- Joined
- Dec 18, 2000
- Location
- The battle born State
Initially this tweak did not work for me. I reported that here: http://forums2.overclockers.ws/forums/UltraBoard.cgi?action=Read&BID=4&TID=4618&SID=19487
Upon further trying it all of a sudden did work, after I completely turned of the PC and cold-booted from scratch.
Now the register read 6B, and it was possible to change it to EB, and after accepting the new register settings the idle temps dropped indeed by about 10C - unbelievable - well actually I had been wondering why they were not lower. On my PIII systems there usually is a difference of about 15C between idle and load, so why not with an AMD system?
Upon further inquisition of the board via MBM 5.05 I made the following observations:
The idle CPU voltage dropped from 1.97 to 1.93.
The 5V voltage, which feeds the CPU rose from 5.10 to 5.30.
The latter observation is an indication tht the CPU is drawing considerably less amperage while idle.
Upon loading SETI, the values dropped back to their original numbers, and the temperature under load rose back to its original value, may be 1C less but nothing serious.
The above observations indicate to me that this register indeed seems to invoke a CPUidle routine, now I am just wondering why this had been overlooked by virtually all board manufacturers...
Do they know something that we don't (e.g. is there possibly a downside)?
Yo
Upon further trying it all of a sudden did work, after I completely turned of the PC and cold-booted from scratch.
Now the register read 6B, and it was possible to change it to EB, and after accepting the new register settings the idle temps dropped indeed by about 10C - unbelievable - well actually I had been wondering why they were not lower. On my PIII systems there usually is a difference of about 15C between idle and load, so why not with an AMD system?
Upon further inquisition of the board via MBM 5.05 I made the following observations:
The idle CPU voltage dropped from 1.97 to 1.93.
The 5V voltage, which feeds the CPU rose from 5.10 to 5.30.
The latter observation is an indication tht the CPU is drawing considerably less amperage while idle.
Upon loading SETI, the values dropped back to their original numbers, and the temperature under load rose back to its original value, may be 1C less but nothing serious.
The above observations indicate to me that this register indeed seems to invoke a CPUidle routine, now I am just wondering why this had been overlooked by virtually all board manufacturers...
Do they know something that we don't (e.g. is there possibly a downside)?
Yo
