I was reading this thread which actually is about the P35C-DS3R and something about overvolting ram, I have the NON-C version, P35-DS3R, is this something I need to be concerned about? I'm running 6400 crucial ballistix.
thx
thx
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Does the P35-DS3R overvolt ram?
- Thread starter Techx
- Start date
honolululu
n00b
- Joined
- Jul 19, 2007
- Messages
- 44
The p35c overvolts plenty. Especially considering most people think they are starting from 1.8V in BIOS, which is only true if you have RAM meant for 1.8V. It will start you at over 2.0V if that's what your RAM wants. I talk about this in my last post in that thread.
Does the newest BIOS for your board tell you the exact voltages and not pass/fail?
EDIT: I've seen speedfan read the RAM voltage accurately as Vcore2. I'd check there if it's not in your BIOS.
Does the newest BIOS for your board tell you the exact voltages and not pass/fail?
EDIT: I've seen speedfan read the RAM voltage accurately as Vcore2. I'd check there if it's not in your BIOS.
BillParrish
Supreme [H]ardness
- Joined
- Aug 25, 2006
- Messages
- 7,519
Having read every Gigabyte post on this forum for the past 16 months I would say no and would very much like to see the "There is evidence..." evidence. I have never heard of such a concern before on any Gigabyte board. However the C version of the board is probally not nearly as common so it is possilble and I do not have one to test. The amount of "guessing" and incorrect information in that entire thread is cause for concern.
For example:
The Memory Performance bios option has nothing to do with overclocking profiles but is instead a method of reducing latencies in the memory subsystem by eliminating stages of data buffers. This is the PAT (Performance Acceleration Technology) Intel used as exclusive technology for the high end 875P chipset and then Asus figured out how to implement on lesser chipsets. A Google of "PAT Asus" will bring up reams of relevant data.
There is way too much made of voltages read by 3rd party software.
Be aware I am not throwing rocks at anyone, Gigabyte does a horrible job of documenting this feature along with several others. I had to inquire repeatedly to Gigabyte tech support as to this features function, I was first told it enabled EEP profiles. That was easy enough to test and discredit. Several follow ups apparently shamed the tech support people into actually talking to an engineer and educating themselves as well as me. Bandwidth testing with the different options seems to support the assertion it is PAT. However to this day I know of no Gigabyte official document or publication that fully describes Memory Performance Enhance and so I could be wrong.
There could be a memory voltage issue with the C version, I am almost certain it does not exist with any other Gigabyte or we would be hip deep in "Gibabyte killed my ram" like we are in "680i killed my ram, twice!) posts.
or I could be totally wrong.
For example:
The Memory Performance bios option has nothing to do with overclocking profiles but is instead a method of reducing latencies in the memory subsystem by eliminating stages of data buffers. This is the PAT (Performance Acceleration Technology) Intel used as exclusive technology for the high end 875P chipset and then Asus figured out how to implement on lesser chipsets. A Google of "PAT Asus" will bring up reams of relevant data.
There is way too much made of voltages read by 3rd party software.
Be aware I am not throwing rocks at anyone, Gigabyte does a horrible job of documenting this feature along with several others. I had to inquire repeatedly to Gigabyte tech support as to this features function, I was first told it enabled EEP profiles. That was easy enough to test and discredit. Several follow ups apparently shamed the tech support people into actually talking to an engineer and educating themselves as well as me. Bandwidth testing with the different options seems to support the assertion it is PAT. However to this day I know of no Gigabyte official document or publication that fully describes Memory Performance Enhance and so I could be wrong.
There could be a memory voltage issue with the C version, I am almost certain it does not exist with any other Gigabyte or we would be hip deep in "Gibabyte killed my ram" like we are in "680i killed my ram, twice!) posts.
or I could be totally wrong.
honolululu
n00b
- Joined
- Jul 19, 2007
- Messages
- 44
Are you considering the BIOS as 3rd party software?
Cause here I am with F5 , 1.8V=normal right? then +0.2V set in BIOS=2.144V?
Yet with my cheapo kingston pc2-5300 normal sets me at 1.89V.
But again I read your disclaimer about lack of experience with the C. I almost wish I had a lack of experience with this board since I don't see myself buying new RAM before I buy a new board. Live and learn I guess.
BillParrish
Supreme [H]ardness
- Joined
- Aug 25, 2006
- Messages
- 7,519
Shit, wish you had linked this first.
http://forums.tweaktown.com/showthread.php?t=25275
YOU ARE CORRECT.
BTW sorry if my posts come off a bit "edgy" most here are used to me. As an engineer my social skills are a bit lacking and it annoys me when I do not understand something. I am just trying to get to the bottom of things or at least come up with a rational understanding of what is going on as I have (unintentionally) self appointed myself the resident Gigabyte. http://www.hardforum.com/showthread.php?p=1030795318 I have no doubt you saw what you describe, I am just tying to figure out, if its is your board, all board, your memory, all memory etc. etc. Plus as I do not have one of these on hand your input is critical and as I investigate and post my findings here I would hope other "C" owners or even non-"C" owners who measure ram voltages that seem "off" would chime in. The 680i has a rep for eating memory. I hope the Gigabyte does not acquire one unless deserved.
The correct answer based on my dealing with them is "all of the above"
After doing much more investigation than is reflected here I agree with you. I would be VERY wary of any Gigabyte board with support for both types of ram and would be checking ALL Gigabyte boards. Speedfan Vcore2 typically reports Vdimm fairly accurately.
The P965's do not do this. P35C's it seems certain they do have the problem (bios version might matter), other P35 and X38's unknown.
This should not be an issue with settings of +.2V and below as most ram will take 2.1ish volts without damage but I would be very wary of settings of +.3 and above. I would therefore (and it would be silly to do it anyway) not slam in +.3 or +.4 without determining what speedfan reports Vcore2 (Vdimm really) to be. It is also wise to check the ram temp, even with just your finger, Hot is not a difficult concept and if the ram is hot, back off Vdimm and/or get some air on it.
Thanks for posting this here.
To the OP, just watch it and it should not be a concern. As long as you know what is going on, you can conpensate.
http://forums.tweaktown.com/showthread.php?t=25275
YOU ARE CORRECT.
BTW sorry if my posts come off a bit "edgy" most here are used to me. As an engineer my social skills are a bit lacking and it annoys me when I do not understand something. I am just trying to get to the bottom of things or at least come up with a rational understanding of what is going on as I have (unintentionally) self appointed myself the resident Gigabyte. http://www.hardforum.com/showthread.php?p=1030795318 I have no doubt you saw what you describe, I am just tying to figure out, if its is your board, all board, your memory, all memory etc. etc. Plus as I do not have one of these on hand your input is critical and as I investigate and post my findings here I would hope other "C" owners or even non-"C" owners who measure ram voltages that seem "off" would chime in. The 680i has a rep for eating memory. I hope the Gigabyte does not acquire one unless deserved.
The correct answer based on my dealing with them is "all of the above"
After doing much more investigation than is reflected here I agree with you. I would be VERY wary of any Gigabyte board with support for both types of ram and would be checking ALL Gigabyte boards. Speedfan Vcore2 typically reports Vdimm fairly accurately.
The P965's do not do this. P35C's it seems certain they do have the problem (bios version might matter), other P35 and X38's unknown.
This should not be an issue with settings of +.2V and below as most ram will take 2.1ish volts without damage but I would be very wary of settings of +.3 and above. I would therefore (and it would be silly to do it anyway) not slam in +.3 or +.4 without determining what speedfan reports Vcore2 (Vdimm really) to be. It is also wise to check the ram temp, even with just your finger, Hot is not a difficult concept and if the ram is hot, back off Vdimm and/or get some air on it.
Thanks for posting this here.
To the OP, just watch it and it should not be a concern. As long as you know what is going on, you can conpensate.
honolululu
n00b
- Joined
- Jul 19, 2007
- Messages
- 44
I looked on my pc health bios screen and where you have voltage values I just have "OK" next to vcore, ddr18v, +3.3v, and +12v.. so how am I suppose to know if there is a problem or if i have to make some adjustment? I haven't built a new system in years so if anyone can 'newbify' the conversation I'd apprecaite it.
Thanks guys
honolululu
n00b
- Joined
- Jul 19, 2007
- Messages
- 44
Do as BillParrish advised and monitor using Speedfan Vcore2, and don't go above +0.2V in BIOS until you're sure of what your board is doing.
Do you have the latest BIOS? It was only the latest F5 version for the p35c that replaced the OK/FAIL with the actual voltages. And with the old BIOS the RAM voltages would read FAIL when set to +0.3V in BIOS.
Funny I actually just updated to the latest bios, NOW i can see those values. I have
Vcore ------ 1.332v
DDR18v --- 1.808v
+3.3v ------- 3.376v
+12v -------- 11.985v
does this look ok