CUDA error DRIVER: '701' in func 'Solve' line 153

[Speeddymon]

Hi, I installed SWARM custom miner from GitHub - MaynardMiner/SWARM: Profit Switching Mining Administrator For HiveOS/Linux & Windows: HiveOS Integrated and configured it with nicehash, and it is working well for the ethash algorithm with nbminer.

Now I am wanting to take advantage of the other algorithms that can be used with nicehash and so I started looking at the output from the miners in swarm. In there, I see a lot of cuda errors like below when bminer is running and these algorithms don’t show any hashpower from the benchmarks while ethash does.

I ran nvcc -V and I can see that my cuda version is 9.1, but I read here that the latest drivers (which I have 470.86 installed) are for cuda 11.4. Edit: I also read that these drivers supposedly support cuda 10.x, but I haven’t seen any indications that cuda 10.x is available on the system anywhere so far.

Is there a guide somewhere for how to install cuda 10.x or 11.x in 211205?

[INFO] [2022-01-03T23:47:18+02:00] Bminer: When Crypto-mining Made Fast (v16.3.7-d9521d5)
[INFO] [2022-01-03T23:47:18+02:00] Watchdog has started.
[WARN] [2022-01-03T23:47:18+02:00] Failed to detect platform
[INFO] [2022-01-03T23:47:21+02:00] Starting miner on 5 devices
[INFO] [2022-01-03T23:47:21+02:00] Starting the management API at 127.0.0.1:40001
[INFO] [2022-01-03T23:47:21+02:00] [D0] Starting miner for NVIDIA GeForce RTX 3090 at 0000:01:00.0
[INFO] [2022-01-03T23:47:21+02:00] Connected to kawpow.usa.nicehash.com:3385
[INFO] [2022-01-03T23:47:21+02:00] Subscribed to stratum server
[INFO] [2022-01-03T23:47:21+02:00] Set nonce to 3c255b
[INFO] [2022-01-03T23:47:21+02:00] Started miner on device 0
[INFO] [2022-01-03T23:47:21+02:00] [D1] Starting miner for NVIDIA GeForce RTX 3090 at 0000:03:00.0
[INFO] [2022-01-03T23:47:21+02:00] Authorized
[INFO] [2022-01-03T23:47:21+02:00] Received new job 00000000a058c8d5 period #696334 target 0000000b33b85b7c
[INFO] [2022-01-03T23:47:21+02:00] Started miner on device 1
[INFO] [2022-01-03T23:47:21+02:00] [D2] Starting miner for NVIDIA GeForce RTX 3090 at 0000:05:00.0
[INFO] [2022-01-03T23:47:21+02:00] Started miner on device 2
[INFO] [2022-01-03T23:47:21+02:00] [D3] Starting miner for NVIDIA GeForce RTX 3090 at 0000:06:00.0
[INFO] [2022-01-03T23:47:22+02:00] Started miner on device 3
[INFO] [2022-01-03T23:47:22+02:00] [D4] Starting miner for NVIDIA GeForce RTX 3090 at 0000:07:00.0
[INFO] [2022-01-03T23:47:22+02:00] Started miner on device 4
[INFO] [2022-01-03T23:47:25+02:00] Cached kernel for #696312
[INFO] [2022-01-03T23:47:25+02:00] Cached kernel for #696336
[INFO] [2022-01-03T23:47:25+02:00] CUDA error DRIVER: '209' in func 'ReloadModule' line 41
[FATA] [2022-01-03T23:47:25+02:00] Fatal cuda error in GPU 3. Terminate soon...

I can also only find 9.1 related libraries in the OS and no instructions anywhere on how to update the cuda libraries for hiveos in specific.

I did try to upgrade cuda using ubuntu instructions but ended up breaking the miner and had to do a hive-replace to reinstall the os.

Any help is appreciated

[Speeddymon]

I did a clean reinstall via hive-replace as I mentioned earlier, after trying to upgrade cuda using instructions for ubuntu, so everything is working as originally described (only ethash, with nbminer, other miners crash with cuda errors)

So to be clear, I have 3 questions.

Question 1) Does HiveOS 0.6-212@211205 have CUDA 9.1 like nvcc -V shows, or does it have CUDA 10.x like the Hive website says, or since Nvidia says that the 470.82 driver comes with 11.x, does Hive have that?
Question 2) Regardless of which version I have installed, let’s assume its the right one, what is causing the error?
Question 3) This is the opposite of question 2. Let’s assume this is outdated and needs an upgrade, how can I do that?