I'm frustrated... please help

[27MaD]

ok guys i've just did the following:

put the 1650 back in
wiped the dust of the pci of the card
turned off "3dmark01 enhancement" option in BIOS
did loosen the timings a little bit i changed them from 16-15-17-36 to 17-15-17-37 (for some reason BIOS applied it to C18 not C17 as i entered)

idk but things seem to be better just launched some games and things seem good, i'm still not sure imma have to test it even further later.

 

[TheoneandonlyMrK]

Corsair Vegetable's omg I'm dieing

Damn phone.

 

[80-watt Hamster]

Damn phone.

I vote we refer to them as Corsair Vegetables from now on.

 

[27MaD]

I vote we refer to them as Corsair Vegetables from now on.

As an owner of the kit, i also strongly vote.

ok guys i've just did the following:

put the 1650 back in
wiped the dust of the pci of the card
turned off "3dmark01 enhancement" option in BIOS
did loosen the timings a little bit i changed them from 16-15-17-36 to 17-15-17-37 (for some reason BIOS applied it to C18 not C17 as i entered)

idk but things seem to be better just launched some games and things seem good, i'm still not sure imma have to test it even further later.

Problem SOLVED!

as expected, it wasn't RAMs.

i guess turning that 3d mark01 enhancement thing in the BIOS did the trick

Thank y'all guys so much for helping i rly appreciate it.

 

[27MaD]

Ok guys same issue is back after i updated my driver from 466.11 to 466.27, currently downloading the previous driver and'll see how things go.

Ok guys same issue is back after i updated my driver from 466.11 to 466.27, currently downloading the previous driver and'll see how things go.

Just reinstalled the old driver still crashed, i remembered i had installed a windows cumulative update i had received earlier this day and just uninstalled it and it seems that things got back to where they should be i'm not pretty sure yet gotta test it even more.

if it's true then wow, windows updates are still causing issues like that after almost 6 years, and i thought i was the only one who didn't face any problems with it.

 

[Hemmingstamp]

Just reinstalled the old driver still crashed, i remembered i had installed a windows cumulative update i had received earlier this day and just uninstalled it and it seems that things got back to where they should be i'm not pretty sure yet gotta test it even more.

if it's true then wow, windows updates are still causing issues like that after almost 6 years, and i thought i was the only one who didn't face any problems with it.

It is true. Just keep it as it is until MS fix the update issue. That might be a long wait though.

 

[TheoneandonlyMrK]

Have you put your memory timing back to the tighter normal you originally ran.

 

[27MaD]

Ok cool we're now still crashing even after rolling back drivers and uninstalling the last windows update, as always just more and more disappointment.

Have you put your memory timing back to the tighter normal you originally ran.

Yes i did

 

[phanbuey]

Get a crappy non-corsair kit of ram, run that kit. Even something cheap.

Or downlock, overvolt and loosen the timings on your ram. Your memory controller HATES that ram. I had a ryzen 1700 on a b350 that would crash once a week (bone stock everything -- was a work devbox) with that ram regardless of what we tried.

 

[Deleted member 205776]

as expected, it wasn't RAMs.

I'm still not convinced. Have you ran 1 pass of MemTest86? I usually get errors in the first 10 mins.

 

[phanbuey]

Try taking a stick out and run it single channel for science. If it doesn't crash you know that's what it is.

Also have you looked at your event viewer for WHEA errors / log errors?

Nvidia drivers uninstalling and issues coming back out of sleep happens to me with unstable cache / ram.

 

[Hemmingstamp]

Ok cool we're now still crashing even after rolling back drivers and uninstalling the last windows update, as always just more and more disappointment.


Yes i did

Do as @Vanny says and run MemTest86 for a few hours to be certain it isn't your RAM. You're gonna have to be patient. It isn't a fast process.

 

[27MaD]

No guys i didn't try MemTest86 yet.

Ok i just dropped the latency and'll see how things go, went from 16-15-17-36 to 16-18-18-38 and increased voltage from 1.3 to 1.35V.

16-18-18-38

Crashed, went down to 2933MHz, same voltage.

Just tried to run each RAM alone in each slot, samething no luck.

 

[Deleted member 205776]

I would just get the G.Skill kit I mentioned earlier at this point to make sure RAM isn't the issue, and if it isn't, would you be able to return it?

 

[27MaD]

and if it isn't, would you be able to return it?

No i'd have to sell it.

 

[Mussels]

Just wanted to pitch in and say i've seen this same crap with early LPX ram as well on ryzen with instability that wont go away and is hard to detect - every single time its been C15 LPX ram.

The C16 and newer stuff has been fine, but basically if it wont run XMP stable out of the box you're gunna have a bad time and its not worth using that RAM.

 

[Ferrum Master]

Because everyone is having hard time understanding how RAM IC datasheet corelates with actual timings. Everyone shouts some nonsense, yet they have no experience themselves and have failed. Place starts to remind reddit.

No one even asked for full real timing screenshot actually.

First fault is to drop primary timings from the faulty XMP profile, that probably has a dud elsewhere, like hell primaries matter on rare random errors. Those are the culprits on constant errors, cold boot bugs etc often it cures with adding voltage and the OP imho already did up the voltage. Other timings are less sensitive on that.

First thing to do... always if something fails... Run you RAM at slow JEDEC rated speeds. If those fail, then you really have a possibility you have a bad RAM. If those work fine, then then we can only declare PBCAK. JEDEC timings are there for a reason.

LPX reacts on secondary and tertiary timings that are actually very important, I always remember Dave when saying this, at least he knew how things are done in RAM department as a reviewer. You have to read up and understand the sum of timings and how to set them up. For rare errors the usual culprit is the tRFC. The recommended values are found in the datasheet of your own specific DRAM IC model and if you run them on faster speeds, you have to increase the value, if no datasheet is found just start at very large values like 800. The second area where you should relax the IMC is the tRDWR and tWRRD and tRDRDC with tRDRDSD. Those timings has correlation in pairs and various combinations react differently on various motherboards Just trial and error no magic tools can help you with that as we own different HW combos usually.

You have to spend many hours on testing each setting. Do fast setting search = running 10 loops on linpack on max memory on each timing alteration step. Then the final settings leave for 12-24h. RAM overclocking has never been different really. There are just people who think their systems are stable unless they don't BSOD, yet they spew rare RAM errors, but for their workload it doesn't cause any problems so they declare YOLO and call themself experts. ECC memory should fix this at last for those lazy people as the log would show them also a ECC correction event I really hope DDR5 will have it in desktop segment too. Actually overclocking ECC enabled memory is such a breeze, I had the pleasure with DDR3 Samsung RAM sticks that I could push almost twice the speed and running it for months you knew it is stable. As you put incorrect timings the event log spewed ECC related events.