Asus P7P55D + Ripjaw DDR3-2000 (F3-16000CL9D)

[gpb500]

Hey Bishop...you might try when you do your next cold boot, to boot from the memtest disk and skip to "test 5" (i assume that's where you're seeing errors)...if you get errors, power off and back on (do not do a reset). After the computer is off for 5 to 10 secs, restart and run "test 5" again.

To skip to "test 5"...once memtest starts, press "c" (for configure) and then options 1, 3, 5 and 0. If you get no errors you will have recreated the problem I have. Now having said that, this morning using only my 2nd pair of ram, I had no errors on the cold boot and have been stable so far (if i have no memtest errors, then i'm good all day) which seems to point directly at the memory. I'm waiting for the experts to chime in. Hope that helps...helps me think I'm not crazy.

[GSKILL TECH]

gpb500, We can definitely have that set replaced and it should be much better. Once we receive your defective set, you should receive new replacements within 4-7 business days. Unforunately we do not cross-ship. NewEgg has a 30 day exchange/return policy so if you are still within that period simply contact them for an exchange.

bishop,
Keep your voltages on AUTO and see if that will give you different results. They are set too high at the moment.

Thank you
GSKILL SUPPORT

[bishop]

Thank you for the reply. I am confused about the the voltages being too high. In this previous thread, you said that the settings should be:

AI Overclock Tuner - Manual
CPU Ratio Setting - 18.0 for 2.80GHz (can set to 20 for a little CPU overclock)
BCLK Frequency - 160
DRAM Frequency - DDR3-1600MHz

Fixed Voltage - 1.30V
IMC Voltage - 1.20V
DRAM Voltage - 1.65V
CPU PLL Voltage - 1.90V

DRAM Timing Control (submenu)
CAS Latency - 7
RAS to CAS Delay - 7
RAS PRE Time - 7
RAS ACT Time - 24

DRAM Timing Mode - 2N
All others can be AUTO

That person has the same setup as me.

I have tried auto (via XMP profile) and the blue screens persist.

I appreciate any help...

[elysoto]

Hello, I'm having the exact same problem as the the thread starter. Just about every exact detail.

I have a Gigabyte GA-p55-UD4P motherboard with

8GB of F3-16000CL9D-4GBRH

Cold boot causes memory problems and memory errors in MemTest86+.

I found a very intersting thread in another memory vendor's forums.
Seems that they discovered that the 'Round Trip Latency' value was changing
across cold boots when in Auto mode. I'm going to try forcing it to a static value and see what results I get.

http://forum.corsair.com/v3/showthre...t=82480&page=1

I was just about to RMA my memory to NewEgg but now I have something else to try.

[bishop]

Quick question to anyone who can answer it (hoping G.Skill Tech Support can provide an answer here):

If we are running XMP at 1600 (BLCK 160), do we have to disable C1E and Intel SpeedStep? I've read in some support forums that when these features are on while overclocking the ram to run at the 1600 spec, that the voltages could decrease on idle causing the BSOD and memory errors? What is G.Skill's recommendation on these two settings? I see that ASUS' most recent bios is trying to address this:

P7P55D 0711 BIOS
1. Disable EIST and CIE functions if BCLK is higher than 150MHz

Thanks,

Bishop

[GSKILL TECH]

bishop,

Sorry about the confusion, the voltages recommended were for the 3.2GHz overclock but I forgot to mention that. If processor speed is default, then AUTO voltages are fine.

C1E and SpeedStep should definitely be disabled. They are good ideas for normal users to save power, but the minimal power save is not worth the errors. With those option enabled, the computer jumps frequencies, thus causing power to fluctuate, and basically everything thus causing instability. It is simply unnecessary especially for people that are looking for performance in their machines.

Thank you
GSKILL SUPPORT

[bishop]

Many thanks for the reply and clarification. I will try the XMP profile with C1E and Intel SpeedStep disabled when I get home and let you know how I make out...

[gpb500]

Hey Bishop, another person that mentioned they too were having the same cold boot problem pointed me to this thread. Sounds like they are experiencing the same as us. As it turns out, the motherboard/bios is not keeping the Roundtrip Latency setting constant. I checked with the G.Skill Tech on this and he advised 47+. My stable settings were defaulting to 51-53 (when set to auto). I tested 49-51 with my 8GB and I had 2 errors. Changed it to 50-52 (NORMAL option) and I seem to be golden. Need to still test the whole cold boot tomorrow morning but this kind of behavior certainly explains why we're stable sometimes and not others.

You might jot down all your "auto" settings and when you go unstable, compare them to see which (if any) changed. That's my plan. In fact, I'll post mine here for our board. Note we are not using the same RAM...so your settings will differ.

1st: 9-9-9-24-6-109-15-10-30 (all auto)
2nd: 2N-50-52 (all set manually 2N-Normal-Normal)
3rd: 6-6-20-10-10-10-7-6-4-7-7-4 (all auto)
vMem: 1.575
SpeedStep: Enabled
DDR3-2000 @ 1960 (underclocked here slightly to keep my cpu at 3.53)
CPU 18x (3.53 GHz)
BCLK 196

Maybe this will help someone else too...but gotta confirm the cold boot tomorrow morning.

[bishop]

Many thanks for this detailed information. I will make the bio setting change to Roundtrip Latency and see how it goes with a cold boot tomorrow after work. I tried a cold boot when I got home today and received errors. I just updated my bios to the latest [P7P55D-ASUS-0711] (be warned the automatic asus update utility does not update to the latest - you must get it from their Web site and manually flash the bios). The new bios will disable some settings by default when you are clocking memory 1600+. In my case, it disabled C1E and SpeedStep....for the time being I have been running stable without any errors in memtest at 1600 7-7-7-24. Again, the test tomorrow with these changes I made and the one you recommended to Roundtrip latency will prove whether this is fixed for good. Thank you for your time in helping me with this. I hope it works for you tomorrow!

[gpb500]

I just updated my bios to the latest [P7P55D-ASUS-0711] (be warned the automatic asus update utility does not update to the latest - you must get it from their Web site and manually flash the bios). The new bios will disable some settings by default when you are clocking memory 1600+. In my case, it disabled C1E and SpeedStep....

A word about BIOS updates. I would recommend NOT using the update utility from windows, but instead update the bios from within the bios using the EZ Flash (look under Tools in the BIOS menu). You might also dial the O/C down just in case.

You should be able to re-enable SpeedStep unless you are doing massive overclocking. Good luck with the cold boot. I hope this fixes the problems.

[elysoto]

Arg, fixing the 'Round Trip Delay' to a static value didn't work for me.

Back to square one.

[gpb500]

Nope...didn't work for me either. Memory packaged for RMA.

[bishop]

Didn't work for me as well...I will be RMAing this back to the retailer I bought it from for an exchange. I am hoping that I don't have this problem with the 2nd pair...This is my first time trying G.Skill and would be very disappointed. I will keep you updated.

[gpb500]

Didn't work for me as well...I will be RMAing this back to the retailer I bought it from for an exchange. I am hoping that I don't have this problem with the 2nd pair...This is my first time trying G.Skill and would be very disappointed. I will keep you updated.

Yeah the whole RMA process is a pain. Newegg makes it easier, and I've been lucky having only one RMA in the past. I used 8GB of G.Skill on my prior Asus P5W DH Deluxe board and it worked great. The tech here seems very good...I think it should work out. I'll report back when I get my replacement. Good luck.

PS - I'd still like to know what's up with this cold boot thing.

[GSKILL TECH]

Very weird, we have this board here and never encountered that issue. Will contact ASUS to see if they have any idea.

Thank you
GSKILL SUPPORT