more devices over what you would see on a micro atx model. But you could also compare various micro models to see which ones have more things packed onto certain ones over others.
 
I also really don't think that there is any issu that relates high temperature to the motherboard. It may be the case of the casing used or the location of fan that might be causing the high temperature.
 
A bigger motherboard is not going to run a higher temp. In reality a micro board will probably run a little hotter since everything is crammed together closer.
 
Different chipsets will see variations as well as how the air flow is directed. Serious Chatter has a good point there as far as air flow is concerned. One thing I like about the Antec 900 model case so far is not seeing warm air trapped at the top plus having twin 120s for front intake. The 939 board from the old would love being in the Antec.
 
You must log in or register to reply here.
Back
Top