This thread is so huge that it's hard to find something. I tried finding his bios settings and I found a bunch of stuff from him just not the bios settings. Also, I need 8GB on my system so 1GB sticks aren't really an option for me. 4x2GB is what I need...just trying to figure out if it'll solve my OC problem or not...and it's hard to find a decent deal on 2x2GB sticks with decent timings...G.SKILL on newegg has PC-8500 2x2GB for about $55 CDN but the timings are 6-6-6-18. 5-5-5-15 are there too but that'll cost me another $30 on top. Plus their shipping and handling charges are ludicrous for us Canadians at least. $30 just to get it shipped to Toronto for a package not even weighing half a pound.
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Asus P5E-VM HDMI: best matx ever?
- Thread starter pvhk
- Start date
This thread is so huge that it's hard to find something. I tried finding his bios settings and I found a bunch of stuff from him just not the bios settings. Also, I need 8GB on my system so 1GB sticks aren't really an option for me. 4x2GB is what I need...just trying to figure out if it'll solve my OC problem or not...and it's hard to find a decent deal on 2x2GB sticks with decent timings...G.SKILL on newegg has PC-8500 2x2GB for about $55 CDN but the timings are 6-6-6-18. 5-5-5-15 are there too but that'll cost me another $30 on top. Plus their shipping and handling charges are ludicrous for us Canadians at least. $30 just to get it shipped to Toronto for a package not even weighing half a pound.
If I was using 2GB modules on this board I would try these fiirst:
http://www.newegg.com/Product/Product.aspx?Item=N82E16820146726
With they way he has his HS mounted (parallel to the backplane) I'd guess he's using the SLI version of the HR-05. If you can mount the HS perpendicular to the backplane (P180 Mini, for example,) then you can use the non SLI version. Or, you could use the Noctua Noctua NH-U12P HS & NC-U6 NB combo.
Also, for those thinking about getting this board I just saw a post on [H]otDeals, it's $99.99 + free shipping right now.
http://www.newegg.com/Special/ShellShocker.aspx
http://www.newegg.com/Product/Product.aspx?Item=N82E16820146726
With they way he has his HS mounted (parallel to the backplane) I'd guess he's using the SLI version of the HR-05. If you can mount the HS perpendicular to the backplane (P180 Mini, for example,) then you can use the non SLI version. Or, you could use the Noctua Noctua NH-U12P HS & NC-U6 NB combo.
Also, for those thinking about getting this board I just saw a post on [H]otDeals, it's $99.99 + free shipping right now.
http://www.newegg.com/Special/ShellShocker.aspx
This hot deal is for the Americanos only.
I think it's actually the non-SLI version he's using if you compare the pics at:
http://www.thermalright.com/new_a_page/product_page/product_chipset_cooler.html
<< tags to finish your quotes.[/QUOTE]
Ya, you do need it...but I'm just using the Quote button so it does it automatically...unless I'm deleting it without noticing it. It works fine for everyone else, including this post. Ok nm.
Still, I've never seen a ratio of more than 9 on Q6600. Can anyone else confirm if there are versions of this cpu with a ratio of more than 9? I think he's got the same revision and all as I do so it should be the same.
Oh, you need to find an Americano compadre, willing to ship stuff to you.
You're right, you can see how it lines up in the post just before his in another setup (with more pics,) but also using the HR-05 non-SLI.
Perhaps he has an "Extreme" chip? Or got real lucky with an unlocked chip?
http://www.tomshardware.com/forum/241554-28-unlock-q6600-multiplier
Oh, you need to find an Americano compadre, willing to ship stuff to you.
You're right, you can see how it lines up in the post just before his in another setup (with more pics,) but also using the HR-05 non-SLI.
Perhaps he has an "Extreme" chip? Or got real lucky with an unlocked chip?
http://www.tomshardware.com/forum/241554-28-unlock-q6600-multiplier
Who knows in that case. He didn't prove it either using CPU-z. Missing or wrong info on it...then he reposts it again with the multiplier at 6x then at 9x so he proved nothing to me.
So there's an Extreme version of the Q6600 out there that comes with the multiplier above 9? I didn't know that.
Btw, my Q6600 is also unlocked...says so in the BIOS but it's still only 9x max (see my image on P64, post# 1267.
Hmm a few thoughts:
officermartinez is using the non-SLI version of the HR-05 IFX.
The 10.0 CPU ratio IS a mistake. As you said, 9.0 is the max on the Q6600.
Your last statement about your setup working with CPU ratio 8.0 1.30V, what settings did you say your RAM was at?
Also, the pencil mod to do is the one that officermartinez is showing. If you did another one, that could be dangerous!! I suggest you use an eraser and take it off.
officermartinez is using the non-SLI version of the HR-05 IFX.
The 10.0 CPU ratio IS a mistake. As you said, 9.0 is the max on the Q6600.
Your last statement about your setup working with CPU ratio 8.0 1.30V, what settings did you say your RAM was at?
Also, the pencil mod to do is the one that officermartinez is showing. If you did another one, that could be dangerous!! I suggest you use an eraser and take it off.
Oh, you need to find an Americano compadre, willing to ship stuff to you.
You're right, you can see how it lines up in the post just before his in another setup (with more pics,) but also using the HR-05 non-SLI.
Perhaps he has an "Extreme" chip? Or got real lucky with an unlocked chip?
http://www.tomshardware.com/forum/241554-28-unlock-q6600-multiplier
I believe I said 1.8V but now I'm not 100% sure anymore. When I get home I can try it again with IBT for 5 loops and see if it passes with vCPU 1.3 and vDRAM 1.8 unless you want me to try something else.
I figured the 10 cpu ratio must be a mistake.
As for the pencil mod. I'll definitely clean off the previous attempt with some isopropyl alcohol on a lint free cloth before I try it again. But I have to find a 2B pencil and that's easier said than done without going out to buy one somewhere. I don't think an eraser would get in very well there. Those transistors are VERY small.
After cleaning the old attempt, I'm just going to fill that transistor black again and blow off any excess lead from the pencil with one of those air compressed cans. In my 1st attempt I did the 2nd transistor on the right (in the circle) and that's what I'll do the 2nd time. I'm going to have to find a magnifying glass though so that I can tell this time that I filled it completely (though I know you don't have to do that).
Don't worry too much about the pencil mod. Ultimately it will be needed to achieve the results you want (if your hardware plays nice) but for the immediate it shouldn't matter too much. We just deal with the vdroop. My guess is 5-4-4-15 2.0V or 5-6-6-15 1.8V. Pick one
By the way I just checked page 59 of this thread and it's clearly the resistor to the right in the picture above for the pencil mod.
By the way I just checked page 59 of this thread and it's clearly the resistor to the right in the picture above for the pencil mod.
Don't worry too much about the pencil mod. Ultimately it will be needed to achieve the results you want (if your hardware plays nice) but for the immediate it shouldn't matter too much. We just deal with the vdroop. My guess is 5-4-4-15 2.0V or 5-6-6-15 1.8V. Pick one
By the way I just checked page 59 of this thread and it's clearly the resistor to the right in the picture above for the pencil mod.
Do you still want me to remove the pencil mod I tried the 1st time or leave it?
Well, I'm pretty sure I did one of those settings if not both and it passed at cpu ratio of 8 with both memory timings. So what's next? That way when I get home in about 4 hours I'll start working on your next suggestion. Or do you still want me to run IBT with one of those timings and ratio 8x to be sure it passes?
Let me explain what I'd like to do:
First I wanna make sure that everything is stable with whatever settings you have at 8*333, RAM running at 800MHz.
Second you change the CPU ratio to 9 and by trial and error you find the CPU voltage that'll keep you stable. So now we are 9*333 stable, memory at 800.
Third: drop ratio to 7, push bus to 400, RAM at 800, CPU and DIMM voltage unchanged. Now You need to stabilize this by adjusting FSB, NB, PLL voltages and maybe the references (Clue: you can try and use officermartinez's settings for those or maybe a notch or 2 lower)
Four: increase CPU Ratio. Either you go straight to 9, or you go to 8 first, and find the proper CPU voltage to keep it stable. My guess is that if you don't do the vdroop pencil mod you might be better off trying 8 first.
Once you are there, you should have 8*400, RAM at 800. Not too shabby. You just need to get the right voltage for a ratio of 9 and that's it, Bob's your uncle.
One caveat though: your CPU cooler could become a problem when you reach step 4. Then at this point you can make the decision to change it. To be honest I doubt you'll be able to do all of this in one day, even if things go according to plan.
Mod-wise, if you think you got the wrong one in the first place, do take it off. An eraser should do the trick, don't you think?
Let me clarify one point that I think you've misunderstood. I never said I did the wrong one. I was simply writing down how to pencil mod, so that in case I'm wrong someone can correct me. I know it was the RIGHT one all along and that is the one I did the first time. When I said I plan to do it again I mean do the same resistor a 2nd time but this time with a different pencil, in this case grade 2B.
I don't know about the eraser but if I can do it I will. I just remember that when I tried running over the top of that resistor with the pencil that there was hardly any pencil movement possible (back and forth) because it's so tiny. Maybe I really need to sharpen the pencil to the max and make it almost pinpoint in order to be able to fill the top of it with lead.
Anyway, I hear what you're saying and I'll try to work from that. You didn't say what to stick with so I'm not sure if I should keep using IBT or Orthos or Prime95. Obviously, anything other than IBT will take much longer to find out whether it's stable or not. On the other hand, IBT is too stressful in my opinion and I'm not sure that a system would undergo such stress even under 24/7 loads which I will be doing. Like I said, those temps. are extremely high, and no other program that I plan on using will raise the temps that high during a load.
I'll work on those steps when I get home then.
You are right: I had misunderstood. I do seem to remember pvhk saying somewhere that the pencil grade was important. So yeah give it a go. Not sure there is a need to remove the other pencil mod.
I'd say for step 1 2 3 stick with IBT. Step 4 start with IBT but if you see that your temps get out of control go back to ORTHOS (or Prime95). You will probably know by then if your cooler is good enough.
Roger that.
Actually I just thought about the choice of stability testing program: IBT not failing doesn't mean that your system is stable: it just means that it hasn't failed yet. So if you have a positive result for IBT (no failure after a certain number of loops), then you should have a longer run of either IBT or Orthos (for Orthos: 8-10 hours to be on the safe side, but 4-5 might be enough??).
Have you used IBT yet? It's extremely stressful, hence the reason why the temps go crazy high! I feel that if Orthos needs 2 hours to crash IBT will do it within ten minutes.
I think what I'll do is run Orthos or Prime95 for a very long time (1 or 2 days) once my goal has been achieved...3.4GHz at an acceptable temperature. I'd like to do that with the hardware I have now...and if I need to upgrade the cooler and/or memory in order to get at least 3.6GHz then so be it. AT this point however, I'm quite skeptical but as I said, I'm not giving up yet.
Ok, shutting down and I'll try those steps you mentioned. What time you going to bed? EST.
I did the vdroop mod a 2nd time, this time with a 2B pencil. Can someone confirm when it worked or when it hasn't?
I set my vCPU to 1.3 and cpu-z reports 1.272 with no load. Once I run IBT it goes down to 1.264. Is that still acceptable? Or did the vdroop mod not work? I expected them a little closer, but this is still only a difference of 0.036 when compared to BIOS and CPU-z under full load.
I have a feeling that it's working because I've noticed a big change. Temps are now showing 87C on the cores in IBT with vCPU @ 1.300 whereas before my temps were probably around 73C at the same voltage.
I set my vCPU to 1.3 and cpu-z reports 1.272 with no load. Once I run IBT it goes down to 1.264. Is that still acceptable? Or did the vdroop mod not work? I expected them a little closer, but this is still only a difference of 0.036 when compared to BIOS and CPU-z under full load.
I have a feeling that it's working because I've noticed a big change. Temps are now showing 87C on the cores in IBT with vCPU @ 1.300 whereas before my temps were probably around 73C at the same voltage.
I'd say it worked. What you set in BIOS as vCPU does not look anything like what the BIOS reports under the Hardware tab, nor what CPU-Z shows. Since you are seeing .008 vdroop, that's great. I haven't done the mod and I see .06 vdroop running IBT.
Wrong. That's 0.036 drop, not 0.008. So I don't know if it worked. And I don't know if we're supposed to compare CPU-z's voltage when under full load or idle.
One thing though. Before doing the pencil mod I set vCPU @ 1.3 and vDRAM @ 1.8 with CPU Ratio 9, 333FSB and 5-4-4-15 timings and after running IBT for a few minutes I got the BSOD. Now, after the vdroop mod, the same settings seem to be stable.
1.27200 - 1.26400 = 0.008
That's only one step difference in what CPU-Z reports, i.e., the minimum.
Have you tried reseating your cooler with new thermal goop? I looked at a review or two on NewEgg and although they weren't OC'ing as high as you, their temps were way cooler.
Also, looking back at your SpeedFan screenshot, I noticed that "CPU" temp is lower than your core temps. I have a feeling that "CPU" is really the reported value of the NB temp, and in any case I've never seen that value lower than my core temps. Maybe this is a typical difference between Quads and C2D's, but my hunch is no. And combined with the NewEgg reviews it's a double no.
That's only one step difference in what CPU-Z reports, i.e., the minimum.
Have you tried reseating your cooler with new thermal goop? I looked at a review or two on NewEgg and although they weren't OC'ing as high as you, their temps were way cooler.
Also, looking back at your SpeedFan screenshot, I noticed that "CPU" temp is lower than your core temps. I have a feeling that "CPU" is really the reported value of the NB temp, and in any case I've never seen that value lower than my core temps. Maybe this is a typical difference between Quads and C2D's, but my hunch is no. And combined with the NewEgg reviews it's a double no.
Oh. I thought you were supposed to compare the vCPU set in BIOS to what CPU-z reports under load. You're saying it's what cpu-z reports with no load and compare it to a full load. In that case yes, there's barely a difference. And this time I'm running IBT with 8GB of ram whereas before I always ran it with 4GB.
Yes, the temps are high...higher than before the pencil mod. Is this normal?
I wonder if I was to go lower on the vCPU would the system still be stable but report lower temps? Oh and yes, I've just replaced the goop 2 times but after reading a review of how it should be layed out on the cpu I'll do it one more time soon.
Right, you've got vdroop under control.
Cool beans.
Probably, yes. Because you are running at a higher effective voltage (for a given setting in BIOS) now that there's no vdroop.
It would definitely be cooler, but stability is the question. That's the whole idea, reduce voltages until you find the line between stable and unstable.
I still think your cooler is not performing well. Read the first pages of NewEgg reviews. Also, like I wrote I think your NB temps are less than your core temps and that points to a problem with CPU cooling IMO.
I appreciate all your input. Maybe you're right about the cooler. But I don't understand why the temps are so high. Even now, I dropped the vCPU from 1.3 to 1.2875 (one notch in BIOS) and temps are still high, reaching a max. of 81C on the cores and 59C on the CPU (or NB if that's what it really is).
I'll keep lowering the vCPU until my system crashes. I'll then raise it to the lowest vCPU where the system is still stable. Then I'll remove the cooler and reapply the goop with some OCZ Arctic Freeze since that's all the pc store had (out of AS5). According to some reviews, the AS5 and Arctic Freeze are about the same.
Well, I went down to 4GB and vCPU @ 1.2875 and IBT passed. I'll drop it another notch and report back soon.
I don't think I would run it up that high any more.
http://processorfinder.intel.com/List.aspx?ParentRadio=All&ProcFam=0&SearchKey=q6600
I would pull the cooler and reapply the goop. Here's some good instructions for AS5 that I'm sure apply to the OCZ stuff as well.
http://www.arcticsilver.com/pdf/appinstruct/as5/ins_as5_intel_quad_wcap.pdf
Be sure to clean up the cooler and CPU with isopropyl alcohol and a lint free cloth. I recently found a cheaper replacement on Amazon for the cloths but I used to use the eyeglass cloths you get at the drugstore for like $3/ea. Don't use the wetwipes.
(Or a paper towel.) You want to avoid finger prints and the like coming into contact with the mating surfaces, let alone micro lint and threads.If re-seating the HSF doesn't make a pretty drastic difference in temps I'd say either the mounting is off so it's not making proper contact or the fan isn't spinning enough RPMs.
Battlestar
Weaksauce
- Joined
- Feb 19, 2008
- Messages
- 71
Hi Bing,
Last night, I have confirmed that my mobo is 1.00G version but the pencil mod work perfectly
As you have seetings above, were almost the same except for the DRAM Freq, CPU Voltage and PLL Voltage. My PLLv can run 1.54v. As I have said few pages ago, my mobo can run 500x8 settings but failed on OCCT after 1hr & 10 mins. I tried to lower the FSB and into 445x9 and now stable in OCCT for 24hours. Here's the screenies
OCCT with Real|Core Temp|CPU-Z
CPU-Z Validation
Let's see if those settings is also stable on Linpack and Prime95. My next objective now is to run these stable on 500X8. If any of you guys have run E8400 E0 on 500x8 please share your BIOS please?
Thanks - Battlestar
MassiveOverkill
[H]ard|Gawd
- Joined
- Jul 8, 2005
- Messages
- 1,748
Anyone try Windows 7 yet on this board? I can't get the Atheros network drivers to load, (yes I've tried compatibility mode).
I don't think I would run it up that high any more.
http://processorfinder.intel.com/List.aspx?ParentRadio=All&ProcFam=0&SearchKey=q6600
I would pull the cooler and reapply the goop. Here's some good instructions for AS5 that I'm sure apply to the OCZ stuff as well.
http://www.arcticsilver.com/pdf/appinstruct/as5/ins_as5_intel_quad_wcap.pdf
Be sure to clean up the cooler and CPU with isopropyl alcohol and a lint free cloth. I recently found a cheaper replacement on Amazon for the cloths but I used to use the eyeglass cloths you get at the drugstore for like $3/ea. Don't use the wetwipes.
If re-seating the HSF doesn't make a pretty drastic difference in temps I'd say either the mounting is off so it's not making proper contact or the fan isn't spinning enough RPMs.
I don't think I would run it up that high any more.
http://processorfinder.intel.com/List.aspx?ParentRadio=All&ProcFam=0&SearchKey=q6600
I would pull the cooler and reapply the goop. Here's some good instructions for AS5 that I'm sure apply to the OCZ stuff as well.
http://www.arcticsilver.com/pdf/appinstruct/as5/ins_as5_intel_quad_wcap.pdf
Be sure to clean up the cooler and CPU with isopropyl alcohol and a lint free cloth. I recently found a cheaper replacement on Amazon for the cloths but I used to use the eyeglass cloths you get at the drugstore for like $3/ea. Don't use the wetwipes.
If re-seating the HSF doesn't make a pretty drastic difference in temps I'd say either the mounting is off so it's not making proper contact or the fan isn't spinning enough RPMs.
Ok, I'll be reapplying the heatsink very soon.
Well, it seems the lowest I can go on the vCPU is 1.2275V and vDRAM 1.80V. CPU-z idles at 1.224V. The pencil mod definitely made the difference as I couldn't go below 1.30V or 1.35V before. DRAM is at 5-4-4-15 and 800MHz (as per its rating) and FSB is 333 for a total of 3GHz.
Proceeding to step#3 as per NicholT's instructions. Though I guess I should reapply the HS first.
Goop has been reapplied yet again. No difference.
Be sure to clean up the cooler and CPU with isopropyl alcohol and a lint free cloth. I recently found a cheaper replacement on Amazon for the cloths but I used to use the eyeglass cloths you get at the drugstore for like $3/ea. Don't use the wetwipes.
Before reapplying goop I cleaned cpu and cooler using Isopropyl alcohol and a lint free cloth (I've got a whole roll of it).
Hmm..RPM's seem fine. Last time I checked it was spinning as per its rating. If what SpeedFan labels NB as CPU then are you sure it's supposed to be lower than the cores temps (or about the same)? I suppose it makes sense but then why such a big difference between my NB temp. and the cores' temp?
I mean, if I was to replace the HS with Xigmatek HDT-S1283 for example, I don't see it reducing it by THAT much...we're talking nearly 20C difference between the cores and the NB.
"Pros: It is cooling my Intel Q6600 at ~46 Deg Celcius when overclocked to 2.88Ghz. It will top out at ~51 Deg Celcius when overclocked to 3Ghz."
Is this the review comment you were referring to at Newegg.com regarding my cooler?
Maybe he was using SpeedFan like me and was reading the temp. for what SpeedFan reports as CPU but is actually NB? Because if that's the case then I'm getting the same temperature. And he probably wasn't using IBT, and IBT will raise your temp. by up to 20C over other loads (they say so on their website).
I'm now using another program as a load and its been running for 20 mins now, topping out at 67C for the cores (50 for NB)...which is still high but at least it's not 82C like I get with IBT. And this is @ 3.2GHz (400*8).
Pencil mod worked great with the proper pencil. It was worth the trip to Wal-Mart ;-)
I'm running @ 3.2GHz now (400*8) and it seems stable but will run it overnight to be sure.
I haven't tried *9 yet as I know the temps will be overkill. I need to resolve my issue about why my NB is 50C at full load and the cores go up to 82C (with IBT). As per 'bing', he says the cores shouldn't be that much higher than the NB.
Not too shabby indeed. All at 1.312V vCPU and 2.1V vDRAM with 5-4-4-15 timings. I raised the PLL a little along with NB but I left FSB as per what you originally suggested. I also set Clock Overcharge to 0.80V.
As you can see by the time of this msg., I did this in one night (will pay for it in the morning when I won't be able to get my ass out of bed for work) and I don't regret it so far.
I'd never get this far though without the vdroop working properly...and of course your help. Bing is trying to help too so that's great...but I've stumped on why the cores are so much more hot than the NB and I doubt the xigmatek will make all the difference....it's too much of a difference in degress. We're talking around 20C!
I looked at your voltage numbers: they're quite good. I really need to do it too. I've just received my HR-01 PLUS so I'm gonna start pushing my CPU again.
Hmm I don't think the NB temp is actually reported by the Winbond chip on this board. On HWMonitor, you only have 2 temps: SYSTIN which is a 'board temp', which is indicative of how well the hot air is being taken out of your case. CPUTIN which is (as the name would suggest) a CPU temp that, I think, is taken at the CPU heatspreader (the area of contact between the CPU and the cooler IIRC). But I'm not 100% sure about all this.
I don't really know myself what Clock Overcharge does. But if it works, who cares what it means... also I see that you pushed your vDRAM to 2.1V. That's good. I was thinking to myself that according to your RAM specs 1.8V would never give you stability. Maybe one of the things to do at one point is to try lowering that voltage a couple of notches to see if the system is still stable. But make sure that everything's stable in the first place. Actually that goes for most of the voltages set. The lower the better.
Hmm I've just run IBT on my system and the max Cores' temp and CPUTIN value were the same (+- 1 or 2 degrees). Is it an indication that your cooler is struggling to evacuate all the heat? I'd like to get some input from someone who actually owns a Q6600 and with a decent O/C.
Is your PCI express frequency too high? Needs to be 100mhz.
It's 100 on mine so that's not it.
Are you getting the same temps as me even with this new cooler? Goes without saying then that getting a better cooler will not help me much.
Hmm...I don't know. I guess bing wasn't 100% sure either. Your explanation makes sense too though. What OC do you have? 3.4GHz? Or 3.2 like me? If it's the latter and you have the same temps as me then that's an indication that perhaps it's not my cpu after all. Either we're both doing something wrong or this is normal as per the cooling we have.
Yeah, I was following the steps you outlined and I was doing everything in increments so I already know I'm running at the absolute lowest voltage, at least on the vCPU. I also removed the additional 4 GB of ram to leave me with 4 because with 8GB IBT took much longer to finish. So last I remember is my system is running stable at 1.312V vCPU and vDRAM at 2.1V...I could probably go less on the ram because when I upped the ratio from 7 to 8 I upped everything you said at the same time (vCPU, vDRAM, vPLL), with vDRAM from 1.8 to 1.9 and then next to 2.1 rather than 2V. I'm not sure if that's the way to do it but it was getting late and I was almost finished.
I put back the other 4GB in my system (8GB total) and with an overnight run with a full load I was relieved to wake up this morning and not find my system has rebooted. I did not increase the voltage on the DRAM, I left it at 2.1V.
Other than my temps on the cores, everything you said worked like a charm after my 2nd pencil mod and I'm quite happy even though I didn't achieve 3.4GHz. However, I see lots of room to get to 3.4GHz and beyond as I've been able to get to 3.2GHz at only 1.312V on the vCPU whereas before I needed like 1.4875V. And all this time, I never even knew that Bob's my uncle
But I still need to figure out my temp. problem though and it looks like it'll be tough to do.This forum is great, especially this thread for those that have the same board. It's the best forum I've ever joined! With users that help out others like you, bing and the rest on here, it's no wonder!
Makes two of us now.
I don't have a Q6600. I own a C0 E8200 that I'm running 24/7 at 3.6GHz. I've pushed it to 3.8 once, seemed stable but when running ORTHOS my temps were getting too high for my taste (that was with my Ultima 90). Now with the HR-01 PLUS, I've taken 8 degrees or so off my IBT temps, so that'll give me enough room to up my voltages again. But my temps are not comparable with yours. Different CPU, different frequency. To give you an idea, IBT will push my core temps and CPUTIN to 52 degrees Celsius.
When you just change your CPU multiplier, You're only impacting the CPU so vDIMM shouldn't have to change when you go from 7 to 8. I'm not sure about vPLL, but I operate on the assumption that you shouldn't have to change it either. You can then separate the whole thing into 3 different sections: Mem, CPU and 'The Rest' (essentially the NB). So far we've always kept your memory at rated frequencies so we've never had to worry about memory parameters. When changing the BUS frequency, lowering the CPU ratio prevents you from raising the CPU frequency (you then don't have to worry about CPU voltage). Then you just have to fix 'The Rest'. When that's done, you raise CPU ratio and just have to fix the CPU voltage.
When you just change your CPU multiplier, You're only impacting the CPU so vDIMM shouldn't have to change when you go from 7 to 8. I'm not sure about vPLL, but I operate on the assumption that you shouldn't have to change it either. You can then separate the whole thing into 3 different sections: Mem, CPU and 'The Rest' (essentially the NB). So far we've always kept your memory at rated frequencies so we've never had to worry about memory parameters. When changing the BUS frequency, lowering the CPU ratio prevents you from raising the CPU frequency (you then don't have to worry about CPU voltage). Then you just have to fix 'The Rest'. When that's done, you raise CPU ratio and just have to fix the CPU voltage.