GA-X79S-UP5-WIFI and 1.6V+ DIMMs?

[RootWyrm]

I've been searching for about an hour or so, and I'm not finding any answers, so may as well ask.

How is the GA-X79S-UP5-WIFI with higher than spec voltages? (Anything over 1.5V is higher than spec.) I'm rifling through a stack of potential candidates for 32GB in 4x8 which respond differently to voltage bump/drop. Given the throttling issues, which are still largely uncorrected, I'm a bit skeptical of stability at 1.6-1.7V, and given the bugs, kind of nervous about 1.35V LVDIMMs not getting overvolted.

Any experiences to share? Not interested in the overclocking bits or such - your results will differ from mine, of course - just whether or not stability is possible and voltage is stable.

[wazza300]

Re: GA-X79S-UP5-WIFI and 1.6V+ DIMMs?

you mean memory dram voltage? you'll be fine to use anything from stock 1.3v to 1.65v depending on how much needed to remain stable for the memory amount/speed ect

[RootWyrm]

Re: GA-X79S-UP5-WIFI and 1.6V+ DIMMs?

you mean memory dram voltage? you'll be fine to use anything from stock 1.3v to 1.65v depending on how much needed to remain stable for the memory amount/speed ect

Ayup, VDIMM.

My biggest concern with <1.5V is overvolting - the X79S-UP5 for some reason, seems to struggle to stay stable at 1.4V. Lots of wiggling around 1.38-1.43. The banding at 1.5V is much narrower - 1.49-1.51V. I'm at a loss as to why that might be other than BIOS bug, but honestly it's at the point where the best fix is to scrap F4 and start over. (And maybe fix the damned OROM and AC97 disable issues.)

Mostly though, I'm looking at some kits which demand 1.6 or 1.65 in XMP (4x8GB, as I mentioned) around 1866 and I'm somewhat concerned that anything over 1.5 is going to negatively impact total stability. So hoping some of you out there have been using 1.6-1.65's and have results to share. If it BSoDs regularly, it's not gonna fly. It was hard enough addressing the RSTe issues, so I'm understandably skittish with this board. It really seems to be operating at/near margins on the north end of things as is. :(

Still better than the others I tested - holy crap, what moron thought just turning the SAS ports back on was a good idea? - but still one of those 'handle with extreme care unless you like BSoD' situations.

[wazza300]

Re: GA-X79S-UP5-WIFI and 1.6V+ DIMMs?

Yh with 32gb its gonna be a struggle for stability with the CPU overclocked,more CPU/vtt and system agent voltage will help as will looser timings

[RootWyrm]

Re: GA-X79S-UP5-WIFI and 1.6V+ DIMMs?

Yh with 32gb its gonna be a struggle for stability with the CPU overclocked,more CPU/vtt and system agent voltage will help as will looser timings

*points to specs, specifically the 'modified' part*
Not struggling in the least. If it was struggling, I wouldn't have stable VDIMM at 1.5 vs. 1.4. (Seriously, if anyone can figure out what the hell is with the undervolting being weird, I am ALL ears.) The problem is that the stuff I have will go CL9->CL8 and a HUGE tRAS drop from 29 to 24, if I can get it to hold <=1.4. At 1.42, CL craps out. It's perfectly happy at 1.5V relying on XMP's SPD with CL9 and tRAS 29. And seriously - don't even bother trying to overclock this board unless you're ready to get hip deep in it - especially the BIOS.
TL,DR: board's fine at stock voltages and Intel VDIMM spec (1.5V), even pushed hard. No component issues other than personal dislike for the IOH/VRM sink. BIOS just sucks.

Most of the reports I've found on 1.65 is for 8 DIMM configurations for anything 1.6V+, with nothing on 4 DIMMs at 1.6-1.65 so I have no idea whether it's a sum draw issue or the vregs and/or trace layout and/or obnoxious baseline BIOS just not liking 1.6. That's what I need to know: if it will hold stable voltages with 4 DIMMs or if it's going to behave badly at 1.5V+ like it does with 8 DIMM configs.

[stasio]

Re: GA-X79S-UP5-WIFI and 1.6V+ DIMMs?

Did you try to replace (flash) your BIOS "SHRIKE/GEN6 (F4V Derivative)" with simple F4x?
Btw,
all DDR3 memory are capable to overvolt (24/7) at least 3-5%.

[RootWyrm]

Re: GA-X79S-UP5-WIFI and 1.6V+ DIMMs?

Did you try to replace (flash) your BIOS "SHRIKE/GEN6 (F4V Derivative)" with simple F4x?
Btw,
all DDR3 memory are capable to overvolt (24/7) at least 3-5%.

There's a reason I'm running modified - and it ain't 'cause stock is well-behaved. Or stable. As I said, frankly F4's such a wreck the only answer is clean slate, but that is a LOT of cash outlay without an existing core license. (Even before the LPC IP core licenses, ugh.) But no, the board isn't any more stable under 1.5 with stock BIOS. It just does not seem to like LVDIMMs and I can't find a reason why, unless it's just this board.

And whether or not it fries them is irrelevant; the issue I'm concerned with is stability and undervolting above 1.5. 64GB in an 8x8GB at 1.65V apparently ends up actually getting around 1.62ish. Which is fine - the only DIMMs that work for that config hit specs at 1.60-1.65. (And yes, you can do 4x16GB if 1. you can afford it 2. you don't mind a drop to 1333 with CL9.)

Problem is, as I have said, a lack of data on a 4 DIMM 1.65V configuration and whether it shows the same undervolting issues. It's not droop - it's just plain undervolt apparently. My suspicion is that the 8 DIMM configuration experiences the drop as the trace layout can't really support 8, but I don't have any data to support that or confirm it. So, the question is whether it's going to undervolt or have unstable voltage at 4 x 1.65 knowing that it does undervolt at 8 x 1.65.

It might be more helpful to understand this: I'm looking at high spec clock 4 set DIMMs that require stable 1.65V with banding of -0.02+0.03 to hold timing and I absolutely will not accept looser timings than manufacturer spec. Unstable voltage screws timing and that's not going to fly. And the work done on this machine absolutely hammers memory, so timings matter a LOT more than clock.
So is anyone running 1.65V quad and watching their voltage? Is it stable at 1.65V or is there undervolt and swings? Or is everyone just running 1.5V and overvolting?