Broken Master Branch GPU Support

[kev1n]

I've been trying to mine an efficient contract address with create2crunch. I followed all the instructions found in: this blog post, with little luck.

I used the following configs:

• nvidia/opencl
• 3090s/4090s/3070s/etc
• sudo apt install build-essential -y; curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y; source "$HOME/.cargo/env"; git clone https://github.com/0age/create2crunch && cd create2crunch; sed -i 's/0x4/0x40/g' src/lib.rs
• export FACTORY="0x0000000000ffe8b47b3e2130213b802212439497"; export CALLER={MY_CALLER_ADDRESS_INSERTED_HERE}; export INIT_CODE_HASH="0x640c5a7534c78cec935e63288b849fdfb379a57f064c1f013884d721830e8bc5"; export LEADING=5; export TOTAL=7; cargo run --release $FACTORY $CALLER $INIT_CODE_HASH 0 $LEADING $TOTAL

Here are some of the things I've tried:

• Varying combos of GPUs and CPUs
• One GPU vs multiple GPUs
• Choosing different GPU targets (i.e. 0, 1, 2, 3,..)

But the program would just hang on this message indefinitely
Setting up experimental OpenCL miner using device {}...

None of these worked, so my next idea was to go through the commit history until I found one that worked.

Cloning Vectorized's create2crunch fork, I tried the following branches:

• optimize: stalled
• optfixes: worked (ran within 1-3s)
• optfix2: worked (ran within 1-3s)
• master: stalled

I'm not sure if this is just an issue when on vast.ai or a similar GPU renting platform, or if something I just did wrong like not waiting long enough, but I hope this helps.

@Vectorized

[MRHarrison]

I also have the same issue, but it is intermittent. It will sometimes work or sometimes be stuck with the message below. I'm also using vast.ai

Stuck on:

Setting up experimental OpenCL miner using device

[kev1n]

Have you tried the other branches? They tended to fix the issue for me @MRHarrison

[MRHarrison]

@kev1n yeah I used optfix2 on vectorized clone. That works! Thanks!