Question Upgraded GPU, now the DRAM LED keeps coming on ?

[Joe Soap]

Good Day To All,

So to start off, here is the current PC Build:
CPU: 2600k

MOBO: ASUS P8Z68-V

PSU: Cooler Master Elite V3 P600

SSD: SAMSUNG 870 EVO 1TB

RAM: G.Skill F3-1866C10D-16GSR Sniper 16GB (2x8GB) DDR3-1866MHz (HERE IS THE PROBLEM)

GPU: GeForce® GTX 1080 G1 Gaming 8G


I very recently got a Gigabyte Aorus 1080Ti Extreme 11GB , so I swopped this with my Gigabyte GTX 1080 G1. Switched on the machine, booted up no problem.

Loaded all the drivers required. Shutdown the PC after I was happy with setting up the new card. Next morning I try to switch the machine on again, it hits me with a DRAM LED out of nowhere after just working the previous day.

So luckily I had old Corsair Vengeance 2x4Gb dual channel kit lying around which I then swopped the G.Skill with to test. At this point I have also removed the new GPU again. So now the machine boots up to a screen that says Overclocking Failed after I have not made any overclocking attempt with anything?

I then go to BIOS and restore to safe defaults and machine boots back up into windows. I then switch off and retry the G.Skill, hits me with DRAM LED again. Switch back Corsair Ram get overclocking failed. Same process tried and repeated couple of times.

I then took a chance with putting the new GPU in with the Corsair Vengeance Ram, boots up fine and can log into windows. I check everything out and all looks good. Then switched off the pc again via normal shutdown. Here comes the funny part, I switch on the machine the next day and it hits me with the DRAM LED on the Corsair Ram too?

So now I ask if anyone has had anything similar happen to them with such a board or any suggestions, tips or advice please?

 

[punkncat]

You don't mention your PSU. What is it and how old?

If you were running an XMP profile it generally requires a little more power. If I were going to take a guess, I would think that the power supply in this system is aging and couldn't supply the new card and the power required for the VRM/XMP load.

I could be wrong.

 

[Joe Soap]

You don't mention your PSU. What is it and how old?

If you were running an XMP profile it generally requires a little more power. If I were going to take a guess, I would think that the power supply in this system is aging and couldn't supply the new card and the power required for the VRM/XMP load.

I could be wrong.

Howzit,

Thanks for chipping in, PSU was list in my specs at the top as Cooler Master Elite V3 P600 @ 600 watts

Bought it new along with the G.Skill kit, December 2022.

I was also wondering about the XMP profile of the Rams not perhaps messing about, but how could that even be linked between the gpu and ram😑

I have sent both ram kits do go through memory testing and see what they come up with. Was also wondering if the BIOS isn't freaking out with that specific GPU and RAM combination?

 

[punkncat]

Not sure how I missed that. I even looked again...preoccupied, I guess.

Linked only by power delivery and demand. The motherboard is using more power with an XMP profile set and the ti model of the 1080 is certainly drawing more than the vanilla version. I hope you can get it worked out as that is still a very relevant graphics card.

 

[DSzymborski]

Howzit,

Thanks for chipping in, PSU was list in my specs at the top as Cooler Master Elite V3 P600 @ 600 watts

Well, that's a pretty horrifying sign to start with. This PSU is very cheaply made, group-regulated PSU and had no business ever running with the old GPU, let alone the slightly more powerful one. I might put that PSU into my dad's PC, but only because my dad died in 1997, before discrete gaming GPUs took off and CPUs weren't yet based on +12V power. Your problem may not be caused by this PSU, but it's a priority to fix this; you've been very fortunate it hasn't fried your parts.

 

[Joe Soap]

Not sure how I missed that. I even looked again...preoccupied, I guess.

Linked only by power delivery and demand. The motherboard is using more power with an XMP profile set and the ti model of the 1080 is certainly drawing more than the vanilla version. I hope you can get it worked out as that is still a very relevant graphics card.

No problem at all haha,
But I do believe the PSU is capable of handling the 1080Ti? Unless it is borderline and that makes the RAM freak...

Also to note the BIOS on this motherboard is still the OG version it came out with, so no updates on that yet. Opinions in whether I update this that it might "resolve" this problem from occurring?

Not sure how I missed that. I even looked again...preoccupied, I guess.

Linked only by power delivery and demand. The motherboard is using more power with an XMP profile set and the ti model of the 1080 is certainly drawing more than the vanilla version. I hope you can get it worked out as that is still a very relevant graphics card.

 

[Joe Soap]

Well, that's a pretty horrifying sign to start with. This PSU is very cheaply made, group-regulated PSU and had no business ever running with the old GPU, let alone the slightly more powerful one. I might put that PSU into my dad's PC, but only because my dad died in 1997, before discrete gaming GPUs took off and CPUs weren't yet based on +12V power. Your problem may not be caused by this PSU, but it's a priority to fix this; you've been very fortunate it hasn't fried your parts.

Meaning that you think the PSU is perhaps messing with the RAM and is not suitable for this setup?

I can say that it did run the 1080 G1 for a solid 8 Months and all the other components remained the same..

 

[DSzymborski]

Meaning that you think the PSU is perhaps messing with the RAM and is not suitable for this setup?

I can say that it did run the 1080 G1 for a solid 8 Months and all the other components remained the same..

No, meaning the PSU is a giant, horrifying issue, even if everything were visibly working. If an appropriate PSU is used and you still have problems, that just means you had two (or more) problems rather than one.

You're gambling here. Using junk PSUs are the PC equivalent of driving drunk. Most people driving drunk get home safely...until the day they might not.

This isn't even a decent budget PSU or even a decent entry-level one. There's a reason it's in the bottom of the formerly curated PSU list here at "Tier E: Potentially dangerous in multiple scenarios." Cooler Master actually has some competent PSUs on the higher-end these days, something which wasn't always the case, but dirt-cheap Cooler Master (or Thermaltake) PSUs are some of the nastiest PC components ever sold by a mainstream PC component company.

 

[Joe Soap]

No, meaning the PSU is a giant, horrifying issue, even if everything were visibly working. If an appropriate PSU is used and you still have problems, that just means you had two (or more) problems rather than one.

You're gambling here. Using junk PSUs are the PC equivalent of driving drunk. Most people driving drunk get home safely...until the day they might not.

This isn't even a decent budget PSU or even a decent entry-level one. There's a reason it's in the bottom of the formerly curated PSU list here at "Tier E: Potentially dangerous in multiple scenarios." Cooler Master actually has some competent PSUs on the higher-end these days, something which wasn't always the case, but dirt-cheap Cooler Master (or Thermaltake) PSUs are some of the nastiest PC components ever sold by a mainstream PC component company.

Ah perfect okay now that makes way more sense, in summary the PSU is a major problem in this system and should be replaced👍🏻

And I was considering getting a more decent PSU Unit before doing the 1080Ti upgrade.

Will definitely look at a improved unit, thanks for all the feedback and info man, much much appreciated.

Do you think the mobo needs the BIOS updated ?

 

[Joe Soap]

Just another update on this situation.

So the Corsair Vengeance and G.Skill kits got tested on another mobo and everything booted up fine. Switched off the machine and tried switching on again, now got the same DRAM won't boot issue on this mobo as well.

Why would the Ram be able to boot up once on the mobo and then never again after that?

 

[Eximo]

When you installed the GPU did you jostle the CPU cooler around at all? Possible you need to re-seat the CPU in the socket. Probably due for a repaste anyway.

1080Ti may be drawing more power through the slot than the 1080 which is causing a power deficit in your motherboard, causing the memory to have issues starting.

I agree with the power supply advice as well. I would not want to run a 1080 or 1080Ti off that PSU. Even though by today's standards the 1080 doesn't use a lot of power, the 1080Ti certainly pushes well into the higher power requirements.

 

[Joe Soap]

When you installed the GPU did you jostle the CPU cooler around at all? Possible you need to re-seat the CPU in the socket. Probably due for a repaste anyway.

1080Ti may be drawing more power through the slot than the 1080 which is causing a power deficit in your motherboard, causing the memory to have issues starting.

I agree with the power supply advice as well. I would not want to run a 1080 or 1080Ti off that PSU. Even though by today's standards the 1080 doesn't use a lot of power, the 1080Ti certainly pushes well into the higher power requirements.

I tried to be as careful as possible not to knock anything else with the installation of the card, but anything could happen. So I will give the cpu reseat a try as well.

I am in agreement that the 1080ti would be pulling more power because of it's physical build etc. And I would like to believe now that the crappy PSU I have running in that case is a huge problem... I just need to understand how come both ram sets were able to boot safely into windows once each with the 1080ti?

Also these ram kits booted into windows once safely each on a complete other motherboard, cpu etc. then also couldn switch on a second time on this board...

 

[Eximo]

Could be a coincidental failure, it is a fairly old system.

Try clearing the CMOS after you put it back together, might just have a lingering setting.

May also be a good time to see if there are any BIOS updates available.

 

[Joe Soap]

When you installed the GPU did you jostle the CPU cooler around at all? Possible you need to re-seat the CPU in the socket. Probably due for a repaste anyway.

1080Ti may be drawing more power through the slot than the 1080 which is causing a power deficit in your motherboard, causing the memory to have issues starting.

I agree with the power supply advice as well. I would not want to run a 1080 or 1080Ti off that PSU. Even though by today's standards the 1080 doesn't use a lot of power, the 1080Ti certainly pushes well into the higher power requirements.

Could be a coincidental failure, it is a fairly old system.

Try clearing the CMOS after you put it back together, might just have a lingering setting.

May also be a good time to see if there are any BIOS updates available.

It must be very coincidental because it affected two dual channel sets, but not in such a way that the ram boots up once on a other system.

The cmos battery I had left out overnight with the clear cmos jumper set and no luck there yet.

The BIOS of this board is still the first version with which it came out...so I am considering updates on it as well now

 

[Joe Soap]

Good Day Everyone,

I trust all is going well.

So I have managed to find the solution to my problem, which in a Nutshell was that the RGB Fusion 2.0 Software for the Gigabyte Aorus 1080Ti Extreme 11GB was "killing" the DDR3 Ram Sticks. The software was overwriting the RAM SPD profiles, rendering them unusable. Luckily with perseverance I came across this useful and very helpful link that saved all my RAM and got them working again on linustechtips.

RGB FUSION 2.0 killing DDR3 RAM

Kindest Regards to All and hope this helps someone else out.

 

[Joe Soap]

@DSzymborski @Colif if you could please flag the above as the best answer/solution, that would be appreciated thanks 🙏

 

[Colif]

its not normal for op to be given the award but I will this time since you found answer and it might be handy.

I know rgb fusion has issues with ram, I have it installed but it doesn't run my ram... it actually has problems seeing it so I let Icue handle it.

Not heard of rgb software killing ram before.

 

[Joe Soap]

its not normal for op to be given the award but I will this time since you found answer and it might be handy.

I know rgb fusion has issues with ram, I have it installed but it doesn't run my ram... it actually has problems seeing it so I let Icue handle it.

Not heard of rgb software killing ram before.

Thank you very much for awarding the solution. Yeah so the ram wasn't being destroyed completely, it was more being corrupted by the RGB Fusion Software, specifically the Ram SPD was being written to. So whenever I switched the PC on again after a boot with Ram that worked before, it would make the BIOS go into a "overclock failed" boot error. And somehow I managed to come across the link and use the indicated software, RWEverything and Thaiphoon Burner to fix up the RAM again by correcting the RAM SPD's.

 

[Colif]

I am surprised gigabyte software can change values in an Asus bios. I could understand if you had a Gigabyte board but you don't. I only use it as I have a Gigabyte motherboard and it runs the rgb fans on my AIO

 

[Joe Soap]

I am surprised gigabyte software can change values in an Asus bios. I could understand if you had a Gigabyte board but you don't. I only use it as I have a Gigabyte motherboard and it runs the rgb fans on my AIO

I am surprised gigabyte software can change values in an Asus bios. I could understand if you had a Gigabyte board but you don't. I only use it as I have a Gigabyte motherboard and it runs the rgb fans on my AIO

It is more that it is able to change or rewrite the SPD values of the RAM. Which also causes a CRC check fail. Apparently the RGB Fusion Software has had some sketchy things that it does in the background from articles I have come across. Just know that I won't be using RGB Fusion Software anytime soon.