RAM Issues on X58 Boards

This happened on my EVGA X58 SLI when I was pushing the BLCK too high. I'd change the voltages around, lower the blck, eventually it would detect all 3 DIMMs again.

Same here, it seems that in this case it is more of issues of the vcore, qpi, and pll getting in sync as the blck changes.
 
Motherboard: ASUS P6T Deluxe X58 v2
CPU: i7 920
RAM: OCZ Platinum 1600 MHz

Problem: When 12 GB of RAM is installed, ran at default system settings, only 10 shows up in BIOS. I had to RMA 3 of these motherboards in a row, all having the same problem, to know that it had a glitchy A2 DIMM slot.

Notes: When overclocking the BClock, 12 GB of RAM shows up. Strange, but had to be noted.

Solution: Bought a Qualified Vendors List memory set (12 GB Corsair Dominators), and it's been without a hitch since.

Conclusion: This motherboard may be intolerant of non-QVL memory.
 
Hi long-time lurker, and finally decided to register and share some information.

I am working at a retail store and have a customer who ordered the following parts (I am detailing serial and batch numbers for reference... if it's not okay, please edit them out).

Mainboard
Asus P6TD Deluxe <Green> - Serial no. : A1M0AC876468
LGA 1366 socket manufactured by LOTES
BIOS @ 0501

CPU
Intel Core i7 980X - FPO/Batch 3013A557
Using stock Intel air cooling unit

RAM
OCZ PC10666 OCZ3G1333LV6GK 6x 2GB 1.65v

Population as follows:
DIMM_A2 010431004000743-1
DIMM_A1 010431004000736-1
DIMM_B2 010431004000743-2
DIMM_B1 010431004000736-2
DIMM_C2 010431004000743-3
DIMM_C1 010431004000736-3


PSU
Coolermaster GX750W

Graphics
Leadtek Quadro FX1800

Constants
Ai Overclock Tuner – Manual
CPU Ration Setting – Auto
Intel SpeedStep Tech – Enabled
Intel TurboMode Tech – Enabled
BCLK Frequency – 133
PCIE Frequency – 100
UCLK Frequency – Auto
QPI Link Data Rate – Auto

Tested Parameters
DRAM Frequency - DDR3-1333
QPI/DRAM Voltage - Auto
DRAM Bus Voltage - Auto
Result – 4088MB registered

DRAM Frequency - DDR3-1333
QPI/DRAM Voltage – 1.275v
DRAM Bus Voltage – 1.64v
Result – 4088MB registered

DRAM Frequency - DDR3-1333
QPI/DRAM Voltage – 1.275v
DRAM Bus Voltage – 1.66v
Result – 4088MB registered

DRAM Frequency - DDR3-1333
QPI/DRAM Voltage – 1.275v
DRAM Bus Voltage – 1.64v
DRAM Timing – 9-9-9-24-Auto-Auto-Auo-Auto-Auto-Auto
DRAM Timing Mode – 2N
Result – 4088MB registered

DRAM Frequency - DDR3-1333
QPI/DRAM Voltage – 1.275v
DRAM Bus Voltage – 1.64v
DRAM Timing – 9-9-9-20-Auto-88-Auo-Auto-Auto-Auto
DRAM Timing Mode – 2N
Result – 4088MB registered

DRAM Frequency - DDR3-1333
QPI/DRAM Voltage – 1.35v
DRAM Bus Voltage – 1.64v
DRAM Timing – 9-9-9-20-Auto-88-Auo-Auto-Auto-Auto
DRAM Timing Mode – 2N
Result – 4088MB registered

Ai Overclock Tuner - XMP
DRAM Frequency - DDR3-1333
QPI/DRAM Voltage – 1.275v
DRAM Bus Voltage – 1.64v
DRAM Timing – 9-9-9-24-Auto-Auto-Auo-Auto-Auto-Auto
DRAM Timing Mode – 2N
Result – 4088MB registered

DRAM Frequency - DDR3-1066
QPI/DRAM Voltage – Auto
DRAM Bus Voltage – Auto
Result – 12280MB registered

DRAM Frequency – DDR3-1066
QPI/DRAM Voltage – 1.275v
DRAM Bus Voltage – 1.64v
Result – 12280MB registered

Take note for DDR3-1006.... After a few cold reboots, the RAM will mysteriously disappear and show only 4088MB or 2048MB.

I send the entire motherboard and RAM back to the distributor. They did their testing and (according to them) swapped out several motherboards and memory modules... and no solution.

Annoyed, and with my customer breathing down my neck (it's not a gaming machine), I asked for the return of the motherboard and RAM... and to do my own testing and troubleshooting.

Tried all the methods and suggestions documented on various sources (Asus, OCZ and so on), to no success and much frustration.

For shit and giggles, I drew out some el-cheapo Kingston KVR1333D3N9/2G 6x 2GB (1.5v) RAMs.

Constants
Ai Overclock Tuner – Manual
CPU Ration Setting – Auto
Intel SpeedStep Tech – Enabled
Intel TurboMode Tech – Enabled
BCLK Frequency – 133
PCIE Frequency – 100
UCLK Frequency – Auto
QPI Link Data Rate – Auto

Tested Parameters
DRAM Frequency – Auto (DDR3-1066)
QPI/DRAM Voltage - Auto
DRAM Bus Voltage - Auto
Result – 12280MB registered

DRAM Frequency –DDR3-1066
QPI/DRAM Voltage - 1.275v
DRAM Bus Voltage - 1.64v
Result – 12280MB registered

DRAM Frequency – DDR3-1333
QPI/DRAM Voltage - Auto
DRAM Bus Voltage - Auto
Result – 12280MB registered

DRAM Frequency – DDR3-1333
QPI/DRAM Voltage - 1.275v
DRAM Bus Voltage - 1.64v
Result – 12280MB registered

Finally the fracking thing works... I tested the Kingston memory with Prime95, BurnInTest and LinX (on Win7 Ultimate 64-bit), with no returned errors.

I will test it further when I get in another 6x OCZ Titaniums and update here as well.
 
I was messing with my personal system, here last week doing some cabling that I needed to do in order to move this fire spewing 480 SLI box away from my knees and I gave the case a good jar and it locked up. I rebooted, and only 8GBG of 12GB showed. Rebooted over and over, no go.... Opened up the side, reseated all of the DIMMs, 12GB of the next POST no problem.
 
This is really strange. I had it happen with the first P6TD I bought. I've seen plenty of posts about this problem. I wonder if the memory controller isn't fully baked on the i7's. Hope Gary's emergency wasn't too serious.
 
My brother's Asus P6X58D Premium has had memory issues (showing only 4gb of 6) even after an RMA. He did some research and found that it may be due to a short caused by an incompatibility with the mounting system of the TRUE Spirit heatsink and his mobo. Last I heard he was going to pick up some nylon washers and remount to see if that fixed things.

My EVGA x58 SLI has had minor RAM issues as well. I have 6gb installed and from cold boot it frequently takes a ridiculously long time to post. From what I can tell it retries POSTing a few times until everything checks out, sometimes it goes on the first try sometimes it takes a full minute, but it's always from cold boot. The error code readout displays F3 F6 and F8 during each 'cycle' of the boot attempts. Last time I looked it up I believe I was told these were RAM related despite having the problem persist with 2 sets of ram and being completely stable in windows 7, prime95, and memtest.

Update on this:

Asus P6X58D
My brother has confirmed the TRUE Spirit was shorting on the Asus P6X58D and disabled a slot and the ram stick in that slot.

EVGA x58 SLI
I picked up another set of RAM (OCZ1600) for troubleshooting and it did not fix my problem. Going to try reseating the cpu in the near future.
 
Got an Asus Rampage II Gene, i7-950, and 3x4G G.Skill RAM (DDR3-1333, 1.5v), PC&P 750W ps. All 12G are showing up.

Mainboard reports BIOS to be 0501, so will update that shortly. Everything is stock while I do the first round of sanity testing.
 
had this issue on an EVGA 760 Classified, with a corsair 12gb kit 9-9-9-24
 
I have the EVGA X58 SLI LE(BIOS ver. 74) with the I7 930 and 6GB(3x2GB) of GEIL 1600 8-8-8-28. I have Windows 7 64 bit installed it shows 6 GB with 3.99 GB useable.

And if I had not seen this thread I would not have checked mine.

Well mine is showing all 6 GB now. I was messing with some settings in the BIOS and the system ended up not booting up. So I end up reseting the BIOS and putting my settings back and now i am good to go. :confused:
 
New build:

MSI Big Bang Xpower
Core I7 930
BFG GTX260 (216core)
W7 64bit

I started out with 3 x 2 GB of Patriot PVT36G1600ELK. No problems. After a few days, I bought another set of the same memory. Upon first install it showed 9.99 GB of memory. Shut down, reseated memory, then it showed 10 GB.

The board shipped with the standard shipping BIOS of 1.2. I updated the system to Beta BIOS 1.43, and problem is now solved. System now shows 12 GB of memory.

This is all at stock speeds. Will overclock later, and report if any further problems occur.
 
Last edited:
P6T Deluxe (first version)
Core i7 920
6 x 2 gigs G.Skill 1600
Win7 64-bit

Only once did I not have all 12 gigs available, instead 8 were. Reseating the DIMMs did the trick, and I haven't had the problem since.
 
all fun and games but until someone takes a misbehaving board out of a case where a logic analyzer or O-scope can be used to "buzz" out the memory address and data buss along with a good knowledge of exactly how and where (what lines/traces/pins) the bios uses to detect and size a memory stick along with the JEDEC design document for the sticks, I dont think a universal solution is going to be found. It would also be interesting to know if Intel does test each cpu with a full load on the memory controller during final functional test.

or get a bigger hammer.
 
all fun and games but until someone takes a misbehaving board out of a case where a logic analyzer or O-scope can be used to "buzz" out the memory address and data buss along with a good knowledge of exactly how and where (what lines/traces/pins) the bios uses to detect and size a memory stick along with the JEDEC design document for the sticks, I dont think a universal solution is going to be found. It would also be interesting to know if Intel does test each cpu with a full load on the memory controller during final functional test.

or get a bigger hammer.


I do not think it is a memory controller / CPU issue. I have been able to replicate the issue with 2 year old i7 processors that have NEVER shown the issue previously.
 
I had this issue on my EVGA 758-a1 board when I first got it in Jan 08 with my i7 920 c0. I was using 3x2gb OCZ DDR3 gold. It would only detect as 4gb. After some time, I figured out if I manually set the timings and voltages everything was fine again.

Eventually one of the newer BIOS updates fixed the problem anyways, it was just annoying after clearing the CMOS to have to go set those timings. Save the BIOS settings. Reboot. Then shut down and do a cold boot. Finally after that I could have all 6gb.

Interestingly enough when I used my 6gb of kingston hyperx the problem never existed. I know the hyperx has XMP profiles, but the EVGA board doesn't support them anyways.
 
Wow I somehow missed this thread. I guess it's because I don't come to the MB forum much but anyway I wanted to contribute. I have heard of this issue as well across various forums since socket 1366 was introduced and it pretty much mirrors what we are seeing here, very fucking odd to say the least.

Anyway In my personal experience with this rig in my sig, which I built in jan 2009 orignally, I had some issues in the begining as well. I'm running the orignal vanilla EVGA X58 3xSLI board. This however is my second board since the first one was RMA'd way back about a month after I first built the system for the same reason you guys are talking about here. Sometimes windows would detect 4gigs of my dominators and sometimes on some boots 6gigs. After various bios's and speaking to EVGA techs for days to no avail I decided to RMA the board and it's been fine ever since(knocks on a big hard piece of wood). I haven't updated the bios since may of that year 2009and the machine has been fine. Between then and now I have changed out my 3 gtx 260core 216's for 3 gtx 470's this past april as well as just 2 weeks ago I replaced my CM 1000w power supply for the awesome corsair AX1200 and again it's been ok.

So that's my personal experience and yes Kyle something definetly seems rotten in Denmark. Clearly something is fucked up and everyone is passing the buck. I am VERY interested in hearing how this pans put over time as you put the pressure on the vendors/manufacturers for answers(and rightly so).

Peace.
 
Has anyone tired using Memset or other utility that will read the SPD chip on a board that is a "misbehaving" mode (reporting less than installed) to see if the SPD chip can be read on all installed sticks while the "error" is occuring ? Tested Vdimm and other power signals and/or refresh strobe etc. at the socket of a stick that is "missing' ?

Are SPD settings detected by the motherboard or the mem controllers? I assume the motherboard, but since it seems like at least some people (myself included) are having the issue because the timings aren't being autodetected correctly.

The software stored in the bios should detect and size the computers memory configuration during POST. In order to do that the software must access the memory through the memory controller to query the SPD chip and do memory sizing and testing. As a final step the bios will decompress itself and load into high memory (or it used to work that way).

If you really want to get to the bottom of it someone needs to formulate a testing stragety based on how the hardware is susposed to work and find out where it is failing to work properly.

Anyway I dont have a board so have fun, was just trying to help focus the effort as I am worried swapping war stories will help size the problem and define it but do little to get to the root cause.
 
Asus P6t Deluxe v2 and 6 GB of OCZ DDR 3 1600 would only see 4/6 depending on the day and it would change each time you restart the workstation......Asus was trying to tell me that it could be the onboard mem controller on my I7 920 so I RMA that and get a brand new one back same issue...Then they blame the OCZ ram so I RMA that and get a brand new set same issue- last I send in my motherboard and get back a refurbished one (looks new) and never had issues since then :rolleyes:
 
Has anyone tired using Memset or other utility that will read the SPD chip on a board that is a "misbehaving" mode (reporting less than installed) to see if the SPD chip can be read on all installed sticks while the "error" is occuring ? Tested Vdimm and other power signals and/or refresh strobe etc. at the socket of a stick that is "missing' ?



The software stored in the bios should detect and size the computers memory configuration during POST. In order to do that the software must access the memory through the memory controller to query the SPD chip and do memory sizing and testing. As a final step the bios will decompress itself and load into high memory (or it used to work that way).

If you really want to get to the bottom of it someone needs to formulate a testing stragety based on how the hardware is susposed to work and find out where it is failing to work properly.

Anyway I dont have a board so have fun, was just trying to help focus the effort as I am worried swapping war stories will help size the problem and define it but do little to get to the root cause.
I hear ya, no point in knowing it happens anymore (since it seems pretty damn common), but rather how to fix it.

I'm no electrical engineer, but I can reproduce the problem fairly regularly on my board. I will try a few things to see if anything changes.
 
I tried reseating the CPU and the heat sink... no dice. I have tried loosening the tension on the heat sink as well... no go as well.

An update to my previous post, the 12GB of OCZ Titanium worked out ok as well. No problems so far.
 
My EVGA X58 SLI LE would only recognize 4GB of XMS3 1333's. The same thing happened when I tried the XMS3 1600 sticks. Swapping out processors didn't fix it either. I just returned my mobo to MC and got a new one and that one is working great.

Funny thing though, when I put all 6 sticks in, my computer recognized all of them and reported that I had 12GB. I've had no problems with my ASUS P6T with any combination of RAM or CPU.
 
I've been having this issue on my computer... very irritating.

Intel i7 920 D0
Foxconn Bloodrage GTI
Mushkin Enhanced Blackline 6GB (3 x 2GB) #998659

I have 6GB installed, CPUZ confirms this. Windows also sees 6GB but only 3.99 is usable.

My motherboard only has 3 slots (1/3/5).

Troubleshooting steps I have taken:
-Was looking around bios for a "memory remapping" option, couldn't find it, so it must be enabled by default.
-Set RAM to stock timings 9-9-9-24
-Tried a BIOS update from P07 (March 2009) to G42 (July 2010)
-Removed my heatsink/processor to check for any damaged pins (none found)
-Removed RAM modules, cleaned the gold connectors with a Q-Tip and rubbing alcohol, then reseated them in a different order. The pins in the ram slots looked good too.
-Removed two of my RAM modules and left the one closest to the CPU in (should be A1). I was able to boot up just fine, and it registers 2GB.
-Tried going to Start>msconfig>Boot>Advanced>and unchecked the maximum memory box

There are only two things I haven't tried... One is setting voltages and all that nitty gritty stuff manually - I haven't done this yet though because I'm not that well versed in what everything should be set at. The other is just booting up with a single module in my other two slots, although I'm not sure if that is supposed to work anyway IE. if I have a module in slot 3, but not 1 or 5.

EDIT: I may have jumped the gun on this issue...

When I discovered that Windows was showing 6GB (3.99GB Usable), I began looking into it and found MANY stories of other users experiencing the same problems from their motherboards. So naturally I thought it was a motherboard issue. Tonight I tries booting with 2 sticks at a time, and in different slots. I tried a stick in DIMM 1 and 3, then used those same sticks in 1 and 5 and my PC recognized all 4GB. I tried the same except with my 3rd stick substituted in, and was able to dupe the problem. 4GB (2GB Usable). My motherboard problem just became a memory problem methinks...
 
Last edited:
6gb ram, 1.99 available
evga x58 sli original
patriot viper 1600
i7 920, i7 950

swapped out cpu, ram, psu, vid cards to no avail. clean install, no good. all reg fixes i could find - no good. flash bios updates and backdates - no good. all dimm slot combos - no good. cpu socket reseating thing - no good. tried new evga bios "fix" , which is the memory setting "trl", to all values - no good. tried with default settings (no oc, no voltage changes, etc) - no good.

swapped out mb to evga x58 sli3 - fixed.
 
I just added another 12gb of G.SKILL Ripjaws DDR3 DDR3 to my system and it says : 24.0 GB (only 16.0 GB usable). If I play with the BIOS it'll drop to 12, 8 or even 4 GB usable.

The rest of my system :

i7 920 (D0)
Prolima Megahalems with some AS5
EVGA E758-TR 3-Way SLI
24gb G.SKILL Ripjaws DDR3 DDR3 1600 (PC3 12800)
XFX Radeon HD 5870
Intel XM25 80gb SSD
WD Black 1tb (storage drive)
ENERMAX Liberty 620W
Windows 7 Ultimate - 64bit
 
happened to me wit hsame board, same cpu, different ram. swap out mb, unless you already have.
 
I just added a single chip at a time and mixxed them around and got my full 12 gigs istead of 8 gigs. Seems makke no sense why I would have to do this since sticks of ram were idenical make and speed. I have EVGA 3 way classified E-759. Try reseeding your cpu and ram since controller is on CPU for I7's. Try various ram configs till it works.
 
[Rant]
I'm so thrilled to swap RAM configurations, reseat CPUs, stare at the BIOS for hours, unscrew this and that, dominate all the coffee in the office, work past my office hours... just to get the damned thing to see 12GB of RAM consistently.

I can't even bear to bid a conscience-free good luck at customers when they buy a system with 6 slots and load it up with 3 sticks of RAM... because I will be praying every night that somewhere down the line they won't add another 3 sticks of RAM.

If the MB manufacturers can't replicate it, I can replicate it until the seas dry up... and so can a lot of people on their own support forums and this very forum as well.
[/Rant]
 
Asus P6X58D-E & 6GB Corsair XMS3 7-7-7-20

Installed: 6138MB
Shows: 4088MB

Oddly after a few reboots it showed fine, 3 days later and it cropped up again. When it shows 4GB the machine BSOD's under load, when it shows 6GB it does not. RMA is in progress.
RAM is fine, survived 4 hours of MemTesting.
 
I have a DFI UT X58-T3eH8 and a 920.

My first kit of Ripjaws would only detect 4gb out of 6gb on cold boots. I had to have the computer running for a bit, then restart and it would detect all 6gb. Reseating the CPU did not help.

2nd from RMA: only 4gb out of 6gb is detected when I moved my computer case. I reseated the CPU and from that point on all 6gb were detected. Never any problems since.

I think the CPUs are just picky about sitting in the socket properly. Maybe the memory controller is finicky.
 
EVGA 132-BL-E758 (identified via PC Wizard)
6X 2048 Corsair DDR3 1600 CM3X2G1600C8D (Set to factory specs - 8.8.8.24 - 1.65V - 2:12)
Intel i7 920 @ 2.66GHz

Computer started hard locking, so I did some routine maintenance (new AC5 tim, cleaning, etc.) put everything back together = 3.99 ram issue. Took it all apart AGAIN and reseated the cpu, RE-applied the AC5, put it all back together and running 6.00gb now. God I hate LGA sockets. Too easy to bend pins. I wish Intel would go back to a ZIF socket.....
 
Well I feel stupid.

After disassembling the entire rig to RMA the motherboard I found a small case screw lodged in between the pins on the back of the DIMM slot that was giving me trouble, seems to be running fine now.

Doh.
 
Well I feel stupid.

After disassembling the entire rig to RMA the motherboard I found a small case screw lodged in between the pins on the back of the DIMM slot that was giving me trouble, seems to be running fine now.

Doh.

Don't admit it! :eek:
 
ON 2010.09.22 @ 0145 AM I posted about inaccurate RAM count during POST. Since then a couple things have happened. First I replaced my 12G Black Dragon GB36GB1333C7TC PC3 10660 with 6G Kingston KHX2333C9D3T1K3/3GX and my inaccuracies during POST RAM count seaced. So I took the Black Dragon SIMMS and ran MemTest86 v4.10 on ea SIMM. This test showed one of my 6 SIMMS faulty. I RMA'D the Black Dragon, received new replacement and now am running 12G Black Dragon 1333 in a wide range OC from 2.8GHz @ 1333MHz C6 to 4.01GHz @ 1504MHz C7. So perhaps it was just a faulty SIMM all along. I am still operating with the GA-X58A-UD5 MoBo so the faulty RAM count during POST on the P6T (which I do not have any more) is still unexplained at this time for I also used some very good Patriot Viper on that MoBo and still got the inaccurate RAM count during POST. What I recommend from my experience is: first, use the MemTest86 (I just utilized it from a FDD) to check your memory and if that shows nothing amiss, commence swapping out your RAM with known good stuff. After all, what we are all speaking of here is "Easter egging" when it comes to HDWR. One other thing, and this is a error on my part during trouble shooting; I flashed the GA-X58A-UD5 BIOS before I utilized the KNST RAM. What I should have done was run the KNST stuff and flashed the BIOS afterwards. So at this point I cannot say one way or the other if the flashing of the BIOS had anything to do with the faulty MeM count during POST. -end-
irvinj in Virginia Beach, VA w.GA-X58A-UD5 LGA1366, i7 920 @ 4.01GHz, 12GB DDR3 1333C7 @ 1504MHz
 
Noticed my Sabertooth X58 just started showing 10 GB, after it posted 12 for a while. It has the latest BIOS. I have 6GB of Patriot and 6GB of G Skill, both with the same specs.
 
Just found this thread, but I've been having this issue for several months. Original EVGA X58 SLI. I have 2 triple channel sets of GSkill (6GB each) and with 6 GB it will read 3.99 GB usable, with 12 GB it will read either 3.99 or 7.99 GB usable. I've had a C0 i920 and a D0 i930 and the problem pops up with both. A hard CMOS reset (disconnect power, pull the battery for several minutes) will sometimes correct the problem for 1 boot, then it's back.

I also have an Asrock X58 with SATA3 and USB3 and it always correctly displays the amount of RAM.

A lot of people are running into this problem, and nobody from any of the mobo or RAM companies (or Intel) is providing any solid answers, just an endless, variable stream of "solutions" that might, maybe, intermittently work.
 
Back
Top