[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#1031121: rocm-smi -S crash with assertion val_vec VDDC_CURVE ...



Control: tag -1 upstream
Control: forwarded -1 https://github.com/RadeonOpenCompute/rocm_smi_lib/issues/116

Hi Witold,

Witold Baryluk, on 2023-02-12:
> user@debian:~$ rocm-smi -S
> 
> 
> ======================= ROCm System Management Interface =======================
> ================== GPU Memory clock frequencies and voltages ===================
> python3: ./src/rocm_smi.cc:1131: rsmi_status_t get_od_clk_volt_info(uint32_t, rsmi_od_volt_freq_data_t*): Assertion `val_vec[kOD_VDDC_CURVE_label_array_index] == "OD_VDDC_CURVE:"' failed.
> Aborted (core dumped)

Thank you for your report!  On my end I see it doesn't work but
for a different reason apparently:

	$ rocm-smi -S
	
	
	======================= ROCm System Management Interface =======================
	================== GPU Memory clock frequencies and voltages ===================
	ERROR: 2 GPU[0]: od volt: RSMI_STATUS_NOT_SUPPORTED: This function is not supported in the current environment.	
	================================================================================
	============================= End of ROCm SMI Log ==============================

Reading:
> Kernel: Linux 6.2.0-rc5 (SMP w/32 CPU threads; PREEMPT)
> Kernel taint flags: TAINT_UNSIGNED_MODULE
I wondered whether the difference could be hardware or kernel
related.  For reference, I tested so far with:

	$ rocm_agent_enumerator
	gfx000
	gfx1030

	$ uname -srv
	Linux 6.1.0-3-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.1.8-1 (2023-01-29)

In any case, I ended up seeing that there is an issue open
upstream, so linked the BTS to follow that, we'll see when it
gets sorted in a newer version eventually.

Have a nice day,  :)
-- 
Étienne Mollier <emollier@emlwks999.eu>
Fingerprint:  8f91 b227 c7d6 f2b1 948c  8236 793c f67e 8f0d 11da
Sent from /dev/pts/2, please excuse my verbosity.
On air: Ferrigno, Leal & Kuprij - Ethiopia

Attachment: signature.asc
Description: PGP signature


Reply to: