Yeah, was a boot sequence issue mostly, the BCM needs time to take over the sxm5 card
these are my logs:
sudo dmesg | grep nvidia
[ 4.674546] nvidia: loading out-of-tree module taints kernel.
[ 4.674551] nvidia: module verification failed: signature and/or required key missing - tainting kernel
[ 4.719369] nvidia-nvlink: Nvlink Core is being initialized, major device number 511
[ 4.720085] nvidia 0000:01:00.0: enabling device (0000 -> 0002)
[ 4.749968] nvidia-modeset: Loading NVIDIA UNIX Open Kernel Mode Setting Driver for x86_64 570.211.01 Release Build (dvs-builder@U22-I3-AM27-29-5) Tue Nov 25 16:11:49 UTC 2025
[ 4.806113] [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver
[ 4.806116] [drm] Initialized nvidia-drm 0.0.0 for 0000:01:00.0 on minor 1
sudo dmesg | grep gsp
[ 67.812295] NVRM: kfspSendBootCommands_GH100: version=0x1, size=0x35c, gspFmcSysmemOffset=0x178f00000
[ 67.812297] NVRM: kfspSendBootCommands_GH100: gspBootArgsSysmemOffset=0x130bb4000
[ 67.812319] NVRM: nvCheckOkFailedNoLog: Check failed: Call timed out [NV_ERR_TIMEOUT] (0x00000065) returned from kfspSendBootCommands_HAL(pGpu, pKernelFsp) @ kernel_gsp_gh100.c:834
so, this "Call timed out" I assume kills the process. My understanding H100 refuses to boot, have no clue how to force it to boot.