Nagamaki/Whitebook CPU temps with Prime 95

Disclaimer: Read through the entire post before making a conclusion, people seem to jump to conclusions as I like to tell the story, not just the end result. Read on

I ran about 27 hours of Prime95, with a 12 hour block, some shorter 45 minute blocks, and a final 18 hour block. This was done without turning off the computer, and only about an hour total not running Prime95. During my intial testing, I had some minor issues while I had the 11.5 multiplier set. I reset back to a 9,5 multiplier and had no issues. I have not to date had an issue while gaming with an 11.5 multiplier overclock (3.06 Ghz), but I find that in gaming the CPU is only used about 65-70%, and runs about 66 degrees C.

After 12 hours of running, this is what I got. However, be warned, I realized it was getting warm in my room after the first 10 mins, and I remembered my A/C was off. May have contributed to my earlier problems with the overclock. Anyway, I turned it back on, and I never saw temps get higher after that than the current temps in this screenshot.


When I woke up this morning, I then stopped the test I was running (Blend, temps above) and started running the small FFT's, (max FPU torture). With the A/C running, this was the highest I had. Screenshot was taken at about 45 mins in, I let it run for an hour and a half without seeing any change from this-


After this test, I started the test which is supposed to create the max cpu temps. I also decided to do some surfing and photo editing at the same time. Probably not a good idea, I ultimately think the core errors I got ( http://www.guruofgaming.com/modules.php?name=Forums&file=viewtopic&t=2630 ) may have something to do with this. With that much CPU usage and windows relatively poor memory management, it's possible that something got overwritten, or, with the FPU's at max temp, perhaps using other portions of the processor concurrently create localized processor overheating that they just can't handle. Either way, I then ran the Max CPU temp test through Prime 95 for a whopping 18 hours (without doing other work on the computer) and had no issues.



TZ00 CPUID graph of temp over time


Core 0 CPUID graph


Core 2 CPUID graph


If anyone wants to see all 4 core graphs, let me know, just figured that would be a waste of space, they're all pretty similar.

Interesting things I noticed-

Although the temps spike up to as high as 77degrees C, the fan doesn' t go to high until right about 70 degrees is hit. I initially thought this had to do with the fan coming on a little late, but from the graph, it just looks like various parts of the test produce different loads on the processor. It would still be nice to have the fan go to high a couple of degrees earlier so it doesn't spike as much, though.

Over the course of the 18 hour test, you can see that the average temperature steadily goes down. I attribute this to the slowly decreasing temperature in my room throughout the day and night, hitting it's lowest average in the morning right before I shut it down. As a side note, I'm located in Hawaii, so when the A/C is off, the outside temp even at this time of year is in the low 80's. With my A/C on, the temp in the room can at times get in the low to mid 60's.

Not each thread is completed at the same time. My core #3 and #4 have most of the background load, via the K|N COP, and it shows as they were farther behind in their tasks. Or maybe some processor cores aren't as smart as others?

It seems that in every test I run, the highest core temp is in the 77 degree range. I don't know how that is considered temp wise, but I didn't get any errors during this, and the computer still works fine. After over 27 hours of total test time, performed consecutively with no real rest in between, I have to say it's pretty solid. I would like it to do this with the 11.5 multiplier, but it doesn't seem to be in the cards for now. Not sure why that was, temps didn't get as high as they did here, maybe voltage is an issue? I may do more testing with intermediate multipliers, but for now, I think this shows some pretty robust capability from the Nagamaki and the Q9300.

 

Was this not interesting/ too technical? I won't post this stuff in the future if no one finds it useful, that is, I don't want to waste anyone's time, least of all mine! Thanks, have a great Navy day,

 

Naw Bro,...... GOOD stuff! Keep posting! ALL info is useful!

 

This is very useful actually because it gives us a sense of the thermals in terms of the load temperature and such. The Whitebook volts the Extreme series at the maximum threshold (around 1.2v), and I'm sure lowering them would decrease these temps in return for less stability when multipliers are of higher value.

 

Yes this is definitely useful. I don't even have the ocz whitebook and i still find it interesting and a good tool for comparing, plus also gaining knowledge about one of the best laptops available now.

 

Just wanted to make sure as there were no responses. I am in California right now and it's significantly cooler here, I may run a quick check to see how different it is.