Cold boot failures x58A-ud3r Rev2 bios FB

[Dark Mantis]

I think from looking at your timing values it was purely a matter of data corruption because of the too fast closing of the bank. Your timing settings should be more like 8-8-8-24 although you might get away with tighter timings if it is good quality RAM

[FunkyErn]

The specs for the RAM at 1066 are 8-8-8-22, but could try 8-8-8-24. Does changing the RAS value from 20 to 22 or 24 fix these kind of BCCode 101 problems?

I am also running this at 1066 with 1.5v as per spec. I do not believe I need to up the memory voltage to 1.65v as I am not running at 1600.

Unfortunately I lost the timing values the tech guys at the supplier used due to the POST errors I was getting and having to clear the CMOS to get round that.

[Lsdmeasap]

BSOD 101 means Vcore is not set properly, this can be too much or not enough ( << generally this)

Recovering DRAM size means a channel was dropped, if you are seeing this on cold boots but not warm try increasing QPI/Vtt, or set Dram voltage to 1.66V if it's not already.

If you are underclocking the ram, as you mentioned 1066 instead of stock 1600, then 1.5 may be ok or may not be, you might need to try 1.55-1.6 as the memory was not designed to run that slow so 1.5V may or may not work.

[FunkyErn]

The actual message for BC Code 101 is "A clock interrupt was not received on a secondary processor within the allocated time interval". I do not believe this is restricted to vcore voltage being over or under. My latest research suggests there are three categories of issues that can cause this:
1. bios issues (wrong settings, inc. voltage settings)
2. hardware problems (overheating, defective components)
3. device drivers

Update to saga so far: got home last night, cold booted into MemTest86 and no memory errors! Logged on to Windows 7 and no BSODs! At this stage I have always been running bios FB.

However, one of my two graphics cards started intermittently going blank or showing an all blue screen. I had encountered this issue several weeks before when flashing the bios to FC1. So I swapped both card over to make sure they were seated correctly. And as expected the issue now switched over to the other monitor. As I had this issue one and by flashing to a different bios it stopped, I thought I would flash to bios FC3 to see if that would fix that. All went fine and with FC3, both graphics cards working fine. Run Sisoft burn in to stress test it - all OK, no BSOD.

First thing this morning I cold booted (with bios FC3) and ran MemTest86 and - lo and behold MEMORY ERRORS are back. They lasted for over 20 mins. I then gave up flashed back to bios FB, and no memory errors. All rather confusing. Too many variables.

This is new replacement memory which was tested as ok by both supplier tech support and me, with bios FB. As soon as I flashed to bios FC3 memory errors came back on cold boot. I am so sick of this - all I wanted to do was get on and use the new SONAR 8.5 music software on a decent machine and make music and here I am wasting my time with this...

The graphics card might be a red herring. So what I have done is flashed back to FB and will see what happens tomorrow morning after a overnight cold boot.

[Dark Mantis]

I think that it would still be a good idea to increase your memory voltage as this generally increases stability. If it is not a settings problem it is normally down to a bad chip on one of the modules that once warmed up will work reliably.

[FunkyErn]

Dark Mantis - the saga continues....

I have tried your suggestion and increased DRAM voltage to 1.64v and QPI/VTT to 1.2v (as per the XMP settings) and run at both 1066 and 1600 - with identical results. Cold boot MemTest86 errors consistently! I have tried all permutations of optimised and fail-safe defaults without any joy.

Either this second set of RAM is also bad, or there is a motherboard issue. I have emailed the supplier's tech support saying that the RMA did not address the problem and asking what we should do next.

I am considering purchasing Kingston KVR1066D3N7K3/6G which is on both Gigabyte QVL and Kingston memory configurator as supported. This is 1066 memory which is fine as I do not want to OC. Does anyone know if there have been issues with using KVR1066D3N7K3/6G with the GA-X58A-UD3R rev 2.0 motherboard?

I will try and RMA the motherboard once again. I am also seriously considering ditching Gigabyte and going for the MSI-X58A-GD65 instead. Judging by posting on various forums the ASUS P6X58D-E which is of comparable features and price also suffers from cold boot issues with memory.

[Dark Mantis]

I would say that if you are still getting cold boot issues and Memtest is recording errors thewn the memory is faulty. If the Kingston memory you mentioned is on both lists it should be fine.

[notshy]

Have you tried with the voltage at 1.66 and QPI/VTT at 1.355? I've just finished running through a guide to OC-ing and got my system stable after running through a lot of the settings. I have Corsair TR3X6G1600C8D RAM - I'm running BIOS FB and here are my settings - they are overclocked but stable. Also see that there is a link below to the Corsair recommended settings from their forum


BLCOCK ................................... 200
Load Line calibration .................... Level 2
Standard Vcore ........................... 1.25v
Dynamic Vore ............................. +0.05000v
QPI/VTT voltage  ......................... 1.355v
IOH Core ................................. 1.100v
DRAM Voltage ............................. 1.660v
Enabled EIST, C1E , other C States
CPU Multiplier = x19
SPD Multiplier = x8
Uncore = x16
QPI = x36
XMP Profile = Profile 1
RAM Set at 8-8-8-24 - everything else on Auto

I tested my system with CPU at x15 (Vcore=normal), and BLCOCK from 150 - 200 with RAM at x6 and x8 - it was stable once I'd increased the QPI/VTT to 1.355 (From the Corsair forum they tested the RAM with QPI/VTT of 1.35 to get stable - http://forum.corsair.com/v3/showthread.php?s=89e1dacb2680af46b8e7c0f945fafa54&t=81035)

Then to get a higher CPU, I did x21 CPU with Block from 150->180 and that's when I had to up Vcore to 1.25.

I'm sorry if this reads like an OC guide but there are a bunch of other BIOS settings that I'd switched from Auto to a manual setting and I thought they might help.

Feel free to ignore it if I'm talking rubbish - I'm new to OC and manual settings in BIOS.

[notshy]

QPI/VTT has to be within 0.5 of the RAM Voltage - so if you are upping the DRAM Voltage to 1.66 then you have to have at least QPI/VTT = 1.16v

But i've found my RAM stable up to 1600Mhz only with QPI/VTT up to 1.355 (I had it around 1.25/1.27 when I had Bclock on 133 - but upped to 1.355 when I upped the Blcock to 200).

I was told that you could run QPI/VTT up to about 1.45 but that is quite high - I've had no issues 1.355 with only a couple of case fans to cool my machine and Mobo.

[theboyfold]

notshy, did you have any issues booting from cold as well?

[notshy]

I had issues with some OCZ 2000 Mhz Gold Series (OCZ3G2000LV6GK) - they would give me issues with a cold boot but would pass 10 hours Memtest and 24 hours Prime95 when warm.

I've switched to 6GB CORSAIR Dominator DDR3-1600MHZ (TR3X6G1600C8D) and had no issues - I can post my full BIOS settings later on if people would find that useful. As I said earlier, I've been through an overclocking exercise this weekend and it's all working well on cold boots and long stress tests.

Main RAM settings are to use XMP Profile 1 - but manually set voltage to 1.66v, QPI/VTT to 1.355, set only the four timings at 8-8-8-24.

[Dark Mantis]

To be honest I think the Corsair Dominators take some beating on this platform.

[notshy]

I went for them as they were on the supported list, I'd heard good things and Corsair guaranteed that exact model to work with the motherboard - http://www.corsair.com/configurator/product_results.aspx?id=1362102

[FunkyErn]

Thanks for the info notshy!

I tried the QPI at 1.2 v which was within 0.5v of the DRAM 1.64v, so I do not think it is that. The thing that gets me is I am just trying to run at stock not OC speeds and this should not be so difficult.

It is interesting that with the OCZ3G2000LV6GK you had cold boot errors, but once warm it would run fine with MemTest. Exactly the same pattern with my Corsair TR3X6G1600C9. I see you are using the TR3X6G1600C8D, which is the Dominator while mine is XMS - that could be the difference. Switching make of RAM resolved the cold boot issue for you - hopefully it will do the same for me!

I spoke to the supplier tech support earlier and he said there are known issues with the GA-X58A-UD3R and TR3X6G1600C9 under certain conditions (which are not clearly understood yet) which both Corsair and Gigabyte are aware of and working to address. Generally this combo works fine. He thinks that the two components are probably fine individually, but just do not work well together under the current FB bios. He thinks this will be addressed in a future release of the BIOS. He did say that the ASUS P6X58D-E also had similar cold boot issues which have now been resolved by the latest BIOS release. These are their two best selling mobos.

So I went ahead and ordered the Kingston KVR1066D3N7K3/6G this afternoon. I also did a search on Google to see if there were any known issues and all looks good so far. Will let you all know the results in the next few days.

[theboyfold]

I went for them as they were on the supported list, I'd heard good things and Corsair guaranteed that exact model to work with the motherboard - http://www.corsair.com/configurator/product_results.aspx?id=1362102

The thing with that list is that I have tried the TR3X6G1333C9 in my motherboard, and it didn't work!

Stuff

Did you get my PM?