6
Random resets after cpu+GPU upgrade (Arch) (lemmy.world)
submitted 3 months ago by to c/linuxquestions@lemmy.zip
13 comments fedilink hide all child comments
 

Upgraded to amd 5900xt & 7090xt and am experiencing random crashes. No overclock, no xmp/docp, no pbo (not defaults cause Asus has pbo on by default). 1000w seasonic PSU.

Whole screen goes green and reboots. Happens in heavy gaming, watching YouTube and even at idle.

Journalctl lists an mc5 hardware error suggests the cpu, but none of the troubleshooting has fixed it.

BIOS updated, c-states disabled, triple checked all settings, fully up to date Arch, but still randomly crashes.

Where do I go from here?

all 13 comments

sorted by: hot top controversial new old
[–] 4 points 3 months ago (3 children)

What motherboard? The 5900x and 5950x are pretty power hungry CPUs and low end boards might not play well with them.

  • permalink
  • fedilink
  • source
  • hideshow 3 child comments
    •
  • [–] [S] 3 points 3 months ago* (2 children)

    Tuf b550-m (no Wi-Fi), not low end, but certainly not top tier

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 3 points 3 months ago* (last edited 3 months ago) (1 child)

    Is this the board? The VRM situation looks pretty dire, but I wouldn't think that'd be an actual problem. Also why do you have C states disabled? That's not going to be helping your VRM situation.

    Also what does your CPU cooling situation look like? Do you have a tower cooler that blows air over the VRM, or water cooling that leaves the VRM baking? Also generally for best performance you'd want xmp/docp/whatever the other thing is turned on. It's free performance if your RAM supports it.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] [S] 2 points 3 months ago

    Yep, that's the mobo. C states disabled as that was the advice given on arch forums. Apparently dialing c states is supposed to prevent the os from going into low power mode then crashing when voltage sags.

    Cooler is noctua nhd15 with ptm7950, never been above 75c even in stress tests. Second fan on the cooler is as low as possible to blow over the vrm heat sync close to rear i/o.

    Literally all of the advice for crashes says to disable xmp/docp. At the moment performance isn't the concern, it's the crashes

  • permalink
  • fedilink
  • source
  • parent
    • [–] 3 points 3 months ago

    This is what comes to mind for things to check:

    • Check your cables. Make sure none of them are loose, damaged, or obstructing fans
    • Make sure your RAM is seated properly, your GPU isn't sagging, the heatsink is properly attached to your CPU, etc.
    • Watch the thermals and physically check to make sure that your fans are actually spinning when they should be
    • Disconnect the GPU and see if it still fails
    • Run a memtest
    • Try a different kernel (newer is usually but not always better...)

    Good luck. Hope you can get the gremlins worked out soon!

  • permalink
  • fedilink
  • source
    •
    [–] 0 points 3 months ago (7 children)

    Your PSU is too weak.

  • permalink
  • fedilink
  • source
  • hideshow 7 child comments
    •
  • [–] [S] 0 points 3 months ago (6 children)

    Really? 1000w isn't enough? GPU+cpu combined shouldn't use more than 450 at full load, so I should have plenty of room for transient spikes.

    If I really don't have enough now how much is enough?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 6 child comments
    •
  • [–] -1 points 3 months ago (3 children)

    Yes, your combined wattage use at peak is up to 1250, so you need to downclock, or buy a powerhouse PSU.

    Your error literally says devices dropped offline. How would you imagine that happens aside from power drops?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] [S] 0 points 3 months ago (2 children)

    Your error literally says devices dropped offline.

    That's definitely not the wording in the error, so I didn't realize that's what it meant. Is there anywhere you'd recommend I read up on mc5 errors?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] -3 points 3 months ago* (1 child)

    Mc5 hardware error

    Dafuq you think that is, son?

    *Edit: post logs, ya baby ass bitch

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] [S] 1 point 3 months ago* (last edited 3 months ago)

    why the fuck would i ask if i already knew?

    [Hardware Error]: System Fatal error.

    [Hardware Error]: CPU:12 (19:21:2) MC5_STATUS[-|UE|MiscV|AddrV|PCC|TCC|SyndV|-|-|-]: 0xbea00000>

    [Hardware Error]: Error Addr: 0x00007fb9e0b2ab0c

    [Hardware Error]: IPID: 0x000500b000000000, Syndrome: 0x000000004d000000

    [Hardware Error]: Execution Unit Ext. Error Code: 0

    [Hardware Error]: cache level: RESV, tx: GEN, mem-tx: GEN

    [Hardware Error]: System Fatal error.

    [Hardware Error]: CPU:3 (19:21:2) MC5_STATUS[-|UE|MiscV|AddrV|PCC|TCC|SyndV|-|-|-]: 0xbea000000>

    [Hardware Error]: Error Addr: 0x01ffffffc0a06ae4

    [Hardware Error]: IPID: 0x000500b000000000, Syndrome: 0x000000004d000000

    [Hardware Error]: Execution Unit Ext. Error Code: 0

    [Hardware Error]: cache level: RESV, tx: GEN, mem-tx: GEN

    [Hardware Error]: System Fatal error.

    [Hardware Error]: CPU:24 (19:21:2) MC5_STATUS[-|UE|MiscV|AddrV|PCC|TCC|SyndV|-|-|-]: 0xbea00000>

    [Hardware Error]: Error Addr: 0x01ffffffb895f8b4

    [Hardware Error]: IPID: 0x000500b000000000, Syndrome: 0x000000004d000000

    [Hardware Error]: Execution Unit Ext. Error Code: 0

    [Hardware Error]: cache level: RESV, tx: GEN, mem-tx: GEN

  • permalink
  • fedilink
  • source
  • parent