Search results

  1. A

    NVIDIA P4261 — Grace Hopper carrier with no BMC and no SCM slot. Anyone recognize this board?

    @dollop Have you figured out a cooler/heatsink solution for this yet? Building a High-End AI Desktop wish that guy would just post his step files somewhere... It's an amazing cooler design.
  2. A

    NVIDIA P4261 — Grace Hopper carrier with no BMC and no SCM slot. Anyone recognize this board?

    Update, a P4261 is on the way. Let's see how this goes.
  3. A

    NVIDIA P4261 — Grace Hopper carrier with no BMC and no SCM slot. Anyone recognize this board?

    I have ONE HGX HMC, I'm thinking of reading the firmware from it if I can to make sure i back it up, then replicating it and flashing non GH200 HMC to it. I think I'm getting one P4261 soon to actually play around though, since Ihave the HMC I might as well.
  4. A

    NVIDIA P4261 — Grace Hopper carrier with no BMC and no SCM slot. Anyone recognize this board?

    Hell of a rabbit hole I'm going down here. The P4261 is interesting. I've found a potential PCIEx16 card to drive one of the HMCs I have lying around, but overall it seems impossible to figure out how to connect them to the P4261. No way it's a MCIO to PCIE since the pcie link only trains once...
  5. A

    How to fix the AMD Starship/Matisse GPP Bridge PCIE ERR SPAM

    If you notice a AER hardware error corrected error spamming your SEL and chat, try this to silence the issue.
  6. A

    9300-16i PCIE PERR Starship/Matisse PCIE bridge

    Any idea how in a ROME 2d16-2t?
  7. A

    9300-16i PCIE PERR Starship/Matisse PCIE bridge

    i decided to upgrade to a 9305-24i and the 9300-16i ran too hot anyways, but ill definitely give it a go!
  8. A

    9300-16i PCIE PERR Starship/Matisse PCIE bridge

    I believe I narrowed down my insane SEL/IPMI log spam issue down to a PCIE PERR that is repeating. And I believe i narrowed it down to my LSI card- onboard LAN disabled, no other pcie devices installed yet issues continue. This error is also reported on the SEL/IPMI logs, which is why it's an...
  9. A

    ROME 2D16-2T OEM Error spam

    0e25h| 02h| 64b3e559h | 21h| 00h| 04h| c0h| 00h| 6fh| a1h| 10h| 32h| This is all the information in the SEL list. Has anyone seen this issue before? AsRock Rack support has been unhelpful, and their RMA system has been ignoring my emails for nearly two weeks.
  10. A

    Help! Rome 2D16-2T with 2x Epyc 7532

    UPDATE: I don't know what the **** I did but it's working.
  11. A

    Help! Rome 2D16-2T with 2x Epyc 7532

    Things get even wierder. Booting with 4 sticks on cpu 1 and 1 stick on cpu 2 threw memory errors on cpu 2 but booted into the screen.
  12. A

    Help! Rome 2D16-2T with 2x Epyc 7532

    The VCORE problem resolved itself after multiple power cycles.
  13. A

    Help! Rome 2D16-2T with 2x Epyc 7532

    Yup, CPU SOC for CPU 2 is only at 0.6v according to the BMC. However, I'm using an EVGA 1200 P2 Supernova, which should be fine. Contacting Asrock Rack with these questions as well. Both cpus seem to post just fine in socket 0 with 256GBs of RAM. Switching RAM sticks had no effect. (That was a...
  14. A

    Help! Rome 2D16-2T with 2x Epyc 7532

    This is very odd. The system worked fine before with 8x 64GB 2400 LRDIMM, and 1x Epyc 7532. When I plugged in the second cpu, no matter what I do including power cycles and cmos clears, the board always gets stuck in post code A7. Swapping sockets had no effect. Any ideas?
  15. A

    ----

    Indeed, the man sends with OVERNIGHT shipping, although fedex delayed mine a day or two.
  16. A

    ----

    Bumping for great seller. Consistent support. Bought Rome2D16-2T off of them, had a lot of trouble getting it to post at the start, turned out to be a bad solder joint from factory on VGA, BMC advice worked perfectly. I did have a lot of issues from this board, but it was directly from factory...
Top Bottom