CMZ4GX3M1A1600C9 and Gigabyte z87x-UD3H

Hello,

New system, with the ram from my "old system" :)

 

I have two sticks of CMZ4GX3M1A1600C9 (4gb Vengeance 1600mhz)

CPU is the i7 4771, which is not overclocked. Everything in this system is completely stock. GA-Z87X-UD3H board, TX750 PSU

 

Ram is set to XMP mode to get 1600mhz speed. Nothing else is touched, it's all stock speeds.

 

Ran Memtest overnight. After 4 passes everything looked great, had 0 errors. Came back when it was up to pass 6, test 7 - and it started throwing errors (64 total)

 

Is this a definite RMA situation? Or is there another setting I possibly have wrong? (Voltage fine or should it be forced to something higher??) I have not tested the DIMMs individually yet. Thanks for any input.

SPD lists a setting "tRC" with a value of 41. Can't find this in the bios anywhere?

Sean all you need to do is set the first four values and leave everything else on auto.

i set the first four (9-9-9-24), 2T command rate, plus tRFC @ 128. Just can't find the tRC in my bios to ensure it's set to 41. I want to match all of the returned SPD settings. ;):

At first I was lost as to where you got the 128 value from, then I realized your looking at the wrong table.

 

Look at the SPD tab in CPUz. NOT the MEMORY tab. Set the first 4 values under XMP and leave everything else on auto.

 

The MEMORY tab shows you what your memory is currently set to. Actually,Your memory is running at rated speed right where it should be .

i still got errors on the same test 7, after three passes. I even had the voltage bumped up a tad to 1.525

Try at 1.55v and you can go to 1.65v if you need to . but only add what you need to to get it stable.

Well, I've come to the conclusion that the latest MEMTEST 5.01 has some bugs with test number 7 giving false positives. And it's possibly only in UEFI boards.

 

Brand new ram purchased Friday, got here Saturday. (CMY16GX3M2A1600C9) also fails on the same test 7 after a few passes.

 

Neither ram fails in the new Z87X-UD3H with test 4.10, 4.20, or 4.30 versions.

 

"OLD" ram (the ram in question) runs fine (using the same options) back in the old socket 1156/i7 870 machine after 8 passes of test number 7 with memtest 5.01.

 

New system is rock solid stable.

 

Long story short, bought some ram I didn't really need (would have rather bought another SSD drive!) - oh well.

I'm happy to read your post, because i have exactly the same problem with memtest 5.01 in test #7 on my brand new Gigabyte G1.SNIPER Z87 and my "old" 4 modules CMX8GX3M2A1600C9. Those modules always worked correctly in my old system.

 

I'm testing for 2 days now with memtest86+ and it's making me crazy. I always have errors in test #7 - even with a single module. I played arround with all kind of voltage, speed and clock settings. After a short or longer period the errors appear in test #7.

 

I tried memtest (without 86+) and there i had no trouble.

 

And finally i observed that it happens when i used the F2 - SMP option in memtest86+.

 

Random BSOD at the beginning and this #7 errors, gave me a bad feeling and just like you i was considering buying new RAM.

I have come to the conclusion that this board of mine simply doesn't like running the ram at 1600mhz. It's very stable at 1600, but I do get looping sound freezes in HL2 based games. If i set the ram back to default 1333mhz, then I no longer get any looping sound freezes in HL2 games.

Thanks for your feedback.

 

I did extensive tests for two weeks now :

 

- Memtest86+ 5.01 with SMP option enabled (all cores) :

4x4gb (CMX8GX3M2A1600C9) @ 1600-9-9-9 (XMP) = errors

3x4gb (CMX8GX3M2A1600C9) @ 1600-9-9-9 (XMP) = errors

2x4gb (CMX8GX3M2A1600C9) @ 1600-9-9-9 (XMP) = errors

1x4gb (CMX8GX3M2A1600C9) @ 1600-9-9-9 (XMP) = errors (each module tested separately)

2x2gb memory from competitor @ 1600-9-9-9 (XMP) = errors

1x2gb memory from competitor @ 1600-9-9-9 (XMP) = errors

 

4x4gb (CMX8GX3M2A1600C9) @ 1600-10-10-10 - 1.50v = errors

4x4gb (CMX8GX3M2A1600C9) @ 1600-10-10-10 - 1.55v = errors

4x4gb (CMX8GX3M2A1600C9) @ 1600-10-10-10 - 1.60v = errors

 

4x4gb (CMX8GX3M2A1600C9) @ 1600-11-11-11 - 1.55v = errors

4x4gb (CMX8GX3M2A1600C9) @ 1333-9-9-9 - 1.50v = errors

4x4gb (CMX8GX3M2A1600C9) @ 1033-9-9-9 - 1.50v = OK

 

- Memtest86+ 5.01 with SMP option disabled (one core):

4x4gb (CMX8GX3M2A1600C9) @ 1600-9-9-9 (XMP) = OK

 

- Passmark Memtest86 5.01 with SMP option enabled (all cores) :

4x4gb (CMX8GX3M2A1600C9) @ 1600-10-10-10 - 1.55v = OK

4x4gb (CMX8GX3M2A1600C9) @ 1600-11-11-11 - 1.55v = OK

 

The biggest difference i observe on the RAM modules themselves is the temperature. On tests with a single core they are 10°C colder than with SMP enabled.

 

In Gigabyte forums i read that problems often occur on systems with 4 ram modules.

 

Because of not exactly knowing if memtest86+ (recommended here) has a bug or not, i'm using the last settings of 1600-11-11-11 - 1.55v. Passmarks Memtest86 showed no errors and until now my system didn't crash.

My PC stays on closed observation for the moment. Using the Ram @ 1333 or less is not a option for me at the moment.

The biggest difference i observe on the RAM modules themselves is the temperature. On tests with a single core they are 10°C colder than with SMP enabled.

And thats normal. XMP raises some system voltages which in return generates a little more heat. ZGood cooling is a must.

 

In Gigabyte forums i read that problems often occur on systems with 4 ram modules.

That holds true for just about any board. 4 modules is putting a larger strain on the mem ory controller. So to compensate you need to raise the DIMM voltage.

 

If your system is stable at 1.55v then you should be good to go.

Jack - I installed a beta bios last night for my board (8fm I think), and so far it looks as if my problem has been fixed. It's only been a day but it looks promising. Did several dozen passes of just test 7 with SMP enabled (all 8 logical cores) and no failures. This is with the ram set to XMP profile, 1600mhz. I'll report back after a few more days as soon as I see that I don't get any looping sound freezes in my HL2 based games.

I have had issues with BSOD issues intermitantly on my system. I called Gigabyte and they had me go back into the bios and reset the XMP profile to profile 1, where it was already and with everything set at auto. The only difference is that they had me do it in the old style panel instead of the pretty HD panel. It worked for 14 days and yesterday I had one. They are always Kernal Power errors with a bugcheck code of 292. Everything I find relating to this is either Vcore settings or timings or possibly CPU or PSU. I checked both memory modules and they pass. Please see my system specs. I have a I7-4770K, 16gb Dominator Platinum 1600, RM850 PSU and GA-760GTX-OC-4GB. This usually happens after I leave the computer and come back to see Windows rebooted to due to the Kernal Power error. It has never happened while using the computer. I have Bios version F7 since I built it in November. Where did you fing version 8 of the bios. I looked at Gigabytes website and can't find it. System runs great typically and doesn't get hot. Any thoughts?

I have come to the conclusion that this board of mine simply doesn't like running the ram at 1600mhz. It's very stable at 1600, but I do get looping sound freezes in HL2 based games. If i set the ram back to default 1333mhz, then I no longer get any looping sound freezes in HL2 games.

 

Just an update. I continued to have looping sound freezes in CS:GO and DOD:S, but they did decrease quite a bit with an update to firmware F8.

 

The looping sound freezes FINALLY stopped alltogether - when I replaced the ram with G.Skill Trident X running at 2400mhz. I set this ram to the XMP profile 1, and everything is snappy fast and best of all, ROCK SOLID. Neither CS:GO or DOD:S have had the looping sound freeze.

Just another update:

 

I replaced the Z87X motherboard with the GA-Z97X-UD5H-BK board. I just ran 12 passes of Memtest test #7 with the Corsair ram without any failures (previously this would fail). I could never do this on the Z87 board. If I can play HL2 games without any looping freezing - I will place the blame for everything on that Z87X board.