1) Hello, I compiled a new PC a week ago, with an E5200 M0 revision with 1.175v vid. In bios the cpu temp is 31-34 degrees, but hwmonitor and coretemp report 31-34 degrees for core #0 and 45-49 degrees for core #1.
In full load, after a few hours core 0 reaches 55* max and core #1 60*, so the difference is less far apart.
Is the temp sensor acting funny or is it possible i mounted the radiator poorly. When i first asembled the system, I applied paste 2 times (cleaned it ofcourse) and think i installed the radiator 3 times, because i was getting the same temp reads. Note: i use the intel lga775 quad stock cooler and fanis at 87.5% min speed, set in bios.
For Mobo, i use a MSI p35 neo combo flashed to the latest bios version when bought and reflashed jsut last night to make sure it was properly installed.
2) I also OC'ed a bit, to 3ghz 12.5x240 with stock vcore or up to 1.2v. It was very stable in assasins creed 1&2 but stable for 2 hours in modern warfare 2, at some time, I'd hear a windows error sound, game would freeze and after 5sec it rebooted, thought it was a game error.
Last night I tried Tropico 4, and after 20-30mins again i get the windows error sound , game freezes and it reboots, so i think it wasn;t jsut modern warfares 2 problem, it the cpu.
Does the cpu overheat? I don't know any cpu temp monitoring tools that logs data in a file, like gpu-z does, so can't really verify. Or does it need more vcore? My brother says until you set it to 1.35v don't bother me with "why it doesn't work" lol... A bit harsh vcore for a measly 20% oc i think.
You dont need to worry about your CPU being too hot until you see it go past about 75-80. 60-65 is very average for load temps on Intel Systems.
The 15 degree heat difference between cores is quite large, although a difference is not uncommon, 15 degrees is higher than what we would usually see.
Seeing as though the hottest temps recorded are only 60 degree's though, I wouldn't even worry about it
The re-boot problem is almost 100% down to the CPU over-clock. Its obviously not too stable. For a proper stability test use something like Prime95 and leave it running for about 5hours or so and see if any errors are returned.
Edit: ran prime now in small FFTs and it got an unexpected error at the 8th test. Max temps reached were 63 and 68 at 1.225 vcore 300x10. Is this a problem of prime95 stabillity or did windows xp shutdown prime to avoid a blue screen/restart?
It is a problem with your overclock settings, probably the memory.
The E5200 Wolfdales overclock like banshees. With a 1.175 volt VID, you have a better than average chip. My E5200 has a 1.2825 volt VID and the most I can push it with stability (my definition - 24 hour Prime95 run) is 3.87 GHZ. Some E5's will run past 4 GHz.
I am not really familiar with the MSI BIOS, but let's see if we can get this sorted out.
First, forget about maximizing the FSB freq. Set the CPU multiplier to 12.5.
Second, you didn't say what kind of RAM you have, but set the RAM voltage to 2.2 volts. Your option may be "Default + .4 volts). 2.2 volts is safe for DDR2 RAM.
Third, and this is important, set the FSB:RAM ratio to 1:1. At stock speeds, your FSB will be 200 MHz and your memory clock will be 400 MHz. Yes, this will seem low, but 1:1 is like the happy place where CPU and RAM are running synchronomously (a fancy word meaning "exactly together"). This gets you good stability and you won't notice the decrease in memory i/o.
Now you can start working the FSB upward. You should be able to get past 3.75 GHz easily.
You will not need a great cooler, but you will need something better than stock at those frequencies.
You can push the core voltage to 1.45 volts or until core temps at load reach 70 C.
I seem to have made it stable now. Upped the vcore to 1.235 with 300x10 and rams at 750mhz.
Sorry, i did want to post what ram sticks, but it wouldnt alow me to edit the post and i forgot about them. They are geil ddr2 800 5-5-5-15, they are not much but they were on offer with the mobo at a ridiculously low price.