I'have been a proud owner of HP's finnest nc8430 ^^for over a year now. It really is practical monster. It was used for writting assingments for colledge, as well as for playing games such as Oblivion, Counterstrike, M&M Dark Messiah. There is only one problem. It gets really hot when playing games. My temerature sensors indicate, that CPU core temperature lifts up to a 100 celzius while playing games. At first I didn't believe such data could be true, but then I measured the temperature of outgoing air and listed it as 75 celzius. I'm not worried for it to burn itself on fire, however, my left arm feels cooked after so many hours of playing and touching hot left side of the keyboard.
T7200 is listed as 31 W power eater. I have measured following consuptions
idle/monitor off - 32W
idle/monitor on - 41W
both cores burning - 72W
dual prime and continous 3dmark06 - 84W
On intel's homepage I noticed procesors like U7600 that have 3 times lower power consuption. So my question is following:
CAN I ACHIEVE AT LEAST NEAR TO SO SMALL POWER CONSUPTIONS WITH MY T7200 UNDERCLOCKED AND UNDERVOLTED WITH CPU RMCLOCK UTILITY?
If i could lose 10w it would be just awesome.
Other specifiactions of my notecomicbook^^
t7200 2.0 GH 4MB cache
2x1 GB DDR2 667
chipset 945M
mobile radeon x1600 256 MB
......
I forgot to meniton that situation in which it is used is only on AC power mode and haven't been using battery since the day I bought it. Of course, when on battery, consuptions are much lower and temepreatures also.[/b]
Undervolting HP nc8430
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
-
buzzlightyear
- Posts: 149
- Joined: Tue Oct 28, 2003 8:35 am
I have a T2300 (Core Duo) on my Toshiba lappy and I could lower the default voltage from 0.95-1.25v 0.95-1.0125v using rmclock.
In fact, since Intel locked the lower end voltage to 0.95v, I ended up locking the lower-end multiplier to 8 running at 0.95v (instead of default multiplier of 6)
What I am saying is that you can easily lower the high-end voltage of C2D down to, say, 1.1 to 1v.
In fact, since Intel locked the lower end voltage to 0.95v, I ended up locking the lower-end multiplier to 8 running at 0.95v (instead of default multiplier of 6)
What I am saying is that you can easily lower the high-end voltage of C2D down to, say, 1.1 to 1v.
Tnx for sharing your experiences BUZZ. Yeah, I was going for lowering the higher end of speed step configuration. Lowering voltage for 0,075 sounds good. Hopefully my procesor can do that. I was even thinking of slowering it down a bit, lets say 100Mhz, if that could bring voltage down for another 0,05V.
Ok. I'm back with full report. First of all I have to say that I am absolutely astonished by my merom t7200. With rm clockmark I achieved 1.000V at 2.0 GHZ rock solid running 2 PRIME simultanously. It's amazing how that lowers your power consuption. I went forward and found out that it's also stable at 0.950V runnig at 1.8GHz, which I decided was best, cause temperature of both cores, while under FIRE, don't exceeds junction temperature (100 celzium), and therefore TDP keeps frequencies at maximum at all times.
Here is a good example. If I ran one prime that put one core at maximum load at 2.0ghz 1.188V, waited few seconds for a temperature get close to junction temperature and then run superPI 1MB, it calcullated 38 s, which is awful for a core 2 duo. But when I switched to 2.0Ghz 1.000V, ran one prime and one superPI 1MB, it calculated in 27s. If I run only one superPI 1MB with no additional arson, I get 25.5s. Slower times in first example are due to TDP regulating lower frequencies while under load, not to exceed junction temperature and therefore keeping your sistem safe from burning up.
Here is a good example. If I ran one prime that put one core at maximum load at 2.0ghz 1.188V, waited few seconds for a temperature get close to junction temperature and then run superPI 1MB, it calcullated 38 s, which is awful for a core 2 duo. But when I switched to 2.0Ghz 1.000V, ran one prime and one superPI 1MB, it calculated in 27s. If I run only one superPI 1MB with no additional arson, I get 25.5s. Slower times in first example are due to TDP regulating lower frequencies while under load, not to exceed junction temperature and therefore keeping your sistem safe from burning up.