[SOLVED] MSI MAG Tomahawk wifi with i9-11900K, RTX 3090 - BSODs all the time !! What is wrong with it ?

Jul 4, 2021
15
0
10
Hello everyone,

I recently build my new pc. I was excited like a little kid when I finished building it :) Im a CGI artist, that's why this build (full spec on the bottom). I did some research about the CPU and its single core performance is out of this world. And that's what I need for simulations, GPU for rendering. I am no PC expert, but i do use computers for over 25 years. I know enough to build my own and overclock it. I did it 6 years ago with my old pc which is still running perfectly at 4.7ghz (i7-4790k). And I always do a thorough research before any major PC changes. So I was happy with this build. I gone for beQuite stuff, cause they are good enough ,but mostly they are quiet. Stupidly quite I must say :) I might change for water-cooling one day, but i never tried it so I didn't want to risk flooding brand new pc.

!! Tl;dr version !!: BSODs all the time, no matter what settings I choose in bios. My question and only idea i have atm - is 850W enough power to run this pc ? I did checked it before buying in few different places, and all seemed good.
And one more thing that comes to my mind now - this MB has 2 CPU power connectors - CPU_PWR1 (8 pins - 2x4) and CPU_PWR2 (4 pins - 1x4). One next to each other. I also checked this out before even starting up this PC for the first time, apparently I only need that 8 pin connected. But maybe its not enough ? Should i connect it to both ?

In details:
Everything worked almost perfectly from start on the default bios settings. I didn't change anything for almost two weeks.
I got installed few basic programmes (typical windows stuff, but also XMU soft, MSI Center, but quickly removed it and installed MSI Dragon App), and one game to test that beast of a graphic card in action (Call of Duty Warzone)
Game was the first thing i tested. It started to BSOD like crazy straight away. The game didn't even manage to get to the first menu - it was dying on loading screen.
Then i started doing some stress test under different BIOS settings (making changes after each BSOD). I gone through every setting i could think of. if something wasnt clear i googled it, got some info how it should be set. I thought this will be enough. i ve also watched video tutorials about OC'ing this specific CPU. 3 different videos, i tried few combinations of settings. nothing worked.
- AIDA64 Stress CPU and FPU - CPU was fine , no noticable problems with it. FPU test got extremely hot - around 100*C first time, but it didnt kill the pc. But that temperature maybe extreme, but i guess expected to go that high on AVX tests. Next time AIDA just stopped the test with a warning that is overheating.
-MSI CPU-Z - simple stress tests - on stock BIOS settings ( no oc, no XMP for ram, just as it came from the factory or the shop i bought it in) all is good, i can run it for a long times, nothing happens.
I ve manage to get BSOD during these test as well, but only when i set everything in BIOS to run at full STOCK power (lets call it STOCK OC - so whatever maximum power this CPU can reach without actually overclocking it ). So RAM XMP profile was selected at this point, and I either manually set all the CPU cores ratios to fixed values but same as stock, or I used auto ratios values but with fixed vcore. I didn't overclock it at any point during these test :D So max at .3 ghz. - still BSODs all the time. I was searching forums for similar problems, and some say that this MB had some issues, some saying this CPU can't run at 5.3 on turbo, so they disabled it, but it runs at 5.2... on some occasions.
- Cinebecnch23 - Stock setting (right after i build the pc) - all good, i run both test from start to finish no crash. After changing it to STOCK OC i started to get BSODs.

BSODs - most common one - Clock Watchdog Timeout , and one or two more which i cant remember now but if i will get it again , i will update it here.
I have HWiNFO64 running in the background, I'm checking all the parameters. And until today i also had Intel XMU to check all those values (I am not using it for any overclocking) and MSI Dragon Center (just to see what i can customize, not using for any OC whatsoever). I removed these two software's just to make sure they not changing anything in the system.

Yesterday i had some success in running that game for a bit longer but i had to disconnect every single device that drain any power apart from the essentials. I left it with stock case fans (3), cpu cooler , one monitor, no other usb device apart from mouse and keyboard. I also updated BIOS to the latest version. Still BSODs, but after about 30-45 min of playing. i was checking temperatures while playing, nothing abnormal. Around 60-70*C I heard that this CPU can run hot so im not worrying about temperatures too much. It didn't died when it reached 100 :)

Here is a full PC spec:
MB: MSI MAG Z590 TOMAHAWK WIFI (Socket LGA 1200) DDR4 ATX
CPU: Intel i9 - 11900K ( with Arctic MX-5 Thermal Paste - see attached picture on how it was applied - even spread with spatula https://drive.google.com/file/d/1ftHvEFY5lbkTEUPy2dXsyRsYsaunEQcS/view?usp=sharing)
Cooler: be quiet! Dark Rock Pro 4 CPU Cooler
GPU: MSI RTX 3090 Suprim X
PSU: be quiet! Straight Power 11 850W 80 Plus Platinum Modular Power Supply
RAM: Corsair Vengeance LPX Black 64GB 3200MHz DDR4 Dual Channel Memory Kit
SSD: Samsung 980 PRO V-NAND M.2 1TB
HDD: WD Red Plus 4TB 3.5 SATA 128MB HDD
CASE: be quiet! Dark Base Pro 900 Rev.2 Full Tower
FANS: 3 stock with the case (2x 140mm, 1x 120mm) plus extra 1x 140mm and 2x120mm

My question again: is 850W enough power to run this beast ? If that's not the case, could there also be something wrong with the cables ?
And one more thing that comes to my mind now - this MB has 2 CPU power connectors - CPU_PWR1 (8 pins - 2x4) and CPU_PWR2 (4 pins - 1x4). One next to each other. I also checked this out before even starting up this PC for the first time, apparently I only need that 8 pin connected. But maybe its not enough ? Should i connect it to both ?

Sorry for long post. I appreciate any suggestions or ideas to what is wrong with my pc. Thank you.
 
Solution
Clock watchdog is a very common symptom of unstable OC'ing.

My question again: is 850W enough power to run this beast ? If that's not the case, could there also be something wrong with the cables ?
And one more thing that comes to my mind now - this MB has 2 CPU power connectors - CPU_PWR1 (8 pins - 2x4) and CPU_PWR2 (4 pins - 1x4). One next to each other. I also checked this out before even starting up this PC for the first time, apparently I only need that 8 pin connected. But maybe its not enough ? Should i connect it to both ?
1)Yes. Reviews suggest the unit is good, but I can't say more than that - I'm no expert.
2)Make sure they're not loose and seated fully.
3)You should connect both, especially when attempting any...

Phaaze88

Titan
Ambassador
Clock watchdog is a very common symptom of unstable OC'ing.

My question again: is 850W enough power to run this beast ? If that's not the case, could there also be something wrong with the cables ?
And one more thing that comes to my mind now - this MB has 2 CPU power connectors - CPU_PWR1 (8 pins - 2x4) and CPU_PWR2 (4 pins - 1x4). One next to each other. I also checked this out before even starting up this PC for the first time, apparently I only need that 8 pin connected. But maybe its not enough ? Should i connect it to both ?
1)Yes. Reviews suggest the unit is good, but I can't say more than that - I'm no expert.
2)Make sure they're not loose and seated fully.
3)You should connect both, especially when attempting any kind of OC. That's why it's there. If you weren't doing that, the 4pin isn't necessary.

I ve manage to get BSOD during these test as well, but only when i set everything in BIOS to run at full STOCK power (lets call it STOCK OC - so whatever maximum power this CPU can reach without actually overclocking it ).
This can vary by a little, or a lot with motherboards. Intel does not heavily enforce what they consider stock.
 
  • Like
Reactions: Hyyde88
Solution
Jul 4, 2021
15
0
10
I triple checked all connections, they all fine i think. But that cpu_pwr1-2 was bugging me a lot. So I will try to connect it and see what will that do. I didn't want to do it without asking first, I'm afraid to burn it somehow. :)
 

Zerk2012

Titan
Ambassador
I triple checked all connections, they all fine i think. But that cpu_pwr1-2 was bugging me a lot. So I will try to connect it and see what will that do. I didn't want to do it without asking first, I'm afraid to burn it somehow. :)
MSI Dragon Center

This is known to be buggy uninstall it.
The 11900K can draw a lot of power so yes plug in the other CPU power plug.
For the power supply their a bunch of them that cant handle the power spikes of the 3090 video card.
 
  • Like
Reactions: Hyyde88
Jul 4, 2021
15
0
10
Just about the only ways you could possibly burn it:
-force the plug in the wrong way, then power it on.
-spill fluids on it, then power it on.
good to know thanks :)

MSI Dragon Center

This is known to be buggy uninstall it.
The 11900K can draw a lot of power so yes plug in the other CPU power plug.
For the power supply their a bunch of them that cant handle the power spikes of the 3090 video card.
I have uninstalled it today. And i must say i saw a difference in the game 20 min ago. It didn't BSOD, but the game crashed after a while but might not be related to any of this. I was checking temperatures on CPU and GPU - all good - around 60-70*C. GPU reached almost 90 at some point but just for a moment.

Is there a way of checking GPU power spikes somehow ?
 
Jul 4, 2021
15
0
10
One more question about CPU power cables. I have one cable that is 4+4 and one that is 8 pins (can`t split this).
Can i use both ? the 8pin will connect to CPU_PWR1 and 4 pin connector to CPU_PWR2 ?
Both cables are marked 12V P8 connectors.

Uodate: i think i did something very wrong when I connected this cable for the first time. I'm guessing p4 and p8 shouldn't be connected to the same connector. Especially if they are named cpu1 and cpu2. https://drive.google.com/file/d/1g5HFc0oc4Pz4oj9BSHtaRRPVUKGU8DM6/view?usp=drivesdk

Update 2:
I think i must have been hallucinating when plugging these 4+4 cables in first time. Clearly that wasnt right. Now I put only full 8 pin in , and Aida64 CPU+FPU smashing for few minutes now, temps are jumping quite a lot, but average is 80-85C. Although CPU package is at 93-95C. No BSOD so far. But still stock bios settings. nothing has changed.
https://drive.google.com/file/d/1eeQGxpVjM_dy_VUXs6OXXmqckBKEo7yW/view?usp=sharing
 
Last edited:
Jul 4, 2021
15
0
10
Thanks, i thought that would be the case. i did the same with GPU ( it requires 3x6pin connections = 3 seperate cables ) but I'm testing only single 8pin connection and it is working lovely. i think i wont need another 4 added. unless it will start behaving weird. I think those 4+4 are not meant to be plugged both into the same connector. Lesson learned

So to wrap this up - i am an idiot when it comes to connecting cables :) i must have missed it somehow. And ive never seen 4+4 cables before.

I was watching this guy
View: https://www.youtube.com/watch?v=uNVPBH0bgQU
and tried to find same options in my BIOS, not everything is there, his MB is designed for OCing, i choose wrong one :) But it was cheap. Anyway I'm not planning to OC it for some extreme ghz's. ill try 5.5 on a single core ,on this cooler and i might leave it at that if its stable.

I'll drop BIOS screenshot with setting, if anyone is interested. But maybe in a new topic. Ill link it here as well.
Ill keep testing it for next few days and update this post if something new will come up. I hope this will help someone in the future. Just like it would have helped me 2 weeks ago :)

But i would like to thank you guys for quickly suggesting what's wrong.

XMP on, CPU ratios set to fixed values - 2 of the cores at 5.3 rest at 5.1 so thats what should be running at default. And it was stable and very low temps.
Voltage set to Adaptive offset based on the video from the link above. bit lower offset on the last core 0.1 instead of 0.125 but i will push it now, lest see. All fans connected now.
zlE4462.jpg


CPU+FPU tests this is beautiful, 92*C max after 5 min, i know its not too long but before it wouldn't last a minute.
iE8VprQ.jpg


Current Bios settings - stable for now, but test were 3 min long. No BSOD , so that's what i was aiming for.
Cinebench23 multicore test almost matched 16c Threadripper, not too bad in my opinion.
lHh8mb5.jpg

oyCUOBx.jpg


FEYYVgu.jpg

yPqv1sq.jpg
 
Last edited: