I’m going to check that now
Once the drives are added in the GUI (config.yaml), shut done the GUI, and you start GigaHorse (start_farmer.cmd) then run the Chia GUI and the drives and plots should be there.
As for not using drives A and B, I had a problem with that too but didn’t pursue it further.
To add to that, it also depends on the profit margins and has similarities with GPU mining : during bear phases you optimized for higher efficiency (lower power% and lower frequencies), and during bull phases you optimized for higher output (higher power% and higher frequencies). The smart move might be to have the capacity to replot and put more GPUs online for farming higher C values during high prices phases.
From back when I was using Windows, A: and B: are reserved for floppy drives and should not be used for HDDs.
1 Like
What happens if you have a mix of uncompressed plots, and compressed plots whilst replotting, and have the GUI open?
Obviously the GUI won’t farm the compressed plots, but will both try and farm the uncompressed plots?
just as wild guess, could you try to get rid of spaces in folders names? instead of quoted above go like “B:\hdd5”. also you use another slash symbol.
p/s sometime windows like to have double slash as well “B:\\hdd5”, also worth to try
i’m running pc with all letters mounted, including A,B. had no issues.
1 Like
It will farm the uncompressed plots. However you can simply start gigahorse and then the GUI, which gives you the same capabilities without any downside.
3 Likes
Hi ,
SO I solved the problem .
Thanks to drhicom , I went to config.yml and I saw that the differents path were not included in it .
Then I understood that Chia Gigahorse farmer use the path from there .
Therefore I Closed Gigahorse farmer / added the others path / open Gigahorse farmer again and it Worked !!!
Thank you for your answer .
P.S : You can also add the path using spacepool with the option ‘‘add directory’’ ( dont forget to close spacepool before open Gigahorse farmer ! ) .
1 Like
Does Chia/Gigahorse recognise subfolders now?
So instead of adding each subfolder I can just add the route directory for my mounted drives (Windows) and it will find all the plots.
Yes it recognize subfolders .
I had to had each folder .
Maybe There is a trick to do it in another way ?
If you enable recursive_plot_scan: true in config.yaml yes.
In harvester: section.
4 Likes
K33 adds 1.5% on profit compared to K32 at same C-level.
nice, i have also already figured k33/c8 + k32/c9 combo is the way to go
How can one calculate the maximum amount of plots that a single GPU can handle? On Max’s Github there’s a table with Plot Filter 256 for 3060 Ti/K2200/E2695, but how can I calculate the same data for my GPUs? Personally, I have two A4000’s.
And does using two GPUs on a single system double that amount?
However that will use all available GPUs, not a single GPU.
Yes, however the benchmark by default already uses both if available.
A4000 is a bit faster than 3060ti, around 10% more.
I feel really dumb, but after running
time ./ProofOfSpace lookup -r 8 -f *.plot
I get
supermicro@supermicro:/mnt/ssd$ time ./ProofOfSpace lookup -r 8 -f *.plot
operation: lookup
Caught exception: stoi
terminate called after throwing an instance of 'std::exception'
what(): std::exception
Aborted (core dumped)
real 0m0.148s
user 0m0.000s
sys 0m0.006s
I must be doing something wrong, but my limited knowledge of CUDA, Gigahorse and Linux prevent me from finding out what is not right.
You can only use one plot file as -f argument, don’t use the * unless it will expand to single plot only.
1 Like
Is Gigahorse GPU Farmer for windows (not WSL) still not available?