Printable View
ok im running 1.8rev3 and for some reason the voltages are whack.
My 3200 Venice is locked at 1.45v i've raised it to 1.5 and it doesnt go to 1.5 it is reported as 1.45 by every app i can find in windows.
even changing it with clockgen doesnt work.
anyone seen this before ? i cant tell if its a bios issue or a venice vcore lock.
LBBE 0519 APAW 3200+
the board undervolts...usually up to .1v for me, its .1 less than whatever you set, so i set 1.55+10% to get 1.62v.
Quote:
Originally Posted by Lestat
Yes thats normal 1.45 is the maximum the vid will allow for venice san diego chips. To raise the voltage anymore u use the pecentages. Maximum setting is 1.45 + 10 percent which = 1.63 volts or so.
Try 1.45 + 5 percent that makes about 1.52 volts and is a good place to overclock from on air cooling.
xb and xc? Its kind of hard to answer since we dont know the manufacturer (hyinx, samsung, winbond, etc) and you havent given us any other info off the chips to indicate model numbers or speed ratings. Do you have the ram in the two slots closest the CPU or two farthest from it (dual channel on the Neo2) or are they in 1 and 3 or 2 and 4?[/QUOTE]
ok i'll take the ram out later tonite and give you all the info i can from it.
doesnt matter if i run it in single or dual channel. atm i have it on the 2 slots closest to proc. but i can run any combination of using both sticks and it drops back to 166.
let me write everythign down and get back....thanks for the help so far :)
Ok just got done trying the 1.98 bios modded by nfm.doesnt like my mobo/cpu at all, I want to make it clear I am not bashing nfm with the following at all. currently on 1.8 rev3 by nfm, with 1t fix. 250x11 perfectly stable, ram 1:1. 275x10 with ram 1:1=no good, this was what I was trying to fix. now results with 1.98. Basically at stock(with default craptastic timings), it would post, but not load windows. Had to remove my pci ata controller card to get into windows. 250x11 = post but no boot, even without pci card. 275x10 = no post, 1.8 would get into windows, just fail superpi and prime after a while. so my experience with this bios is that its garbage, I await the final, and may even revert to 1.8 if these problems persist, I despise getting down on the floor and changing the cmos jumper.
@Tyrvidar ok so 270*9 is 2430mhz ok, so bump the vcore depending on ur cooling to about 1.55 to 1.65 for the best of air cooling xp-94s and such.
Keep your memory very low and laxed to keep it out of the question, and the htt mulit to keep it at 3.
A fast way to get the max cpu mhz i find is to set cpu to 2300 lets say and 1.65 volts prime look at heat, judge if its to much heat, anything hihger then 55c is to high, then determine ur max vcore then keep bumping up fsb with clockgen and doing superpi 1 mill till it doesnt go anyhigher then back off 10fsb and start priming. :D then once u get the max primeable, start to lower vcore and prime :D.
Did any one tried some mushkin redline 3500?
With a OCZ DDR BOOSTER?
TKS!
I appreciate your Help 4rory alot. I will hold off on testing this to the limits and in the meantime will let it burn in. Will look into one of the xp94s and such over the next month. Stock cooler can only do so much hehe.
Thanks again ;). Some noob o/cer will see this thread and use your post you helped me with.
I recently upgraded my CPU to a manchester (X2 4200+), and no BIOS I've found yet lets me boot all the way into windows with both the CPU and my SATA drives being recognized properly.
1.8 : dual CPU is not recognized
all 1.9 betas: drive on SATA port #3 works fine, but drive on port #4 shows up as "new hardware", and windows wants to install a driver for it(?!), and the partitions aren't available.
anyone else run into a similar problem ever?
have you tried using xp?Quote:
Originally Posted by angra
Not yet - I'd really rather not rebuild my OS right now; I have a ton of apps installed, etc., and it's only been six months or so since my last rebuild.Quote:
Originally Posted by sideeffect
Both SATA ports seem to work fine under 1.8, so I strongly suspect a BIOS issue of some sort. It's just odd.
though I do recognize that XP is a lot more likely to have up to date support and attention paid to it, so maybe I will end up going that way.
fortunately, the drive in question is used only for nightly backups so I can live without it for a while at least.
next test will be to switch 3&4 and see if the problem stays with the drive or the port.
Msi has been working on the sata ports trying to sort out the locking issue so its possible that the bios has messed it up :)
From what ive read though the 1.98 beta bios locks port 1 and 2 so pehaps its worth testing it out and see if you can boot that way but probably not best to plug it in at 280 htt. :stick:
Yeah I noticed comments along those lines in this threat, I think that's worth a try, too, and I will.
Since the drive itself is purely a backup drive, I can afford to take more risks than normal in the data-corruption area.
@Tyrvidar
Glad I could help.
With the 1.36mod bios I can get volts up to +18.3% with venice my max effective vid is 1.450v correct? does this mean that if I apply +18.3% I get 1.71vcore? and then the board undervolts also? but 0.1volt right? that means max I can get is 1.61vcore this seems a bit low, I just want to know if I should try this bios and try the high volts to see my max overclock, I'm on air and am willing to apply a bit over 1.65 for a nice screeny, already got a 3131mhz on a +3200. That was with 1.8mod bios at 10% i think cpu-z reports close to 1.65 volts.
No the vcore reading is accurate. So 1.45 + 18 percent will give you about 1.7 volts. The chip doesnt undervolt it just wont accept a vid higher than 1.45. The percentages take it higher than that.
The newcastles would do 1.55 + 18 percent as their maximum because they had higher voltage defaults to begin with.
The vid is controlled by the cpu not the motherboard.
Ok so my venice with all voltage settings maxed on 1.36mod bios will give me a safeish testing vcore? 1.450+ 18.3% will be ok for venice?
jus for benching.... i wouldn't recomend running it 24/7 with it...Quote:
Originally Posted by 4rory
1.8 BIOS allows u to run 1.7v as well... using corecenter
this whole NEO2 + Venice voltage issue is a joke
MSI needs to fix this.
the voltages are off from what you have them set in the bios usuall .25v
plus most of the settings dont work
i had the bios set to 1.7v and it was only like 1.6 in windows.
or rather Core center said it was 1.7 but reported it as 1.6
totally screwed up
this needs to be addressed and quick before people start frying chips.
i can never tell what im setting the voltaged too unless i reboot and go back into pc health. and then when i get into windows sometimes its even different.
I cant even get 1.475 to 1.575v its a huge dead zone up to 1.45 its ok and from 1.6 and up its ok. thats a lovely hole in the voltage spectrum.
i am almost done with this board. i think this venice will be my last amd64 for a long time unless i win the lottery.
i do believe im gunna have to get me an Ultra-d msi just cant made this mobo with too many issues and now the venice's voltage issue;s yeah...
The only bios i have not tried is the 1.36mod becuase its too old for this venice and i dont feel it would be wise going back to such an old bios.
considering doing a droop mod to my board..
I can see what you're saying about a "voltage hole," but I haven't had any issues at all yet. I know that if you set the VID to anything higher than 1.45v with a Winchester or a Venice, it will only see 1.45v. (i.e. 1.55v VID = 1.45v VID) The 1.37mod or 1.36mod bios, the one that supports 18% over-voltage will give you finer steps than any other bios, but I really don't see that much of a point unless you want over 1.6v for constant use. :shrug:Quote:
Originally Posted by Lestat
If you don't have time to calculate what each VID and %overvolt is, here's a chart for you. ;) Just find the voltage that you want to try and use the selected VID and %....
1.325v * 3.3% = 1.3687v
1.325v * 5% = 1.3912v
1.35v * 3.3% = 1.3945v
1.35v * 5% = 1.4175v
1.375v * 3.3% = 1.4203v
1.325v * 8.3% = 1.4349v
1.375v * 5% = 1.4437v
1.4v * 3.3% = 1.4462v
1.325v * 10% = 1.4575v
1.35v * 8.3% = 1.4620v
1.4v * 5% = 1.47v
1.425v * 3.3% = 1.4720v
1.35v * 10% = 1.4850v
1.425v * 5% = 1.4962v
1.45v * 3.3% = 1.4978v
1.375v * 8.3% = 1.4891v
1.375v * 10% = 1.5125v
1.4v * 8.3% = 1.5162v
1.45v * 5% = 1.5225v
1.4v * 10% = 1.54v
1.425v * 8.3% = 1.5432v
1.425v * 10% = 1.5675v
1.45v * 8.3% = 1.5703v
1.45v * 10% = 1.595v
I have it up to 18.3% on a sheet of loose-leaf, but I've spent too long typing this up as it is....:P And I know that the board won't be as precise with voltage as my calculations are...it will however give you an idea concerning which direction the voltage will want to swing. (i.e. 1.57v will want to swing up and 1.53 will want to drop a little.)
Hope this helps somewhat..... :rolleyes:
(I tried to line it up nicely, but it's getting too late to even worry....5:15am, no sleep as of yet...)
Hi guys its not msi's fault that the winchester venice and san diego cpus wont go above 1.45 vid. Any setting above 1.45 vid will be ignored by the cpu.
Its because newcastle had a stock voltage higher than the 0.9 nm cpus. Neo 2 was built before 0.9nm cpus came out and they need to leave the settings up to 1.55 for the newcastle cpus.
The vcore is about right. 0.25 difference isnt a lot if thats true all motherboards are never spot on because they use cheap sensors.
There is a thread about venice cpus and voltage and they said lots have been frying with 1.75 volts. They didnt recommend more then 1.65 for 24 / 7 usage
and even then keep the chip cool.
I only use 1.65 with my vapochill.
just bought a 6800 gt :banana: :banana: :banana: :banana: :banana:
sale on at www.overclockers.co.uk
Leadtek WinFast GeForce 6800 GT 256MB = 165 pounds inc vat
http://msi.designlab.info/thread.php?threadid=17656
Murdock has some mod's on the 1.97 and 1.98 bios. these are new. they also have a new version of AWD Flash
Murdock also says that they are a new bios layout and such. basically he says its a new code or new sequence layout of the bios.
a guy actually spoke english in there and they replied in eglish.. must be nice i have posted in english and no one replies. ggrr
his issue does not seem to be related to the bios but more so a mod he did to the temp sensor.
has anyone tried either of those 1.97 or 1.98 bios ?
Sideeffect -
i wont run mine above 1.65 i mean be realistic the default voltages are only 1.38 or 1.4 thats a HUGE voltage jump.
mine is water cooled and runs idle 28-32c(depending if im running stock volts or more) and during Prime95 45c MAX.
that is even running 1.65v
the thing is i cant get the bios to give me an accurate setting. basically i have that hole in the vdimm from 1.45 all the way thru 1.6
and .25v different i didnt mean .25 i actually meant .025 but after during more experimenting i found that th ehigher vcore you use in the bios the more the vcore is wrong in corecenter and cpu-z
basically corecenter will report 1.7v for what you have it set,. but its real reading or realtime reading says 1.6
so there is a difference of .025 all the way to .1v thats a major screw up. and its the bios that can control this even though yes many people report this issue but how mnay report this issue with the DFI boards ? or asus or gigabyte ?
And Clockgen will not set the vcore or the multiplier. its like it doesnt work with the venice or something.
Corecenter will adjust the vcore properly BUT dont tell me or anyone else to use it simply cuz thats silly the bios should be handling the voltage settings porperly.
I told people when the venice came out to go buy a NEW board . this damn NEO2 has enough issues running 2 year old cpu's let alone a brand new model.
I guess i should have followed my own advice.
damn pci-express i aint upgrading to something that doesnt give me more performance. not to mention the insane cost facotr of doing so.
but thats the only way im going to get a NEW mobo that ould support this venice properly i suppose. unless of course ALL mobo's are having these voltage issue's.
Lestat are you using the most recent version of corecenter because corecenter, speedfan, systool, mbm, cpuz and the bios all show the same vcore for me. There is a slight fluctuation to the reading but that will be the sensor.
As has been discusssed lots before the sensors on motherboards are cheap and show variations that dont exist when you use a digital multimeter.