Solid Red DRAM Light & did the XMP profile timings just change?

All-a-Mort

Commendable
Oct 1, 2016
6
0
1,510
A few days ago I changed my RAM from 8GB Ripjaws X to 16GB (2x8GB) Trident X 1600CL7. Initially I had the Asus Red Light of Doom, but after using the MemOK! button and setting the XMP profile in the BIOS, the light went out. According to G.Skill, the BIOS and CPU-Z, it is meant to be running at 800mhz dual channel at 7-8-8-24-2N and it was without any issues.

Today I installed a new GPU (XFX RX480 GTR Black Edition) and in the process for some reason the Red DRAM light has now come on solidly and whilst the machine POSTs and seems to run fine, the XMP profile in CPU-Z is now stating that the timings under XMP profile 1.2 should be 7-9-9-25-2T, which is different from before. Pressing the MemOK! button is fruitless as the machine POSTs but I can't fathom what I have to do to solve the supposedly highlighted RAM issue. Equally annoying is that due to an issue with the RX480, I can't access the BIOS without changing back to the old graphics card, which doesn't exactly help matters.

i5-2500K @ stock (I picked a dud)
Asus P8P67 Pro B3 Rev1
G.Skill TridentX 16GB 1600Mhz DDR3
XFX RX480 GTR Black Edition 8GB DDR5
2x Samsung HD103SJ 1TB HDD
XFX 750W Black Edition PSU
 
1| Can you please pass on the SKU's to both your ram kits?
2| You forgot to include your OS.
3| Have you made sure your BIOS is up to date?

It's best if you could resolve your ram issue and then get along with installing the new GPU once troubleshooting has finished.
 

SKU of the old RAM
F3-12800CL8D-8GBXM

SKU of the new RAM:
F3-1600C7D-16GTX

According to G.Skill both of these are on the QVL list for my motherboard


Whoops. I'm still running Windows 7 (Ultimate) SP1 64bit. It'd be ironic if after my efforts to stop Microsoft automatically installing Windows 10 on this machine, I end up having to do it anyway, whilst paying for it.


Bios for my P8P67 Pro Rev 3.1 is 3207, dating from November 2012. To my knowledge that is last Bios update ever issued for that motherboard. As such it is 'upto date', but still pretty ancient.


That's the thing though: the RAM issue appeared resolved at that point, hence I decided to install the GPU.

Ok, yesterday I tried out the F3-1600C7D-16GTX modules in each of the slots to see what happened. Bear in mind that the well known black screen issue with the GPU makes it difficult to be certain of boot status. I've assumed failure to Boot, when the windows welcome screen is not present after a 5 min wait and the front panel power switch instantly turns the machine off, which it wouldn't do otherwise.

Slot A1: inaccessible due to the CPU cooler fan (a push n pull set up).
Slot A2: Module 1 = POST, boot into Windows but red light on
Module 2 = POST, no boot, red light on.
Slot B1: Module 1 = POST, no boot, red light on
Module 2 = POST, no boot, red light on but very faint.
Slot B2: Module 1= POST, no boot, no red light
Module 2 = POST, no boot, no red light


Now according to the motherboard manual, for a single stick, only A2 should allow it to boot. But that only happened with module 1 and the light was on. Conversely, despite not booting using only B2, the light did not go on. Not sure what to make of that.

It seemed to me that the DRAM light when it lit got fainter as I ran the test. At the end when I put the modules back in the normal A2 and B2 slots, it booted fine with no light, only for it to come on again very dimly later. At POST it is a very bright LED, but instead after boot it was like a pinprick of light. At the moment today, the light is not on at all. It makes me wonder if the light itself is faulty.

If I have time tomorrow I'll repeat the tests using the old GPU just to be able to see if anything shows up where the screen is black with the RX480 connected.

 


Old card is a MSI Radeon HD6950 2GB Non-Frozr model (though it was running a 6970 bios).
The PSU dates from 2011, as does the rest of the build but for a few case fans I added later.
BIOS (as stated above) is 3207, the last one released for the P8P67 Pro Rev 3.1 (AFAIK).
 

That is a possibility. I had one odd boot failure recently, though that was after a Microsoft update and they have er, previous in causing wobbles. I also seem to have a case fan connected to CHA_FAN2 that might have issues with power supply as it was reporting that it wasn't spinning, which I thought my mean it was blocked by a stray cable, but whilst I way away from the machine, the PSU reported an issue of undervoltage on +3.3v, reading 1.984v briefly. Not seeing any other suggestions of errors though in the logs the ASUS board creates. I might need to try running something more taxing to see if any errors crop up when the GPU etc are drawing more power.
I'd have to buy a new one rather than borrow one, so I'd rather like to be a little more certain, unless it's worth replacing it anyway after five years of service.
 


Ah. It seems my board is not a Rev 3 in fact, but the very slightly earlier B3 Rev1. I should have paid more attention when I built this thing way back when. As such, the 3602 Bios was rejected as not matching the same model. So, 3207 is what I'm stuck with it seems.

Just read in another thread from a few years ago someone commenting on the fact that whilst gaming, when their GPU ramps up, they were seeing their network connection fail and wondering if that was a sign of a faulty PSU in that it was struggling with the power. That happens most times I load up a game (XCOM2 or Elite Dangerous at the moment) so I'm curious about that, though as I said, there's nothing in the ASUS logs to show anything untoward on power delivery.