Overclock.net banner
19021 - 19040 of 19134 Posts

·
Registered
Joined
·
47 Posts
Try a looser tRCDRD... 21...22... or even 23 Might be low bin kits.
Thanks again for helping! :)

I just tried that now, 21 didn't even get past booting, 22 and 23 got past booting and crashed into windows, just same as before, using 1.45V.

I decided to try setting things slowly to auto to see what the MB would train for, and the results were quite absurd. It does make me wonder how this runs at CL 16 3600.

At those timings, I honestly didn't bother trying to boot, but it was a slightly better situation than not even getting past the memory checks and into the Bios as it did before.

Font Line Screenshot Software Electronic device
Font Line Screenshot Technology Software
Font Line Software Screenshot Technology
Font Line Screenshot Technology Software
Font Line Screenshot Technology Software


I understand the motherboard isn't too happy training 4 dimms like that, but does the timings indicate this could be some other type of IC? I wonder which though. I tried throwing the exported Thaiphoon file into Dramcalc and it claims this is a "good overclocking memory" about the timings (don't think this makes that huge of a difference though).

At this point, is there anything else I should try for?
 

·
Riding on the edge
Joined
·
3,305 Posts

·
Registered
Joined
·
331 Posts
Thanks again for helping! :)

I just tried that now, 21 didn't even get past booting, 22 and 23 got past booting and crashed into windows, just same as before, using 1.45V.

I decided to try setting things slowly to auto to see what the MB would train for, and the results were quite absurd. It does make me wonder how this runs at CL 16 3600.

At those timings, I honestly didn't bother trying to boot, but it was a slightly better situation than not even getting past the memory checks and into the Bios as it did before.

View attachment 2560419 View attachment 2560420 View attachment 2560421 View attachment 2560422 View attachment 2560423

I understand the motherboard isn't too happy training 4 dimms like that, but does the timings indicate this could be some other type of IC? I wonder which though. I tried throwing the exported Thaiphoon file into Dramcalc and it claims this is a "good overclocking memory" about the timings (don't think this makes that huge of a difference though).

At this point, is there anything else I should try for?
Maybe the default ProcODT or RTT aren't very good for four dimms in your case. You could try ProcODT up or down a step, I would think Hynix could like 40. I don't know if B550-F might like something different for RTT.

If tRCD 22t doesn't work for 3733 or 3800, it is probably not the problem. Make sure you are also using tRP 22 - 23. You can try setting WRWRDD to 9t if that isn't what the motherboard tries to train anyway. Most other stuff is pretty high.

Might also check SOC up to 1.125. I think VDDG and VDDP could be adjusted, but I don't think they are necessarily your problem.

For four dimms it probably saves some headache if you can take the time to test each one at a time and figure out what they can overclock to (and secondarily make sure one isn't weak or faulty if not tested already).

This may not matter but try the VRM settings back on Auto. I didn't have very good luck with phase control "Optimized" on my Asus X370, but also 200 for switching frequency seems like it is lower than what I would expect for default (if it is in kHz).

Why does this show Samsung D Die for F4-3600C14Q-32GTZNA G.SKILL Trident Z Neo Series 32GB (4 x 8GB) 288-Pin PC RAM DDR4 3600 (PC4 28800) Intel XMP 2.0 Desktop Memory Model F4-3600C14Q-32GTZNA - Newegg.com
Perhaps I'm misreading or plain not "in the know" about something?

View attachment 2560448
Because most dimms don't have the die stepping / revision programmed into the SPD. I have no idea why thaiphoon makes this particular guess though. Those are B-die.

For G.Skill, look at the code on the heatspreader label starting with 042. It will end 8810B for 8Gbit Samsung B-die.

edit for typos
 
  • Rep+
Reactions: Kuroihane

·
Registered
Joined
·
47 Posts
Maybe the default ProcODT or RTT aren't very good for four dimms in your case. You could try ProcODT up or down a step, I would think Hynix could like 40. I don't know if B550-F might like something different for RTT.

If tRCD 22t doesn't work for 3733 or 3800, it is probably not the problem. Make sure you are also using tRP 22 - 23. You can try setting WRWRDD to 9t if that isn't what the motherboard tries to train anyway. Most other stuff is pretty high.

Might also check SOC up to 1.125. I think VDDG and VDDP could be adjusted, but I don't think they are necessarily your problem.

For four dimms it probably saves some headache if you can take the time to test each one at a time and figure out what they can overclock to (and secondarily make sure one isn't weak or faulty if not tested already).

This may not matter but try the VRM settings back on Auto. I didn't have very good luck with phase control "Optimized" on my Asus X370, but also 200 for switching frequency seems like it is lower than what I would expect for default (if it is in kHz).
I might actually try to get the single dimm testing done. I was using the XMP @ 1.37V, while I don't get any errors, WHEA or whatever, I see stutter happening in moving my mouse at certain situations. I also noticed that it affects both my Logitech G533 and G603, I heard that the audio would stutter at those times on the headphones, so I am not really sure what could be the issue.

Everything was fine just before the GPU + RAM Swap I did recently, I upgraded my GTX 980 to the 6700 XT too (I don't think the GPU could actually be part of those problems), the stutter takes place in game sometimes, where audio and input lag can be noticed, but no FPS issues, SAM is turned in.

My first hunch was the memory, as it usually affects many things system wide. HCI Memtest at 3600/1.37V didn't show a single error on over 1100% test on each thread, but those stutters are really getting to me now. I just formatted my PC, as soon as it restarted for some drivers, the G533 dongle had a USB issue and didn't get recognized (happened most of the times while the RAM is overclocked) but then I noticed the same stutters when the RAM is not, no USB/Driver issues while not OCd yet though.

The 200 khz setting for the BIOS is default, it defaults at Optimized too. I can ramp up the VRM frequencies to 350 at maximum. Which software to test the dimms? Memtest86? HCI?

TL;DR: Still getting stutter even while the ram is not OCd (but less), resetted BIOS, even formatted the PC to no avail. Issues began after upgrading the PC Ram and GPU. I have a Seasonic 750W Prime Gold PSU. Sorry for throwing so many questions... I am just a bit afraid if I'd need to go through RMA in here, it's always not a pleasant experience where I live.

Edit: I went to the BIOS again and reflashed it to make sure just now.
 

·
Registered
Joined
·
331 Posts
I might actually try to get the single dimm testing done. I was using the XMP @ 1.37V, while I don't get any errors, WHEA or whatever, I see stutter happening in moving my mouse at certain situations. I also noticed that it affects both my Logitech G533 and G603, I heard that the audio would stutter at those times on the headphones, so I am not really sure what could be the issue.

Everything was fine just before the GPU + RAM Swap I did recently, I upgraded my GTX 980 to the 6700 XT too (I don't think the GPU could actually be part of those problems), the stutter takes place in game sometimes, where audio and input lag can be noticed, but no FPS issues, SAM is turned in.

My first hunch was the memory, as it usually affects many things system wide. HCI Memtest at 3600/1.37V didn't show a single error on over 1100% test on each thread, but those stutters are really getting to me now. I just formatted my PC, as soon as it restarted for some drivers, the G533 dongle had a USB issue and didn't get recognized (happened most of the times while the RAM is overclocked) but then I noticed the same stutters when the RAM is not, no USB/Driver issues while not OCd yet though.

The 200 khz setting for the BIOS is default, it defaults at Optimized too. I can ramp up the VRM frequencies to 350 at maximum. Which software to test the dimms? Memtest86? HCI?

TL;DR: Still getting stutter even while the ram is not OCd (but less), resetted BIOS, even formatted the PC to no avail. Issues began after upgrading the PC Ram and GPU. I have a Seasonic 750W Prime Gold PSU. Sorry for throwing so many questions... I am just a bit afraid if I'd need to go through RMA in here, it's always not a pleasant experience where I live.
Testmem 5, you can look in this link to a topic about it with config files. HCI is good, if it did 1000% per thread than that is pretty thorough.

Neither of those tests the fabric, if you do not have WHEA, you can still try adjusting voltages (VDDG CCD may be able to go down, IOD may be ok) and see if they help. Geekbench 5 could be used for the voltage steps. The CLDO VDDP may be ok. You can also look at "CPU 1.8 Voltage" and see if small steps up make a difference.

Can also try SAM off, or also check PCIe set to Gen 3 manually to see if that helps.

I don't know what sort of bugs Asus has in different versions of their B550-F bioses, but it is a common motherboard so someone that has one may be able to help with that.
 
  • Rep+
Reactions: Kuroihane

·
Registered
Joined
·
47 Posts
Testmem 5, you can look in this link to a topic about it with config files. HCI is good, if it did 1000% per thread than that is pretty thorough.

Neither of those tests the fabric, if you do not have WHEA, you can still try adjusting voltages (VDDG CCD may be able to go down, IOD may be ok) and see if they help. Geekbench 5 could be used for the voltage steps. The CLDO VDDP may be ok. You can also look at "CPU 1.8 Voltage" and see if small steps up make a difference.

Can also try SAM off, or also check PCIe set to Gen 3 manually to see if that helps.

I don't know what sort of bugs Asus has in different versions of their B550-F bioses, but it is a common motherboard so someone that has one may be able to help with that.
Seriously thank you for providing so much insight and taking your time.

After the BIOS reset I did I left SAM off... But there's a catch, I noticed I am on a Beta BIOS, at this point I can't even recall if I upgraded it via the Asus APP or actually flashed it.

I will continue testing the stutter issues, since that seems to be the biggest problem now. If they go away with SAM off, that's good enough. Sadly this version isn't too popular. There's the 550-F Wifi and Wifi II, the Wifi II's only difference is launch date and the integrated Wifi (it uses mediatek instead of Intel). I complained about this motherboard not having a Memory Voltage readout to Asus, they said the issue could be that 'The memory wasn't QVL'. They didn't tell me if there was either no sensor or if there just wasn't programming/BIOS to read it.

But the Bioses are "different" for some reason.
 

·
Registered
Joined
·
121 Posts
X570 Aorus master rev 1.1 here, F35b (Agesa 1.2.0.3b), 5950x, crucial ballistix 2x32gb micron 16gb rev B 3600c16 kit. I've been running 1900mhz 1:1:1 getting 61.7ns latency on just auto timings and 1.35v, 1T, GDM enabled with over 10000% karhu coverage. I finally decided to tighten the timings since now there is more info on this die. Big story short basically even with loose timings Geardown mode disabled (at least at 3800mhz, didn't try lower frequency) will not boot at all, or get stuck at the post screen. I thought it only rounded timings but it seems it does more things. Why is that and what do you recommend on doing?

here are the auto loose timings that have been running stably for ages as reference:

Rectangle Font Screenshot Parallel Technology
 

·
Registered
Joined
·
331 Posts
Seriously thank you for providing so much insight and taking your time.

After the BIOS reset I did I left SAM off... But there's a catch, I noticed I am on a Beta BIOS, at this point I can't even recall if I upgraded it via the Asus APP or actually flashed it.

I will continue testing the stutter issues, since that seems to be the biggest problem now. If they go away with SAM off, that's good enough. Sadly this version isn't too popular. There's the 550-F Wifi and Wifi II, the Wifi II's only difference is launch date and the integrated Wifi (it uses mediatek instead of Intel). I complained about this motherboard not having a Memory Voltage readout to Asus, they said the issue could be that 'The memory wasn't QVL'. They didn't tell me if there was either no sensor or if there just wasn't programming/BIOS to read it.

But the Bioses are "different" for some reason.
I didn't know they made a new revision, that's interesting. What AGESA was v 0305 (first release)? Had you used that bios before?

Are you on Windows 11? Stutter could also be from TPM, I don't know if it was really fixed for everyone. I have stayed on 10.

Are you using the Mediatek wifi or Intel Lan? One or other could maybe also be an issue.

Gigabyte 1.2.0.7 bios for B550 Master, I assume the X570 gets same treatment. Flashed mine and appears ok so far.
1207 still has voltage limit lowered with raised EDC?
 
  • Rep+
Reactions: Kuroihane

·
Gamer and overclocker !
Joined
·
3,012 Posts
Plant Nature Natural environment Font Biome
 

·
Registered
Joined
·
151 Posts
Gigabyte 1.2.0.7 bios for B550 Master, I assume the X570 gets same treatment. Flashed mine and appears ok so far.
The latest BIOS of my Gigabyte B550 VISION D says 1.2.0.7 too, but I thought maybe it was a typos. So Gigabyte has released the final version of AGESA.
 

·
Registered
Joined
·
2,518 Posts
i got my lesson , also after messing up @mongoled 's A0's Being equaly resposible for community not knowing what to push and what to lower on over-sensitive weak PCBs :)
Bro

sacrifices must be made

🤣🤣

Seriously, we have come along a long way since then, in particular yourself 😃😃
 
  • Love
Reactions: Veii

·
Registered
Joined
·
2,518 Posts
I managed to nab a 3080/3090 active backplate for £55 on eBay. 2nd hand but hasn't been used. Never do auctions for tech stuff folks unless you have something with a high ceiling. Buy it now or risk losing out. An open box active backplate should have got nearer £90+ as long as unused.

So, I still get to watercool something else after the failed NVMe attempt lmao

View attachment 2560302

It's a sickness I tell you. Once you start watercooling your PC you end up wanting to watercool everything.

Time to redo some bends as well and get rid of right angle pieces.

Even although everyone says no point in watercooling the back of a 3080 the active backplate looks nice and my previous nickel backplate with thermal pads still got warm to the touch. So if passive cooling on the back helps a little, water will help a little more.

The addiction isn't just about performance, it's about driving those temp readings as low as possible!
It's a sickness of the mind

😍😍
 
  • Haha
Reactions: Veii

·
Registered
Joined
·
2,518 Posts
just pay2win by trading out your flare-x kits for Patriot Viper 4400 sticks which actually work great in 4x8
View attachment 2560047
as showcased by some gentleman whose name escapes me at the moment, but I'm sure he's still around:
View attachment 2560048

They also work fantastic alongside the Viper 4000 16-16-16 sticks in a 4400-4000-4400-4000 (A0-A1-B0-B1, Bad-Good-Bad-Good slot) config on z690, but I don't know if you can make them work well on AM4 without running different RTT Park (Park/5 on A0/B0, Park/6 on A1/B1) values for the different slots.
Those must be mine 😃😃
 
  • Rep+
Reactions: Bloax

·
Registered
Joined
·
47 Posts
I didn't know they made a new revision, that's interesting. What AGESA was v 0305 (first release)? Had you used that bios before?

Are you on Windows 11? Stutter could also be from TPM, I don't know if it was really fixed for everyone. I have stayed on 10.

Are you using the Mediatek wifi or Intel Lan? One or other could maybe also be an issue.
Honestly, the board came with 0305, but I can't recall the AGESA version, I would say it was 1.2.0.4 or .6.

So, after reflashing the BIOS and keeping SAM disabled, it seems like there was improvement. I left HCI running overnight as always, this time, I changed some of the LLC and Phase controls. Got to nearly 1000% apart from the last thread, that for some reason lagged behind. The phase control I changed it to Manual and set it to "Fast". I think in case of this motherboard, the Switching frequencies are doubled as I am almost sure this MB uses doublers for it's phases, so 200 khz should be 400 instead, 350 which is the max config would be 700, does that sound more accurate?

I will run some tests today while gaming and see if I get stutters. Then maybe I will try increasing the memory freq to 3800 again.
Computer Rectangle Personal computer Font Line


Once again, thank you so much for your insight on this!
 

·
Premium Member
Joined
·
7,053 Posts

·
Registered
Joined
·
47 Posts
So... @hazium233

I am just tagging you because you've been helping me so much, I'd figure you'd be interested in an update on the lagging issues.

After much, much fidgeting and me almost going nuts on this (since the lags really seemed like unstable memory issues, as the mouse cursor would lag, then "teleport" somewhere and the headset would crackle/cut sound...)

I decided to try everything. I reset the BIOS, began trying from 0, unplugged all USB things, began plugging one by one. Everything seemed to work fairly fine, until the mice was in game.

I have a G603, which is both wireless with it's adapter and bluetooth. But aparently, the receiver had some issues and that was causing all the lags, delays and even USB disconnects I was getting. The sound device is another Logitech device, a G533 Wireless...

I don't even know how such a thing can happen, but I guess the receiver fried somewhere along this past week of me attempting to stabilize/improve my RAM OC. Either that or it just decided to commit sudoku.

Removed the receiver, connected the mouse via Bluetooth... No more lags (besides the bluetooth ofc), headset and sound works fine, there's no big delays or jumping of the cursor around anymore.

I already reached out to Logitech, as they usually have an amazing CS and Warranty, I am now waiting for a final response from them, it's under analysis.

With all that said, the ram is stable. The lags were caused by a single USB dork...dongle. My dumb-self honestly never imagined a single receiver could cause so much headache...
 
  • Rep+
  • Haha
Reactions: hazium233 and Veii
19021 - 19040 of 19134 Posts
Top