Thor power pack issues

[maspien]

I’ve been getting over volts since day one. Even with low profile power settings and non-memory intensive LLMS and image gens. I’ve tried several different wall outlets. LLMs are 4 tokens a second and images take around 30seconds. I’ve replaced all @ with !!

grep “” /sys/class/hwmon/hwmon*/oc*_event_cnt

/sys/class/hwmon/hwmon2/oc1_event_cnt:0

/sys/class/hwmon/hwmon2/oc2_event_cnt:14

/sys/class/hwmon/hwmon2/oc3_event_cnt:8429

01-15-2026 03:13:55 RAM 96427/125772MB (lfb 2x2MB) SWAP 0/2048MB (cached 0MB) CPU [1%!!972,0%!!972,0%!!972,0%!!972,0%!!972,3%!!972,0%!!972,0%!!972,0%!!972,0%!!972,2%!!972,0%!!972,0%!!972,5%!!972] EMC_FREQ 2%!!2750 GR3D_FREQ !![314,314,314] NVENC0_FREQ !!314 NVENC1_FREQ !!314 NVDEC0_FREQ !!314 NVDEC1_FREQ !!314 NVJPG0_FREQ !!314 VIC off OFA_FREQ !!314 PVA0_FREQ off APE 300 cpu!!44C tj!!45.625C soc012!!43.781C gpu!!45.625C soc345!!42C VDD_GPU 2364mW/17365mW VDD_CPU_SOC_MSS 6305mW/18071mW VIN_SYS_5V0 6908mW/17224mW VIN 23402mW/63069mW

01-15-2026 03:13:56 RAM 96415/125772MB (lfb 2x2MB) SWAP 0/2048MB (cached 0MB) CPU [17%!!972,0%!!972,8%!!972,0%!!972,3%!!972,3%!!972,0%!!972,6%!!972,4%!!972,0%!!972,6%!!972,0%!!972,0%!!972,0%!!972] EMC_FREQ 2%!!3200 GR3D_FREQ !![1430,1430,1430] NVENC0_FREQ !!314 NVENC1_FREQ !!314 NVDEC0_FREQ !!314 NVDEC1_FREQ !!314 NVJPG0_FREQ !!314 VIC off OFA_FREQ !!314 PVA0_FREQ off APE 300 cpu!!44.031C tj!!45.906C soc012!!43.562C gpu!!45.781C soc345!!42.156C VDD_GPU 4727mW/17223mW VDD_CPU_SOC_MSS 6699mW/17943mW VIN_SYS_5V0 7213mW/17112mW VIN 26526mW/62658mW

01-15-2026 03:13:57 RAM 96388/125772MB (lfb 1x4MB) SWAP 0/2048MB (cached 0MB) CPU [1%!!972,0%!!972,1%!!972,0%!!972,6%!!972,0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,4%!!972,0%!!972,0%!!972,0%!!972] EMC_FREQ 1%!!2750 GR3D_FREQ !![989,989,989] NVENC0_FREQ !!314 NVENC1_FREQ !!314 NVDEC0_FREQ !!314 NVDEC1_FREQ !!314 NVJPG0_FREQ !!314 VIC off OFA_FREQ !!314 PVA0_FREQ off APE 300 cpu!!43.687C tj!!45.312C soc012!!43.406C gpu!!45.312C soc345!!41.75C VDD_GPU 3545mW/17071mW VDD_CPU_SOC_MSS 6699mW/17818mW VIN_SYS_5V0 7010mW/16999mW VIN 23232mW/62220mW

01-15-2026 03:13:58 RAM 96385/125772MB (lfb 1x4MB) SWAP 0/2048MB (cached 0MB) CPU [16%!!972,1%!!972,4%!!972,8%!!972,0%!!972,0%!!972,0%!!972,7%!!972,4%!!972,0%!!972,4%!!972,0%!!972,0%!!972,3%!!972] EMC_FREQ 1%!!2750 GR3D_FREQ !![548,548,548] NVENC0_FREQ !!314 NVENC1_FREQ !!315 NVDEC0_FREQ !!315 NVDEC1_FREQ !!314 NVJPG0_FREQ !!314 VIC off OFA_FREQ !!314 PVA0_FREQ off APE 300 cpu!!43.531C tj!!45.062C soc012!!43.156C gpu!!45.062C soc345!!41.562C VDD_GPU 3152mW/16918mW VDD_CPU_SOC_MSS 6699mW/17696mW VIN_SYS_5V0 7112mW/16891mW VIN 22936mW/61788mW

01-15-2026 03:13:59 RAM 96386/125772MB (lfb 1x4MB) SWAP 0/2048MB (cached 0MB) CPU [2%!!972,0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,13%!!972,0%!!972,3%!!972,0%!!972] EMC_FREQ 1%!!2750 GR3D_FREQ !![314,314,314] NVENC0_FREQ !!314 NVENC1_FREQ !!314 NVDEC0_FREQ !!314 NVDEC1_FREQ !!314 NVJPG0_FREQ !!314 VIC off OFA_FREQ !!314 PVA0_FREQ off APE 300 cpu!!43.281C tj!!44.812C soc012!!42.968C gpu!!44.812C soc345!!41.281C VDD_GPU 2364mW/16760mW VDD_CPU_SOC_MSS 6305mW/17572mW VIN_SYS_5V0 6897mW/16782mW VIN 23508mW/61372mW

01-15-2026 03:14:00 RAM 96387/125772MB (lfb 1x4MB) SWAP 0/2048MB (cached 0MB) CPU [4%!!972,7%!!972,0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,1%!!972,0%!!972,0%!!972,3%!!972,0%!!972,2%!!972,2%!!972] EMC_FREQ 1%!!2750 GR3D_FREQ !![314,314,314] NVENC0_FREQ !!314 NVENC1_FREQ !!314 NVDEC0_FREQ !!314 NVDEC1_FREQ !!314 NVJPG0_FREQ !!314 VIC off OFA_FREQ !!314 PVA0_FREQ off APE 300 cpu!!43.156C tj!!44.656C soc012!!42.937C gpu!!44.656C soc345!!41.156C VDD_GPU 2364mW/16605mW VDD_CPU_SOC_MSS 6305mW/17451mW VIN_SYS_5V0 6908mW/16676mW VIN 23602mW/60966mW

01-15-2026 03:14:01 RAM 96385/125772MB (lfb 1x4MB) SWAP 0/2048MB (cached 0MB) CPU [0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,6%!!972,0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,4%!!972,0%!!972] EMC_FREQ 1%!!2750 GR3D_FREQ !![314,314,314] NVENC0_FREQ !!314 NVENC1_FREQ !!314 NVDEC0_FREQ !!314 NVDEC1_FREQ !!314 NVJPG0_FREQ !!314 VIC off OFA_FREQ !!314 PVA0_FREQ off APE 300 cpu!!43.062C tj!!44.531C soc012!!42.687C gpu!!44.531C soc345!!40.937C VDD_GPU 2364mW/16454mW VDD_CPU_SOC_MSS 6305mW/17333mW VIN_SYS_5V0 6897mW/16572mW VIN 23442mW/60567mW

01-15-2026 03:14:02 RAM 96386/125772MB (lfb 1x4MB) SWAP 0/2048MB (cached 0MB) CPU [0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,5%!!972,0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,0%!!972,6%!!972] EMC_FREQ 1%!!2750 GR3D_FREQ !![314,314,314] NVENC0_FREQ !!314 NVENC1_FREQ !!314 NVDEC0_FREQ !!314 NVDEC1_FREQ !!314 NVJPG0_FREQ !!314 VIC off OFA_FREQ !!314 PVA0_FREQ off APE 300 cpu!!42.843C tj!!44.312C soc012!!42.468C gpu!!44.312C soc345!!40.781C VDD_GPU 2363mW/16305mW VDD_CPU_SOC_MSS 6305mW/17217mW VIN_SYS_5V0 6908mW/16470mW VIN 22572mW/60167mW

01-15-2026 03:14:03 RAM 96382/125772MB (lfb 2x2MB) SWAP 0/2048MB (cached 0MB) CPU [3%!!972,8%!!972,1%!!972,0%!!972,1%!!972,3%!!972,0%!!972,0%!!972,4%!!1080,0%!!1080,2%!!972,0%!!972,0%!!972,0%!!972] EMC_FREQ 1%!!2750 GR3D_FREQ !![314,314,314] NVENC0_FREQ !!314 NVENC1_FREQ !!314 NVDEC0_FREQ !!314 NVDEC1_FREQ !!314 NVJPG0_FREQ !!314 VIC off OFA_FREQ !!314 PVA0_FREQ off APE 300 cpu!!42.812C tj!!44.343C soc012!!42.406C gpu!!44.343C soc345!!40.875C VDD_GPU 2363mW/16160mW VDD_CPU_SOC_MSS 6699mW/17107mW VIN_SYS_5V0 7010mW/16372mW VIN 27286mW/59825mW

01-15-2026 03:14:04 RAM 96379/125772MB (lfb 2x2MB) SWAP 0/2048MB (cached 0MB) CPU [9%!!972,4%!!972,0%!!972,0%!!972,3%!!972,0%!!972,0%!!972,19%!!972,2%!!972,0%!!972,0%!!972,0%!!972,4%!!972,0%!!972] EMC_FREQ 1%!!2750 GR3D_FREQ !![1169,1169,1169] NVENC0_FREQ !!314 NVENC1_FREQ !!314 NVDEC0_FREQ !!314 NVDEC1_FREQ !!314 NVJPG0_FREQ !!314 VIC off OFA_FREQ !!314 PVA0_FREQ off APE 300 cpu!!42.843C tj!!44.5C soc012!!42.312C gpu!!44.5C soc345!!40.687C VDD_GPU 4333mW/16038mW VDD_CPU_SOC_MSS 6699mW/17000mW VIN_SYS_5V0 7213mW/16277mW VIN 26434mW/59480mW

01-15-2026 03:14:05 RAM 96371/125772MB (lfb 3x2MB) SWAP 0/2048MB (cached 0MB) CPU [2%!!1350,12%!!1350,4%!!972,0%!!972,0%!!972,2%!!972,2%!!972,18%!!972,9%!!972,0%!!972,3%!!972,1%!!972,13%!!972,2%!!972] EMC_FREQ 2%!!2750 GR3D_FREQ !![746,746,746] NVENC0_FREQ !!314 NVENC1_FREQ !!314 NVDEC0_FREQ !!314 NVDEC1_FREQ !!314 NVJPG0_FREQ !!314 VIC off OFA_FREQ !!314 PVA0_FREQ off APE 300 cpu!!43.593C tj!!44.218C soc012!!42.187C gpu!!44.218C soc345!!40.562C VDD_GPU 3939mW/15915mW VDD_CPU_SOC_MSS 6699mW/16895mW VIN_SYS_5V0 7213mW/16185mW VIN 26570mW/59145mW

01-15-2026 03:14:06 RAM 96371/125772MB (lfb 3x2MB) SWAP 0/2048MB (cached 0MB) CPU [1%!!972,5%!!972,3%!!972,0%!!972,0%!!972,4%!!972,16%!!972,2%!!972,2%!!1566,4%!!1566,5%!!972,1%!!972,0%!!972,0%!!972] EMC_FREQ 1%!!2750 GR3D_FREQ !![323,323,323] NVENC0_FREQ !!314 NVENC1_FREQ !!314 NVDEC0_FREQ !!314 NVDEC1_FREQ !!314 NVJPG0_FREQ !!314 VIC off OFA_FREQ !!314 PVA0_FREQ off APE 300 cpu!!42.625C tj!!43.906C soc012!!42.468C gpu!!43.906C soc345!!40.5C VDD_GPU 3152mW/15786mW VDD_CPU_SOC_MSS 6699mW/16792mW VIN_SYS_5V0 7112mW/16093mW VIN 25470mW/58804mW

01-15-2026 03:14:08 RAM 96373/125772MB (lfb 3x2MB) SWAP 0/2048MB (cached 0MB) CPU [13%!!1836,4%!!1836,0%!!972,0%!!972,0%!!972,3%!!972,3%!!972,2%!!972,0%!!972,5%!!972,0%!!972,0%!!972,0%!!972,0%!!972] EMC_FREQ 1%!!3200 GR3D_FREQ !![1205,1205,1205] NVENC0_FREQ !!314 NVENC1_FREQ !!314 NVDEC0_FREQ !!314 NVDEC1_FREQ !!314 NVJPG0_FREQ !!314 VIC off OFA_FREQ !!314 PVA0_FREQ off APE 300 cpu!!42.75C tj!!44.125C soc012!!42.031C gpu!!44.125C soc345!!40.437C VDD_GPU 3939mW/15667mW VDD_CPU_SOC_MSS 7093mW/16695mW VIN_SYS_5V0 7112mW/16003mW VIN 26490mW/58481mW

[AastaLLL]

Hi,

Could you share the GPU utilization with us as well?
This can be found in the below command:

$ nvidia-smi

Thanks.

[maspien]

Mon Jan 19 17:20:08 2026
±----------------------------------------------------------------------------------------+
| NVIDIA-SMI 580.00 Driver Version: 580.00 CUDA Version: 13.0 |
±----------------------------------------±-----------------------±---------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 NVIDIA Thor Off | 00000000:01:00.0 Off | N/A |
| N/A 43C N/A 3W / N/A | Not Supported | 1% Default |
| | | Disabled |
±----------------------------------------±-----------------------±---------------------+

±----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
| 0 N/A N/A 3475 G /usr/lib/xorg/Xorg 411MiB |
| 0 N/A N/A 3701 G /usr/bin/gnome-shell 199MiB |
| 0 N/A N/A 5251 C python 14005MiB |
±----------------------------------------------------------------------------------------+

[AastaLLL]

Hi,

It looks like the loading is quite low.
What kind of issues/alerts do you get?

Thanks.

[maspien]

I keep getting the over voltage warnings and throttling messages. I’m getting those over current readings I posted in the first message.

For LLMs it processes 2-4 tokens a second even on a small model. I get over currents through out the response.

Image generation takes up to 20 seconds 1024x1024 with medium quality. I get over currents towards the end of the generation.

It looks like the instant draw of power is the problem. Using the stock 140W power supply over usb c

[AastaLLL]

Hi,

Could you share more details about the warning?
Based on the log you shared, it is an overcurrent event.

Based on your log, the maximum GPU clock can reach 1430 GHz, so the instant power is still high.
So it triggers overcurrent handling (lower the clocks) to protect the device.

Thanks.

[maspien]

When the over currents happen a warning message pops up and states an over current event has happened and the system is throttled. Also where you can select the various power levels it changes from showing the selected power to a red exclamation point that states warning. When I click acknowledge the warning, a window opens say that there was an over current event. When I select it, the fault clears it’s self.

When I’m trying to generate the an image the fault only shows up once per requested generation. When I send a message to a LLM it faults multiple times through the response. I will follow up with some screen shots.

[maspien]

[maspien]

Is there an update to this? I’m trying to get an rma, and they want me to get help on the forum first.

[maspien]

Is this still being looked into?

[AastaLLL]

Hi,

Sorry for the late update.
Just clarify that this is an over-current warning instead of over-voltage.
This is expected since when the GPU is running the LLM tasks, the loading increases, and the “instant” current can be high.
The mechanism is designed for handling this (via reducing the clocks) to avoid damage to the board.

Although the LLM frequency is low, when the task is triggered, the GPU loading is still high (the GPU will try to run the kernel ASAP).
Maybe you can try to lower the clocks if you don’t need the device to run in performance mode.

For example, below is the command to set the device to the 90W mode:

$ sudo nvpmodel -m 2

Thanks.