MadMax performance issue after Chia upgrade

I have a strange problem that started after I was forced to upgrade from 1.3 to 1.5, after a crash.
Everything came back fine, and I like the new GUI, but plotting with MadMax either directly though the GUI, or using the command line, takes three hours, previously it only took 1 hour.
I even tried running the old version of MadMax (not the one shipped with 1.5) with no improvement.
Everything else seems fine.
I am running Ubuntu 20.04.4, and Chia version 1.5.0. I have rebooted a number of times.
Any ideas appreciated.

Which version of madmax is included with Chia version 1.5?

Under Windows, running:
chia_plot.exe --version (two dashes preceding ‘version’)
– or –
chia_plot_34.exe --version

…will output your version number.

I am running Chia version 1.3.3, and it includes madmax version:
1.1.6-2092041

Via Linux’s terminal, the commands should be the same, except that your executables will not have the .exe extension.

What crashed (the OS, the full node, your plotting job, something else?), and how did that crash result in you being forced to upgrade your version of Chia?

Did you check that madmax version number?
The same madmax version might be bundled with different versions of Chia.

1 Like

What crashed (the OS, the full node, your plotting job, something else?),

Had a dirty powerfail, causing box, OS and Chia to crash (I now have a UPS!). After the crash Chia (was version 1.3) showed 1.8XCH in my wallet instead of 3.6. I decided to start from scratch.

I think I know know what the problems is. After some tests on a spare nvme drive, I believe my 1Tb nvme temp drive is knackered…

It has created about 1500 plots, smartctl shows:
Data Units Read: 3,396,295,038 [1.73 PB]
Data Units Written: 3,201,514,324 [1.63 PB]
Host Read Commands: 8,923,810,138
Host Write Commands: 8,043,621,609

Is this likely? Still working, but 3 times slower? I know nvme have a “limited” lifespan, but did not know what the symptoms would be.