Bladebit GPU plotter - beta

Thank you for your comment. But I have read all the posts of this subject which is over 600, made keyword searches and even read the other threads but no way. I also googled and searched github and reddit with no luck. There is no read.me file among the bladebit assets or github comments.
I need a help from domebody who successfully setup bladebit_cuda plot check. I know that I can check plots by “chia plots.exe plot check” but this way I think I can gain some time.

Here you go, I just scrolled up the page, and it is also shown on GitHub, as that’s where I got it from.

Bladebit is very fussy with the command syntax, get wrong and it won’t work properly.

1 Like

This string works
bladebit_cuda -f fffff -c ccccc -n 1 --compress 7 cudaplot --check 100 --check-threshold 0.8 g:\plots\

1 Like

Thank you. Yes that works now.

2 Likes

Hi, i have a similar system. Dual Xeon 2690v2, 512GB DDR3 1600 RAM, GTX 1080 TI, Nvme SSDs, Win Server 2022, and Chia 2.1.1 with Bladebit 3.1
I use a Supermicro X9DRi-F Board
My problem is that plotting C4 using Bladebit Cuda plotting takes ~22mins! :confused:
Compressionlevel doesnt matter, C0 is also ~22 mins
Bladebit RAM Plotting takes ~30 mins

Im not sure if something is wrong set in my Bios or if Windows is the problem

Any suggestion? Do you use ubuntu?

Update. It seems the problem was Windows. After installing Ubuntu Desktop 22 LTS and nvidia 535 driver the plotting with chia bladebit works out of the box in god damn 3,47 minutes per plot! Why…How??? Whats wrong with windows or is it bladebit which is not well optimized for windows?

Its both, Bladebit is not optimised for Windows, and Windows is slower than Linux at these sort of tasks.

Anyone know why i have this error every time i want to plot
STDERR: Fatal Error:
STDERR: VirtualAlloc failed.

Than i need to restart pc and then its working. After i finish ploting and change hdd , error appears again so i need to restart pc again. thank you.

Seems to be a common problem, only used to affect farming with older cards (10 series/Pascal) but seems to be affecting newer cards recently.

Best fix is to switch to Gigahorse.

1 Like