No, this should be universal to all AMD64 processors. Same voltage dividers, same electrical connections, same diodes.Quote:
Originally Posted by Hound
Printable View
No, this should be universal to all AMD64 processors. Same voltage dividers, same electrical connections, same diodes.Quote:
Originally Posted by Hound
different mem controller in rev e however..it could be more voltage sensitive.
yeh, i thought so, as RevE is improved over the previous generation..
you can change the memory controller all you want, but it's the diodes that protect it from the VTT, which shouldn't have to be tweaked when the MC is ;)Quote:
Originally Posted by Otaking71
Small question based on this list....
Vdimm - Vcpu(min)
----------------------------
2.9 V --- 1.12 V (min)
3.0 V --- 1.17 V (min)
3.1 V --- 1.22 V (min)
3.2 V --- 1.28 V (min)
3.3 V --- 1.33 V (min)
3.4 V --- 1.37 V (min)
3.5 V --- 1.42 V (min)
3.6 V --- 1.47 V (min)
3.7 V --- 1.53 V (min)
3.8 V --- 1.58 V (min)
3.9 V --- 1.62 V (min)
4.0 V --- 1.67 V (min)
Is the minimum required vcore the vcore you set in your bios or the sum of the vcore + vidspecial = the vcore shown in CPUZ...??
NC™
You are confused.Quote:
Originally Posted by NightCrawler™
The sensors at least on the DFI boards are just off, don't use them.
You get what you set in the BIOS, and it doesn't matter whether you set it in the primary control with no special or whether you mix something together with the special.
So regardless my combination I set in my BIOS the final output is what is shown in CPU-Z.
So my 1.425vCore should be safe with my 4000 VX Gold @ 3.3v... correct? :)
NC™
No, what CPUz shows is too low, at least on DFI NF4 boards.Quote:
Originally Posted by NightCrawler™
What you get is what you set in the BIOS.
I think you are still safe, as 3.5v / 2 = 1.75, 1.75 - 1.42 = 0.33, which is less than 0.6volts.Quote:
I was using 1.42Vcore with 3.5Vdimm benching @ 3GHz then remembered this thread and upped Vcore to 1.44v, is this safe ??, this as per my sig using my new 146 that is simply an amazing chip
If it died I would cry for a week
OT: I just got my opty 144 and its not very good... 2.7ghz @ 1.55vcore in bios. Read as 1.52vcore in ITE smart guardian, I dont have any multimeter to test with. Is this vcore thing an issue with the board? or is my Enermax PSU not giving enough juice?
Ah... now I get it... LOL Thanks for the heads up Martin... :toast:Quote:
Originally Posted by Martin Cracauer
NC™
what about 2.6-2.7 on vdimm :D
i have my X2 3800+ @2400 1.45V :D
i think it's safe...
hmm... sounds interesting. This might uhh explain why my computer totally crapped out when I was using a ocz ddr booster on my asrock 939dual sataII mobo (EVEN THOUGH I GOT TCCD) n i know they don't need a lotta vdimm, coughs but i tried anyway. And let it sit at 3.2v for a while while vcore was 1.4 and like everything got really really messed up in BIOS. the whole screen was like full of artifacts. And then i turned it off, reset the cmos, and the turned on the power, and my fans and hdd's n stuff started up for about a second and just shut down. I tried pressing the power button multiple times. but it didn't do squat. I was freaking out. So i took out the booster and and prayed nothing bad happened. And so far it seems ay okay. NEVER GONNA USE A BOOSTER EVER EVER AGAIN! (i mean it was my fault for using such high vdimm, but its ocz's fault that the voltage isn't stable, it keeps fluctuating all over the place from 3.2 to 2.4 to 2.8 to 4.2. Its ridiculous. I'm just glad that nothings broken. YET!
Hmm, I had preached that high vDIMM killed CPUs after a # of processors were dying on the old BH-5 chips back in the 130nm days...I finally got some BH-UTT chips and Ozzimark told me about this as well as CCU, but I hadn't really seen the reasoning. I now understand a little better and I'll feel a little safer now when I go to rape this G.Skills. :D
I agree with your logic Micutzu, but where did you learn the assumption that "the CPU doesn't have to do directly with the VDD, but with Vtt/Vref, wich should vary from 1.25 to 2V when you adjust the RAM from 2.5 to 4V"?Quote:
Originally Posted by Micutzu
Particularly I ask because it doesnt sound so secure when you say "... which should vary ..."
Cheers! :)
not all boards supply the proper VTT
also, if we go based on the current (which should depend mostly on the square of the voltage), and 2.8v and 1.1v being the max differential, things look a bit different
but who knows ;)Code:vdimm vcore
2.5 0.901
2.6 0.970
2.7 1.036
2.8 1.100
2.9 1.163
3.0 1.225
3.1 1.285
3.2 1.345
3.3 1.404
3.4 1.463
3.5 1.521
3.6 1.578
3.7 1.635
3.8 1.691
3.9 1.747
4.0 1.803
Vtt/Vref must always be half of Vdd, so at Vdd=2.5V Vtt/Vref=1.25V, Vdd=3V Vtt/Vref=1.5V and so on, at Vdd=4V Vtt/Vref=2V. Maybe that wasn't the best choice of words, replace "should" with "will" :).Quote:
Originally Posted by beta67
Some photos are here for general usage.Quote:
Originally Posted by Fr3ak
For more info download Guidemania.
I always bump Vcore too much and destroyed the IMC of my previous Venice.
Of course it was normal. :D
so no issues with my winchester a 0.97v and gskill@2.7?
Hey guys sorry to bump a dead thread but I just realised something.
The A64 OC guides need to all be changed so that people dont lower cpu multi and leave stock voltage while testing memory!
Thats probably one of the longest periods that this voltage difference occurs.
This is questionable, I OCed my Sempron 64 2500+ E6 cpu with 1.4vcore and a 3.3v DIMM voltage in my hyper-X's and didn't do anything to my cpu. Hmm so now I got a venice DH-E6 but keep it at a super high voltage at 1.725v. Never had a prob like this with ever burning out a cpu. All or most CPU's can take about as much as a DFI can give it 2.1v. Some are crazy enough to run them that high. I'll run my TCC5 DDR600 OCZ's at 3.3v too. Only OCZ's arn't cheap. =P (seen them like upto 3.6v.) They can and love the volts like my cheap hyper x kingstons. So I guess I don't have to worry since I got my cpu running at high vcore with my DIMM.
Uh, this thread says this should be fine, so no suprise here.Quote:
Originally Posted by Serge84
hello guys,
as you can read my configuration in my signature, i would like to have some advise from you about to set up my voltages on cpu and ram
thank you very much.
i like c'n'q+oc :)
i hope my system will be safe:
opteron 170 toledo watercooled
4x512 corsair tccd
dfi expert
cnq enabled
260x10 2.5-3-3-6 2T
idle 260x5 1.1vcore 2.9vdim
load 260x10 1.45vcore 2.9vdim
is it ok?
To all that ask if their config is ok: go and do the maths guys:
Vdimm/2 - 0.325v = Min Safe Vcore
That's the most conservative formula on the full thread (well, the most is the one that implies the square of the voltage that ozzimark said, so simply check that post: http://www.xtremesystems.org/forums/...&postcount=115)
On a side note, and continuing with the OT: what happens with dual cores? I'm asking this because they refuse to oc with much volts (mine doesnt do 2650 with 1.45 but does with 1.375v).
Only the difference between VTT and Vcore counts? I think I read (sorry, I read the thread from beggining to reply and can't remember where) that bumping the volts help heating the cpu thus shielding it. Does that mean that heat also helps (advantage for DCs), or I have totally misunderstood (that point only :D)
1.) I still dont understand why is memory controller of E-stepping can be damaged using high VDIMM. Why are another, non built in, memory controller going not to be damaged?
2.) Where are protection diodes located? 100% not on MoBo (I checked DFI Lanparty)
i recently killed my good ol' 3700+ because i setted 1,42V 41C in bios and changed tccd 2 oldschool bh-5 and set to 3,6V 250 DDR500 2-2-2 then safe cmos and also into the bios 260 3,8V 2-2-2 cmos save 3 secs booting then "pzzzt" cpu burned away still then i didnt use the 4V jumper anymoa but good to know it now!
thx tony (still 2 late)
seEn
So yours burned at 3.8 Vdimm and default Vcore. What was default Vcore on that 3700+. Probably 1.40 but might vary.Quote:
Originally Posted by andL64
yes it was 1,4 bios but i set +104% so i got 1,42V in bios 41C and 3,8V vdimm => it burned...
That reinforces the theory of vdimm/2 - 0.325v = min safe vcore.
3,8/2 = 1,9 - 1.4 + 104% = 0.444v -> Out of range
Sorry for your lose andL64 :( I see you've found a good replacement with that FX-57 :)
that logics does not work.
Min-vCore with CnQ = 1.1v, max JEDEC specced vDimm = 2.85 gives a vRef of 1.425 which then gievs a max diff of 1.425 - 1.1 = 0.325v.
BUT
Would you really think AMD wouldn't have any kinds of safety-bounds?
I runt my vCore is 1.025 and my vDimm is 2.7v which would be a vRef of 1.35 which gives a diff of 0.325 ;)
If you read through the thread, 0.325v is exactly the safety differential between vcore and 1/2 vdimm. And yes, without re-reading all the thread, IIRC someone arrived to that conclusion the same way as you, CnQ vcore versus JEDEC 2.85v gives you 0.325v.
With all this said, what do you meant with "that logics does not work"? You are clearly stating that it works with your 2 next statements ;)
So AMD safety boundis that, 0.325v with max JEDEC specs :)
Sorry for posting this message here..
i have on cpu - +0.3v (mobo epox 9npa+u). but in real it is only +0.23-0.3v. Is it PSU problem, or problem in mobo power system??? thanks.
Well Ladies and Gentleman,
it looks like, ironically, I am the first victim of this.
I got a bunch new BH-5 modules and tested various voltages, including 4.0. Although I am aware of the voltage issue discussed here and hence was careful to always have 1.55+ Vcore, my 3.0+ GHz CABNE is now fried.
It looks like I either forgot to set Vcore when setting Vdimm at some point after a CMOS reload, or maybe that the board applied Vdimm after Vcore during boot after fresh settings.
In any case, said CABNE now still kind of works but driving RAM fast doesn't work at all, and even with RAM at stock speed I get the most violent memory corruption I have ever seen. The filesystem on my clocking/testing installation was scrambled bejond imagination, even wiping out the boot code. This applies to all RAM, tried BH-5, TCCD and UCCC. The CPU's memory controller is fried. CPU still boots, though.
A re-test of the board with a different CPU and some of the same RAM is happy as normal.
I just got my new DFI nf4 Ultra-D board from new egg & it is revision (R.AD0), the one HW changes. Does anyone know if this Vdimm/Vcore issue was addressed with this revision, or is it possible to fry my CPU if not carefull... thanks!
It is not possible to address this particular issue from a board level.Quote:
Originally Posted by OCme
Thanks uOpt!
Quote:
Originally Posted by uOpt
Cancel that report, it was just bad PSU after all. And my CABNE works again after a few days of rest and rebuilding everything with my Powerstream.
I fact, it kicks very well :D
http://www.xtremesystems.org/forums/...&postcount=319
PSU RMAed...
ok, so....
vdimm/2 - 0.325v = min safe vcore
vdimm = 2.65 (actually 2.6 but smart guardian sees it at 2.65)
vcore = 1.648 (1.475 + 113%)
this gives me.. 1v
am i safe running at these voltages? my UCCC chips run pretty low, but my CAB2E needs 1.6+ to run 3 GHz.
i didnt even know about the relationship between vdimm and vcore until now.
actually, one more thing. is it better to run
1. higher FSB with lower multi
2. lower FSB with higher multi
my mem can do 280MHz in case anyone was wondering. thanks
Well it depens on how your motherboard can handle high FSB .
300x9 and 270x10 there are no difference between performance.
Hey guys,
I am not sure if its teh RAM Ive killed, the mobo or the cpu...
Basically I have been running at 1.425vcore with 3vdimm for the past 3 months on my sandiego FX-55, been flawless in windows but in coutner strike it alwaysused to crash out every now and then with the error "the memory could not be "read"".
I installled a vapochill the other week, all running great but after about 5 days of having it on, I kept getting "SYSTEM_FATAL_EXCEPTION" BSOD of some of the kind - and then it would fail to boot up anymore (well it would get past windows loading screen then before login screen jsut BSOD for half a second and then reboot - it then got worse to just restart)...
It has stayed that way ever since pretty much, though if I move the RAM around into different sockets I can get my win2k hdd to boot up (I think i did it with a single stick of VX in dimm #1 on my dfi lanparty sli-d)
now I am just having ussues getting the damn thing to POST so I am letting it warm up (just had it at -32 hehe) and then I shall put back on the thermaltake big typhoon after taking out and reinserting the cpu.
I tried installing windows xp last night on my winxp hdd, and when it got to the detecting hard drives bit, it just BSODED with this message:
PAGE_FAULT_IN_NOPAGED_AREA
and technical info is:
*** STOP: 0x00000050 (0x00000001,0Xf74DA20,0x00000000)
*** ntfs.sys - Address F747DA20 base at F747B000, DataStamp 3d6de5c1
So really, I am not sure if its cold bug, RAM issues, messed up mem controller or mobo / bios issues.
The cpu hasnt had any decrease in performance overclockign wise for sure - still clocks 3GHz at 1.425vcore, though I think the vcore drips to 1.392vcore when fluctuation occurs....
Wonder if any of you could assist me? :)
Thanks so much in advance - I really could do with some help here, been off the pc for bout 5 days almost now and am really starting to feel ill from the stress of it all :p:
Alec
UPDATE:
Just gont onto air cooling and put another stick of ram in, booted fine and now installind windows XP with not a single BSOD in sight! :D
Me hopes it was just either condensation or cold bug then.
well, nice then. my luck im not running those winbond voltage eaters
Yo guys anybody burned a cpu using high but reasonable cpu vcore with high vdimm?
To my knowledge no death of a CPU has been directly attributed to this phaenomenon yet.
I always wanted to take the cheapest 939 CPU I can get and kill it just to verify the theory. At this time I don't have any Winbond memory left, however, so no go.
I ran stock vcore (1.37v) on my 146 on my ultra-d with 3.4v in the RAM with no issues.
none with any other vcore or vdimm (havent gone over 3.6v vdimm)
Hopefully new boards have fixed this problem if it existed (I just got my mobo)
Hopefully I will have 2gb of uccc coming my way eventually anyway
I have an Opteron 144@1,47v and BH-5@3,5v is this safe?
Use the conservative formula to be safe:
vdimm/2 - vcore > 0.325v
Yours is 0.28v so you're OK :fact:
so if my math is correct (and my math sucks) my ram run at 3.5v needs to have at least 1.425v on the cpu core?