X2 3800+ Steppings

[gord27]

bta=BigToe mod (tony aka "BigToe"), but that was leftover from 704-2bta in sig, forgot to delete it lol. Yes actually using 04/06 official now.

 

TCaseMax= 65C (higher TCasemax means more heat is tolerated, not more produced)

TDP= 89w (measured 71.9w by A64MaxTemp), all X2 are except 65w and 35wEE

Thermal Offset= 2C (this is why the big drop in temps, poor core contact with ihs)

 

my temps @3000mhz, 1.550v are roughly:

cpu/pwmic/chipset

idle

28/34/32

load

40/48/40

 

take off ~5C for each when it load [email protected]

 

before i put an 80mm fan @12v on the mosfets, PWMIC temps under load were almost 65C:eek:

 

fwiw, Toledo should run quite a bit hotter than Manchester due to extra cache (even though it doesn't work), seems to be the case from the few 3800+ toledo I've seen so far. I was getting load of 50-52C before popping ihs.

 

actually i made a typo. my tdp is 89w in core temp and tcasemax is 71c like i said before. actually, toledo would have a greater surface area with the added cache and therefore should be cooler no?

 

hmm, maybe a fan on the mosfets would drop my cpu temp as well...

 

btw, no way you idled at 23c at 2.8 if your ambient is basically that.

 

i popped my ihs and temps barely changed. given that i have a higher tdp though i guess it would make sense that mine runs hotter.

[tweakfreak88]

sitting at 2.95Ghz @1.55v... in the bios, 38-42C cpu, 55-60C pwmic, 40-45C chipset under load. I'm not sure whats keeping me from hitting 3ghz @1.55v but it starts acting flakey and seems to really want the voltage for another 50mhz

[hawkeyefan]

Those things did cost an arm and a leg when I bought my "simple" 3800 I did consider the idea for a while, consulted the bank account and dismissed the idea rather quickly

 

mine was $150 patience is a virtue (hehe, first time I ever waited till the "right time" to buy something)

 

actually I couldn't justify the purchase until recently, so it was lucky timing lol

[hawkeyefan]

actually i made a typo. my tdp is 89w in core temp and tcasemax is 71c like i said before. actually, toledo would have a greater surface area with the added cache and therefore should be cooler no?

 

hmm, maybe a fan on the mosfets would drop my cpu temp as well...

 

btw, no way you idled at 23c at 2.8 if your ambient is basically that.

 

i popped my ihs and temps barely changed. given that i have a higher tdp though i guess it would make sense that mine runs hotter.

 

1) I also have an 89w tdp X2, I don't understand what you're getting at. your TCaseMax being higher than mine should mean that your CPU can handle more heat, not that it creates more. Bigger cache (used or not)= bigger core= more heat.

 

2) yes, a fan on the PWMIC area would drop temps there. same as CPU, ram, anything else. even an 80mm@6v would make a huge difference as that mosfets typically don't get good circulation.

 

3) Mine isn't the first A64 that idles right around room temp, though the "-5C" rule in my post doesn't accurately apply to idle, but to load in that situation. Idle more like -3C, so 25-26C idle. but it never ever idles any longer than it takes for a screenshot, so it's just not something I really pay attention to. Full load 24/7 = what's idle?

 

4) did you use AS5 and not ceramique? If so, did you use "GPU method" or "A64 method", because A64 method does not work on naked cores.

 

5) again, you don't have a higher TDP, same as mine exactly, and TDP doesn't dictate heat output anyway. My 67w Venice for sure doesn't run as hot as 65w tdp X2s. My cooling setup appears to simply be more effective than yours? Haven't seen many air setups with better temps than mine tbh, and a few H20 that are actually worse as far as actual temps, though even at higher temps H20 is more effective obviously. Seen a whole slew of air cooled rigs with much worse temps than mine (and yours)

[gord27]

1) I also have an 89w tdp X2, I don't understand what you're getting at. your TCaseMax being higher than mine should mean that your CPU can handle more heat, not that it creates more. Bigger cache (used or not)= bigger core= more heat.

 

4) did you use AS5 and not ceramique? If so, did you use "GPU method" or "A64 method", because A64 method does not work on naked cores.

 

i wouldn't expect a bigger core to create more heat. what i was getting at is with a bigger core you have more surface area to contact the hsf and dissipate heat better.

 

i use as5. i spread it evenly across the core, same as with a gpu or an old axp. is that not right?

[hawkeyefan]

i wouldn't expect a bigger core to create more heat. what i was getting at is with a bigger core you have more surface area to contact the hsf and dissipate heat better.

 

i use as5. i spread it evenly across the core, same as with a gpu or an old axp. is that not right?

 

 

that's exactly how you should use the as5, so that is likely not the problem

 

maybe because more core is in contact with the ihs, the ihs itself becomes less effective at dissipating heat? dunno:sad:

[tweakfreak88]

Well im setting everything up in bios exactly like you have hawkeyfan. We have the same cooler, same exact stepping cpu. I am working on my chassis cooling system. I've been doing research on thermal dynmaics on closed air systems. It seems that a negative pressure case helps dissapate heat faster. The objective a 2:1 ratio but thats not really practically. Hopefully I can get like .7-.8 atm(1atm is atmospheric pressure at sea level). I have to seal off a lot of the lil holes on the chassis. With two 120mm exhaust and one 120mm intake I should hopefully achieve that, lower my temps and more o/c with minimal noise. I will post my results here tonight hopefully.

[tweakfreak88]

I ended up setting everything exactly like you have hawkeyfan, I even flashed to the same BIOS, the only thing we have different would be the manufactor for the memory, although we both have UCCC memory chips. My temps are around 29C-30C idle tops, as soon as it went under load it failed in sp2004.

[DFI N LUV&#39N IT]

Hi guys Im new to the forms and also a noob to comps..well I am building a comp and wanted to know about my cpu steppings..I have the Windsor model with the.. LDB4F 0622WPAW..Is this a good CPU for over clocking..I could only get this cpu to 2.2 but im not sure if it was because my RMAed motherboard..so anyways thanks for you help.

[Guest LithoTech]

LCBIE 0619RPMW still at stock right now, haven't had a chance to play with it most of my cooling (air) is still on the bench making modifications.

 

I noticed a 10 degree difference between the two cores (at ilde) in every software that reads them, A64info Beta will do for this, but exactly matches Everest etc:

 

 

As you can see, I'm fine if I go by MBM, but everything that senses both core temps shows me about 10 degrees spread between the two cores at idle.

 

Under load, the spread tightens up to about 5 degrees difference.

 

 

Not sure if this would be affecting the core temperatures, or if my TIM just hasn't set yet (only been 12 hours lol), but I noticed a few knicks in my XP-120 from the Opteron IHS, right where the corners would have touched. I'm thinking the IHS is high in these areas, but still scratching my head at how it could carve impressions into the nickle plate of the XP-120.

 

Here's one pic of the damage, I'll probably start a thread on this with more pics later:

 

 

Looks like I have some lapping to do. :/ I'm going to get a straightedge on that Opteron IHS too, see just how concave it is.

[Praz]

The temperature difference between the two on all dual core processors is common. My core 0 at idle usually read 3C - 4C below ambient.

[tcw]

Me for now it is 6hr prime stable but i will get about 14 or so tonight because i have class tomorrow morning.