ghostplant

Maybe you can consider if drop-less MOE mode can solve your issue, which is achieved by setting `capacity_factor=0`

Do you have more information? I didn't get what you said. Outputs from different GPUs: ``` STEP-10: loss = 21.11541, step_time = 3.628716 sec, perf = 0.08 tflops. [Summary] Average...

```sh $ lspci | grep Displ 0e:00.0 Display controller: Advanced Micro Devices, Inc. [AMD/ATI] Device 740c (rev 01) 11:00.0 Display controller: Advanced Micro Devices, Inc. [AMD/ATI] Device 740c (rev 01)...

[dmesg.txt](https://github.com/RadeonOpenCompute/ROCK-Kernel-Driver/files/7102542/dmesg.txt) Attached.

```sh $ dpkg -l | grep rock ii rock-dkms 1:4.3-59 all amdgpu driver in DKMS format. ii rock-dkms-firmware 1:4.3-59 all firmware blobs used by amdgpu driver in DKMS format $...

```sh $ dkms build amdgpu/4.3-59 -k $(uname -r)/x86_64 Kernel preparation unnecessary for this kernel. Skipping... Running the pre_build script: checking for a BSD-compatible install... /bin/install -c checking for gcc... gcc...

``` $ ls /var/lib/dkms/amdgpu* 4.3-59 kernel-5.11.0-27-generic-x86_64 $ ls /var/lib/dkms/amdgpu/4.3-59/5.11.0-27-generic/x86_64/log/make.log /var/lib/dkms/amdgpu/4.3-59/5.11.0-27-generic/x86_64/log/make.log ``` [make.log](https://github.com/RadeonOpenCompute/ROCK-Kernel-Driver/files/7109079/make.log)

@kentrussell Is it related to improper BIOS settings?

```sh $ ls /lib/firmware/updates/amdgpu aldebaran_mec2.bin carrizo_mec2.bin hainan_mc.bin navi10_asd.bin navi14_sos.bin polaris10_k2_smc.bin polaris12_mc.bin renoir_mec.bin tonga_sdma.bin vega12_mec2.bin aldebaran_mec.bin carrizo_mec.bin hainan_me.bin navi10_ce.bin navi14_ta.bin polaris10_k_mc.bin polaris12_me_2.bin renoir_pfp.bin tonga_smc.bin vega12_mec.bin aldebaran_rlc.bin carrizo_pfp.bin hainan_pfp.bin navi10_gpu_info.bin navi14_vcn.bin polaris10_k_smc.bin...

Do you know how to check if current SBIOS version is OK and which version to update?