Hmm these temps sound ok??

[stang8118]

Ok i just installed h20 to my computer and the first CPU i used was an athlon 3300 64bit @ 2.75ghz using 1.8v. (only 256k L2 cache 64bit cpu, found in HP/Compaq computers), and it ran at 25c idle 33c load. Now i have just sold that and reinstalled my 2800 64bit cpu and it idles at 38c and load at about 46c @2.5ghz using 1.55v. That seems like a pretty big jump considering the only differance between the two was that one has 256k cache and the other has 512k.

I have remounted my cpu block twice, and added the same amount of AS5 i used for the 3300. I just don't understand how the extra L2 cache pumped up my temps 10-15c, when i am using .25 less volts to boot

My h20 setup is an RBX all copper block, 1/2" tubing, Eheim 1250 pump.

Also i have no kinks in my system either, plus its 99.99% free of air bubbles.

 

[voigts]

I am running an AMD3200 which seems to run rather warmer than some other temps that I see around. I think I have a pretty good setup and my temps are just a little lower than what you are getting. The main thing from what I can tell is more how much of a difference there is between idle and load temps rather than the temps themselves. A good setup from what I read and can tell has no more than around 5c difference between idle and load temps. That is really more important than what your readings are. Different CPUs are bound to give different temp readings, and as you will read on these forums, motherboards differ a lot on how they report temps. If you have it seated well with AS5 and your idle/load temp difference is OK, I wouldn't worry too much about it.

 

[stang8118]

bump any ideas?

 

[Hoot]

Though I don't have any X64 chips to compare to, my first being this 3200 Winchester, I can tell you that they do run hotter than any Socket A chip I have ever owned, regardless of Vcore. Since temperature reading results vary from one motherboard brand to another, trying to obtain useful information comparing one brands temperature reading to another is pointless. Many motherboard manufacturers "Cheap Out" when reading the on-die thermal diode by using Winbond multi-I/O chips for the task. They violate even AMDs recommended reading technique by not using a dual-current method. They are however cheaper to implement and thats the big factor with motherboard manufacturers. Certainly not accuracy.

Hoot

 

[stang8118]

Thx Hoot, but this is on the same exact mobo. I just switched CPU's. Just seems odd that there was such an increase, i am not worried about it but was just curious.

 

[Jas]

what cores were the chips?

Newcastles run hotter than Winchesters.

 

[Craptacualr]

Yeah, it was prolly a Newcastle... The difference between the cores is the manufacturing process. Newcastle = .13u, Winchester = .09u... The smaller your transisor bridge the less heat going across it. Only other reasons I can think of are dust in your radiator, or a significant seasonal ambient temp change?

 

[Jas]

Or computer gnomes.

 

[Craptacualr]

Yeah those gnomes are nasty little bastages XD

 

[stang8118]

I killed all the gnomes that were infesting my house!! They were both Newcastle cores (acually the 3300 was a paris core, but it's just a newcastle with 1/2 the L2 cache), and 1.8v was getting cooler temps then 1.55. This just boggles me as how two CPU's could run at such differant temps on the same exact mobo.

 

[Craptacualr]

huh, both newcastles? thats kinda messed up. since you eradicated the gnomes, the only thing left I can think of would be differences in the heatspreader on the chips?