Quote:
Originally Posted by theteamaqua
You can't judge the retail stepping from one or two pieces.
Printable View
Quote:
Originally Posted by theteamaqua
You can't judge the retail stepping from one or two pieces.
im sorry but unless the conroe has a built in heat sensor those are not ondie temps.
and ,, i dont know.. does it ?
for years boards have read temps way wrong, i dont know of intels history becuase of their lack of overclocking in the past.
when in doubt mount a probe and see what happens....
but i do doubt they are 10-20c off thats just crazy.
and if you guys are using the stock cooler then YES you can be guarenteed that your temps are wAY high... that cooler sucks hairy you know what..
get yourself a decent air cooler. you cant do anything with that intel cooler and intel needs to get off their collective butts and start giving use a decent heatpipe cooler like amd
for those on water if your temps seem right then ok..
also check to make sure your heatsink is contacting the cpu IHS properly.
you cant use math to calculate heat simply because there are too many other variables involved, mainly is the dissipation of the heatsink and if it has proper contact. how fast the fan is moving, is it aluminum or copper, how thick are the fins,, you can calculate heatoutput of the cpu using the raw wattages but again too many variables make that calculation unreliable.
a probe isnt ondie, its on IHS but that is still going to yield a pretty close temp.
someone needs to try and see if the mobo temps are reading incredibly different or only slightly, if its slightly then you can probably assume that the temps are correct since the probe is a ways away from the die, and only testing the edge of the ihs.
Just tried temp reading.
Remember that this is with an 805 not Conroe yet.
Temp reading taken from the side if the IHS which is covered by the base of my Si-120... the base is stopping any air from the fan reaching the probe.
results:
Idle: temp probe 35c AI-Booster temp reading on CPU35c - 37c
Full load temp probe 37c - 38c AI-Booster temp reading on CPU 41c - 43c
This is just one set of results, but it seems to show that the IHS heat gets disipated by the HS ok. I therefore deduce that the motherboard reads the temp from the underside and maybe that does not disipate so well.
Intel chips still have a void and collection of IC's under the centre of the core. AMD dont now. P4's certainly read thier temps from either on board diodes or from this void area. I think heat is building up there and struggling to escape. This may be a failing of the 775 socket?
They are indeed on die temp sensors built into the core.Quote:
Originally Posted by Lestat
The question is whether or not the the sensor chip is reading them right, or the software is reading the sensor chip right. Either way, the reported temps originate from on die sensors.
Thanks for posting that. Have we already forgotten ABit IC7-G?Quote:
Originally Posted by mdzcpa
Quote:
Originally Posted by Donnie27
true true
I too have a Pentium D 805 as I wait for a Conroe, and I am using the same motherboard as you - P5W DH Deluxe, and with the 0520 bios my idle temp in bios was 75C! Surely something must have been wrong, since the CPU was watercooled (Swiftech Storm - perfect contact).
-k0nsl
Quote:
Originally Posted by simonmaltby
From what it looks like your temps are actual temps.
Why did you use Arctic Alumina with Intel's HSF instead of the original Shin Etsu compound ?
My advice : get another thermal paste and apply it evenly and very thinly across the entire HS.
Well you should be able to tell from your overclock whether the temps are correct or not. At 75deg, you wouldn't get very far.Quote:
Originally Posted by k0nsl
Quote:
Originally Posted by savantu
he's not the only one getting some suspect high readings. I think there may be at least a partial failure in the temp reading sensors on at least some of these boards. I beleive someone who works for the company who makes the sensors stated earlier in one of these threads that there was a problem with the sensors. I am not sure if the badaxe and the asus use the same sensor though, but it would seem likley.
off topic.. followed that link and ended up on the dell site and saw thisQuote:
Originally Posted by Verisimilitude
http://www.dell.com/content/products...en&s=dhs&cs=19
Do u think there would anyway to get that case, Dells are standard ATX right?
When i look at it Ferrari springs to mind for some reason
EDIT
Its BTX
no dell isnt standard ATX unless they have changed things.
they use a custom power supply so you cant use a standard psu unless you cut the rear of the case.
as for motheboard fitting yes, typically there shouldnt be any issues as long as you get a tower and not a mid tower or mATX tower. Although the person i sold my 955 XE to has an XPS 600 and she said they had a mini ATX board in theirs and that my P5WD2-E Prem. wouldnt fit in the case so be careful
ATX is an industry standard and dell and others dont stray from that, but they do use other aspects of their cases, such as the psu to make is hard for other components to fit properly
*edit*
since i actually looked at the picture yes it should work fine.
but i doubt you'll find a case anywhere but ebay. you cant just buy their cases.
Well said. IMO the temps are just reading high. On water I'm seeing 68c loaded on my X6800 @ 3.8ghz with only 1.49v on an unmodded Bad Axe 3.04. Idle is at 46c. That's simply too high. My water block never exceeds 30c ever (and, yes, my block and IHS have excellent contact).Quote:
Originally Posted by xgman
In addition, by cranking down my rad fans, I can get the cpu temps into the 74-76c range easily running dual prime at 3.8ghz at 1.5v stable. There is no way it would stay stable at 75c.
Last, at stock volts and speeds, the chip still idles at 43c and loads at 65c. Not likely on water cooling.
I'm guessing the temps report too high by about 10c. Maybe more.
Just touch your heatsink, for example, my TT BT heatpipes were a bit warm only when temp reading were above 60C
Anyone?Quote:
Originally Posted by Mr.X
CPUZ says 6-11 so I think 11.Quote:
Originally Posted by Mr.X
it goes to atleast 13x and if you still want an answer read the entire threads about it. Or pm someone with a conroe if that doesn't work.
REPOST
There will be other boards and with luck, they'll use better equipment or better micro code.
Once I had 3 computers with the same WinBond chip. These were AOpen AX4C, Gigabyte i850E 8IHX or something and Abit IC7-G Max. The Aopen and Gigabyte were very close reading all 3 P4s, a 1.7GHz Willy, 2GHz NW and 2.4GHz (400MHz FSB from a Dell). The ABit and Gigabyte, are still being used BTW, consistantly read all 3 processors 12 to 16C high depending on room temp at IDLE. Under Load was a real joke. Once it showed the 2GHz Northwood that Idled at 32C most of the time, at 46C and my 2.6C at 56C at stock idle when the Aopen showed 49C while overclocked to 3.26GHz on air and 57C under load. I used a CNPS 7000Cu cooler. Newer BIOS did change some of the results but still left it reading higher than anything else.
The folks at Motherboard Monitor warned about this as well. I was on Abit's forum while they did nothing and said. "We're right and everyone else is wrong".
savantu - the thermal *pad* on the retail cooler sucked. It was basically two slender slips. It would definitely have been much worse.
For the naysayers:
A) I tried different grease compounds to no effect
B) I tried different heatsinks to no effect
I am using a TT Tower112 cu - it is a monstrous copper fin heat pipe connected heatsink that doesn't get blisteringly hot (thanks to the 1"1/2 92 mm fan running at 5200 rpm), or even lukewarm.
I am running in a chilled environment. I know what I am doing folks, and I am telling you either the chip is running hot or the badaxe is reading them too high. I have to say that while the badaxe is stable, I'm not too impressed with its OC options considering the contortions I had to go through to get it to OC (or even recognize when I changed the voltage options!)
Thanks for sharing and keep it up. I'm sure you'll find the key to getting her stable and blistering fast (the key is probably the promie or ln2 hehe)Quote:
Originally Posted by redpriest
that said, I might consider different memory if I were you. Only the best for Conroe ;)
What would you recommend? I'm using OCZ Platinum PC6400 with 5-5-5-15 timings. Fry's had some $300 PC8000(!) OCZ memory, but I wasn't going to splurge the extra cash for it.
EDIT: I haven't bothered yet with aggressively tuning my memory yet - I noticed moving from 667-> 800 hardly made a difference, but then again, I didn't change the latency either; I haven't pushed the memory to the max yet, though the BadAxe needs me to send 2.2v into it to get it to stably work at 5-5-5-15 (!).
Im ordering some Team memory tonight.. People have nail 1000 speed with it with about 2.4 or so volts.. Im also buying some G.Skill... I will see how each plays on my board. I dont have much faith in my P5WDH after reading some of these threads. Im picking up a cheap celly to power it up when the ram comes.
off topic, but has anyone been able to get MBM5 to read the 775x? if so, what profile or .ini file?
No, that's wrong.Quote:
Originally Posted by kiwi
Latest Everest build shows exactly the same temperature as Core Temp on each core.
Bios gives a temperature between idle and load because it does not idle the CPU completely.
Everything is just fine. With air-cooling I also get temperatures around 40-50°C idle and do not complain about wrong readings. If I switch back to water, temperature is alright again.
She already is pretty fast at 3.466 ghz =)Quote:
Originally Posted by dr_sharp
My wall right now appears to be at 3.733 - I think if I had water cooling, I'd be able to go higher with NP. Once I transition my prometia over, I think I'll see > 4 ghz.