Overclock.net banner
8,901 - 8,920 of 14,355 Posts
Welp, after a reboot, when my Windows was getting to the desktop, the screen went black and the Q-code showed 00 yesterday.
Tried the basic turning off the PSU, powering it back ON, and clearing CMOS. Still the same 00.
I think I just got a sudden death a la "x99 Rampage V" era. Tore my water loop and checked the CPU and socket pins, nothing. Ended up removing the CMOS battery and going to sleep, defeated.
I'll see when I get home if it still gives me 00 with nothing but the CPU and a stick of RAM. :cry:
 
  • Wow
Reactions: email.acc77
Does anyone else also have the Double Clod Boot issue ?

I'm using the X670E Gene and every time I start my computer, it shut down during the boot process, after like 30second, and then it finally start properly.
It doesn't happen when the PC is already On and I restart it.

Many Bios updates didn't solved this, also happening with default settings, and Asus Support doesn't help much so far. Thanks.
 
Yes true, but not so user friendly.
Yesterday I tried it again. Its terrible.

I have 1 question about ECLK. Try to save bios profile with MEM OC, and ECLK 101 after clear cmos and PSU off. Reboot and restore bios profile. Again reboot. Will the BIOS or OS boot?
I dont understand this question.

Does anyone else also have the Double Clod Boot issue ?

I'm using the X670E Gene and every time I start my computer, it shut down during the boot process, after like 30second, and then it finally start properly.
It doesn't happen when the PC is already On and I restart it.

Many Bios updates didn't solved this, also happening with default settings, and Asus Support doesn't help much so far. Thanks.
Step by Step, first load default settings
unplug all usb devices (not mouse and keyboard)
Test it again...
unplug one DDR5 stick
Test it again...
re-install Windows
and so on
In such cases you should isolate the problem!
Pls try a minimal config of you hardware
I hope you know what I mean.

Feel free to write all your hardware here.
 
Does anyone else also have the Double Clod Boot issue ?

I'm using the X670E Gene and every time I start my computer, it shut down during the boot process, after like 30second, and then it finally start properly.
It doesn't happen when the PC is already On and I restart it.

Many Bios updates didn't solved this, also happening with default settings, and Asus Support doesn't help much so far. Thanks.
Have you tried a different RAM kit or PSU?
 
Does anyone else also have the Double Clod Boot issue ?

I'm using the X670E Gene and every time I start my computer, it shut down during the boot process, after like 30second, and then it finally start properly.
It doesn't happen when the PC is already On and I restart it.

Many Bios updates didn't solved this, also happening with default settings, and Asus Support doesn't help much so far. Thanks.
can you try turn off erp in bios?
from asus raja

The rig starting twice with erp enabled is normal. Just the way power needs to be sequenced.
 
yes, the karhu + 3DMark combination just happens to crash (i.e. reboot) very quick with unstable fclk on my system. I didn't test, but i would think that prime95 large ffts + furmark does the same. In this DDR4 OC guide (MemTestHelper/DDR4 OC Guide.md at oc-guide · integralfx/MemTestHelper), it is also recommended to test IF stability using occt VRAM test + prime 95 large ffts in parallel. Games will work, too, but probably not as consistent.

regarding mismatch: There are setting combinations causing mismatch/match between boots. I think you accidently had a mismatch due to that, not because you ran that tool.
Today screen froze out of nowhere so tested 2200FCLK again w karhu + timespy extreme in 30C ambient, last time i have ambient at 14C. Crashed within a couple minutes :( So it is temperature related as well.

Do you change anything in timespy besides enable looping? Is cpu test disabled and demo ram set auto or below 2GB?

Hwinfo report cpu die (average) & 3 of those SVI3 TFN report report crazy number, like twice the amount of temperature and voltage. It can't be right right? IOD hotspot is 97C and CPU die 140C o_O
Image


Testing Karhu at 2167FCLK atm, no crashing but the speed is only ~161MBps.
 
Does anyone else also have the Double Clod Boot issue ?

I'm using the X670E Gene and every time I start my computer, it shut down during the boot process, after like 30second, and then it finally start properly.
It doesn't happen when the PC is already On and I restart it.

Many Bios updates didn't solved this, also happening with default settings, and Asus Support doesn't help much so far. Thanks.
This only happens to me if you change something to do with ram in the bios, timings side.
 
Today screen froze out of nowhere so tested 2200FCLK again w karhu + timespy extreme in 30C ambient, last time i have ambient at 14C. Crashed within a couple minutes :( So it is temperature related as well.

Do you change anything in timespy besides enable looping? Is cpu test disabled and demo ram set auto or below 2GB?

Hwinfo report cpu die (average) & 3 of those SVI3 TFN report report crazy number, like twice the amount of temperature and voltage. It can't be right right? IOD hotspot is 97C and CPU die 140C o_O
View attachment 2647533

Testing Karhu at 2167FCLK atm, no crashing but the speed is only ~161MBps.
Can you try going lower on vsoc? It was my saviour involving gaming. Your IOD hitting 97c! Some weird max readings you have, are you using amoury crate or other software like icue, nzxt etc those conflict with Hwinfo

I just eked a timespy extreme stress test pass today.
Image
 
Can you try going lower on vsoc? It was my saviour involving gaming. Your IOD hitting 97c!

I just eked a timespy extreme stress test pass today.
Image
There's something I haven't figured it out yet, lower vsoc or vddp cause mismatch. I'm not sure my cpu is degrading or it's something else. Before 1.15vddp fix the mismatch but now has to up it to 1.155. And I think that's IOD temp is faulty reading, other temp is normal.

I only have g.skill lightning software to turn off the ram rgb. Never use any other software to control, all fans & pumps are in other rooms so i never have to deal with noise. Those D5 pumps are built different, I have one that is from switchtech and it's been running since I got the first intel CPU, socket 1366 iirc.
 
Does anyone knows why the BIOS limits the VDDSOC voltage max to 1.234v? I cant seem to run it higher than the voltage stated, offset mode only have negative sign and manual mode doesnt do anything. My BIOS is on the latest 1904 version. I still (not always) occasionally experience the PC random lockup during shutdown (Screen blank but PC keeps running, and the RGB is stuck on whatever last pattern is set). This is a big problem if it gets stuck at shutdown, because I usually remotely turn on and off my PC, if it gets stuck I have no way to reset it.

Does anyone knows how to resolve or experience stuck at shutdown issue? From older AM4 ryzen I heard disable C-States will help, but I wonder will it work on AM5?
 
ultimative FCLK test: karhu + 3DMark
If you want to reduce the amount of time to failure, use y-cruncher Large FFTs with Furmark - that'll bring out FCLK instability much quicker as it isolates the IMC.


The rig starting twice with erp enabled is normal. Just the way power needs to be sequenced.
No it isn't - I have S4+S5 deep sleep enabled and a single cold boot is normal behaviour - it'll be related to another BIOS setting.

Also "edge" stability can cause this with RAM timings as @Goose3080 mentioned, either loosen timings/bump up voltages or deal with the slightly better performance of tighter timings which require multiple training passes - you can check this by monitoring the Q-Code or 4 LED lights, DRAM will show amber when this is happening.


There's something I haven't figured it out yet, lower vsoc or vddp cause mismatch.different, I have one that is from switchtech and it's been running since I got the first intel CPU, socket 1366 iirc.
When tuning my PHY timings I managed to run VDDP 0.03V lower completely stable for hours of memory testing, but suffered mismatch - it needed a bit of extra voltage beyond what was "stable". Other voltage which seemed to affect it was uncoupled VDD and VDDQ on the DDR5 side, running VDDQ (data bus side) too low below VDD (DDR5 memory controller side) caused mismatch. Most kits should be able to decouple VDD/VDDQ and run lower VDDQ on databus side. VDDIO (CPU IMC) and VDD need to be matched however.
 
Today screen froze out of nowhere so tested 2200FCLK again w karhu + timespy extreme in 30C ambient, last time i have ambient at 14C. Crashed within a couple minutes :( So it is temperature related as well.

Do you change anything in timespy besides enable looping? Is cpu test disabled and demo ram set auto or below 2GB?

Hwinfo report cpu die (average) & 3 of those SVI3 TFN report report crazy number, like twice the amount of temperature and voltage. It can't be right right? IOD hotspot is 97C and CPU die 140C o_O


Testing Karhu at 2167FCLK atm, no crashing but the speed is only ~161MBps.
As others have stated, those are readout errors due to conflicting monitoring programs. I didn't pay too much attention to IOD temps, but you should be fine up to ~ 45C or more. My FCLK is unstable at 2200 at way lower IOD temperatures, when stress test is starting.

When using Karhu + 3DMark, I would set 24-26 GB for Karhu, that is plenty left for 3DMark. I am mainly using Speed Way, for Time Spy Extreme, I would use window mode, looping enabled, and probably the Graphic Test 2 (think it's a bit harder).

You can also try y-cruncher Large FFTs + Furmark as @Desolutional suggests.

For matching tPHYRDL, many settings are important, it also depends on your CPU. In my experience the most important is VDDP, where higher values help getting it matched. VDDIO also seems to play a role. Decoupled VDD/VDDQ doesnt matter on my system, neither do I need to couple VDDIO and VDDQ. I have observed that setting phase shifts like this or to 2/2/2/6/6/6 can help with getting matched PHY in 1:2 mode. YMMV, every CPU is different.
Image
 
I have observed that setting phase shifts like this or to 2/2/2/6/6/6 can help with getting matched PHY in 1:2 mode. YMMV, every CPU is different.
Higher TX CK phase shift exactly (clock line phase shift).

In my experience TX phase shifts in 2:1 mode can go from 0 to 3. Higher values are not necessarily better for stability, but worth testing up to 3/3/3.

Running different CA/CS/CK was always worse in my testing, but if someone desperately wants to match tPHYRDL and 3/3/3 results in no boot, I guess worth trying 2/2/3.

My A-dies was always fine with 3 but 24M-die doesn't boot above 2 TX CA (command/address).
 
I dont understand this question.
Gigabyte support said me: It's problem AMD. On a gigabyte motherboard, there is a problem with restoring the profile with configured eclk and RAM after completely shutdown the PC from PSU and clear CMOS. I set ECLK and do reboot. After this I tune RAM and do reboot. This is the only way both ECLK and overclocked memory work. If you clear cmos and shutdown PC from PSU on 1 minute does restoring profile work?
 
Anyone else getting 503 - Service Not Available
from Asus website now?
 
VDDIO (CPU IMC) and VDD need to be matched however.
But my VDD is at 1.56, isn't it a bit high for VDDIO?

As others have stated, those are readout errors due to conflicting monitoring programs. I didn't pay too much attention to IOD temps, but you should be fine up to ~ 45C or more. My FCLK is unstable at 2200 at way lower IOD temperatures, when stress test is starting.

When using Karhu + 3DMark, I would set 24-26 GB for Karhu, that is plenty left for 3DMark. I am mainly using Speed Way, for Time Spy Extreme, I would use window mode, looping enabled, and probably the Graphic Test 2 (think it's a bit harder).

You can also try y-cruncher Large FFTs + Furmark as @Desolutional suggests.

For matching tPHYRDL, many settings are important, it also depends on your CPU. In my experience the most important is VDDP, where higher values help getting it matched. VDDIO also seems to play a role. Decoupled VDD/VDDQ doesnt matter on my system, neither do I need to couple VDDIO and VDDQ. I have observed that setting phase shifts like this or to 2/2/2/6/6/6 can help with getting matched PHY in 1:2 mode. YMMV, every CPU is different.
I'm currently use phase shift setting from @BlueShark 2/1/2/6/7/7/1 ( bios profile screenshots , vddp was 1.15 -> 1.155, vddio: 1.45 and both VDD=VDDQ=1.56v.

But I'm gonna try both see which helps me to get rid of the occasional mismatch and lower voltage. Mismatch seems to appear more frequent at 2200FCLK.
 
The order in which these phase shifts run during DDR5 memory operation is as follows:
  1. CA (Command/Address) Phase Shift: The memory controller adjusts the timing of command and address signals (CA lines) relative to the clock signal. This ensures accurate interpretation of commands and addresses by the memory modules.
  2. TX (Transmit) Phase Shift: The transmit path involves adjusting the timing of data signals transmitted from the memory controller to the memory modules. Proper TX phase shift ensures correct data sampling during write operations.
  3. CS (Chip Select) Phase Shift: The CS signal selects the specific memory module for data access in multi-channel configurations. Maintaining coherence across CS signals is crucial for accurate performance.
  4. CK (Clock) Phase Shift: The clock signal synchronizes data transfers between the memory controller and modules. Precise CK phase shift ensures correct data sampling during read and write operations.
These phase shifts collectively contribute to DDR5 memory stability and efficient operation.
 
8,901 - 8,920 of 14,355 Posts