-
Notifications
You must be signed in to change notification settings - Fork 246
Open
Description
I am unable to deploy an instance using nvidia rtx3090 using the latest Quai-gpu-miner.
Reverting the quai-gpu-miner-run docker image (see the commit) resolves the issue.
But I think with Quai mainnet miners need to be running the v0.5.0 quai-gpu-miner-run image.
Symptom with the current template - console log of instance fails CUDA test on RTX 3090
Log:
[quai]:
[quai]:
[quai]: quai-gpu-miner 1.2.4+commit.0e7d056b
[quai]: Build: linux/release/gnu
[quai]:
[quai]: CUDA Error : forward compatibility was attempted on non supported HW
[quai]: Error: No usable mining devices found
[quai]:
SDL yaml hardware:
compute:
quai-miner:
resources:
cpu:
units: 1
memory:
size: 8Gb
storage:
- size: 1Gi
gpu:
units: 1
attributes:
vendor:
nvidia:
- model: rtx3090
ram: 24Gi
interface: pcie
Workaround:
- Manually edit the yaml to revert the miner docker image from v0.5.0 to v0.3.0. NOTE: It's not clear if it will mine a block given QUAI launched their main net this week.
quai-miner:
image: quainetwork/quai-gpu-miner-run:v0.3.0
Metadata
Metadata
Assignees
Labels
No labels