Help with Optimizing Config for Swar Manager

Hey, guys! I have the following rig.

  • AMD Ryzen 7 3800X (8c/16t)
  • 32GB (2x16GB) DDR4 3200MHz
  • MSI MAG B550 TOMAHAWK
  • 2x1TB SSD Samsung 980 PRO and 970 EVO Plus

Can you help with better settings to optimize the number of plots I can do per day and get the most out of my system, because I think I made little plots per day for this system.
This is the current settings that I use.

global:
max_concurrent: 6
max_for_phase_1: 3
minimum_minutes_between_jobs: 15

  • name: Samsung 980 PRO 1TB
    max_plots: 999
    farmer_public_key:
    pool_public_key:
    temporary_directory: D:
    temporary2_directory:
    destination_directory:
    - E:\HPOOL Plots
    - F:\HPOOL Plots
    - G:\HPOOL Plots
    - H:\HPOOL Plots
    - I:\HPOOL Plots
    - K:\HPOOL Plots
    size: 32
    bitfield: true
    threads: 2
    buckets: 128
    memory_buffer: 4000
    max_concurrent: 3
    max_concurrent_with_start_early: 3
    initial_delay_minutes: 0
    stagger_minutes: 60
    max_for_phase_1: 1
    concurrency_start_early_phase: 4
    concurrency_start_early_phase_delay: 0
    temporary2_destination_sync: false
    exclude_final_directory: true
    skip_full_destinations: true
    unix_process_priority: 10
    windows_process_priority: 128
    enable_cpu_affinity: false
    cpu_affinity: [ 0, 1, 2, 3, 4, 5 ]

  • name: Samsung 970 EVO Plus 1TB
    max_plots: 999
    farmer_public_key:
    pool_public_key:
    temporary_directory: J:
    temporary2_directory:
    destination_directory:
    - E:\HPOOL Plots
    - F:\HPOOL Plots
    - G:\HPOOL Plots
    - H:\HPOOL Plots
    - I:\HPOOL Plots
    - K:\HPOOL Plots
    size: 32
    bitfield: true
    threads: 2
    buckets: 128
    memory_buffer: 4000
    max_concurrent: 3
    max_concurrent_with_start_early: 3
    initial_delay_minutes: 0
    stagger_minutes: 60
    max_for_phase_1: 1
    concurrency_start_early_phase: 4
    concurrency_start_early_phase_delay: 0
    temporary2_destination_sync: false
    exclude_final_directory: true
    skip_full_destinations: true
    unix_process_priority: 10
    windows_process_priority: 128
    enable_cpu_affinity: false
    cpu_affinity: [ 0, 1, 2, 3, 4, 5 ]

A few suggestions, otherwise looks ok.

  1. Global phase 1 to 4
  2. More threads 3-4
  3. Start early =4

Then test timing between jobs. Global is 15min while within each job its 60min.

Bonus: over provision ram. It balloons up and down and never uses peak amount the whole time.

2 Likes

Thanks, I will try it.

A little update. Now I’m doing 21 plots per day which I think is good for that system. :grinning:

I’m still trying to get swar running.
Just curious why you have your cpu affinity as it is?
With 16 threads couldn’t you be using 0 to 11 leaving 2 cores for the computer to do its thing?

I haven’t touch cpu affinity. Like you see is on false command. I’m going to try now MadMax Plotter. :slight_smile:

1 Like

Ah, ok, from what I read cpu affinity is based on logical cores / threads.
So looks like your only using less than half your available cpu!

Hadn’t spotted the false, only been playing with swar for 24 hrs .

But your ahead of me, you got it to run.
I’d appreciate your feedback on how madmax works out for you.

I’ve an i9 11900k and your plotting more than me , so I definitely need to get a handle on this.

Best of luck to ya.

Yeah, I have a 16 threads like you. I’m running 6 parallel plots with 2 threads on each plot. That’s my final configuration for now if that can help you.

global:
max_concurrent: 8
max_for_phase_1: 4
minimum_minutes_between_jobs: 15

size: 32
bitfield: true
threads: 2
buckets: 128
memory_buffer: 3390
max_concurrent: 3
max_concurrent_with_start_early: 4
initial_delay_minutes: 0
stagger_minutes: 30
max_for_phase_1: 2
concurrency_start_early_phase: 4
concurrency_start_early_phase_delay: 0
temporary2_destination_sync: false
exclude_final_directory: true
skip_full_destinations: true
unix_process_priority: 10
windows_process_priority: 128
enable_cpu_affinity: false
cpu_affinity: [ 0, 1, 2, 3, 4, 5 ]

1 Like

I have 16C 32T, 64 GB Ram, 2TB WD Black SSD.

Below is my config…

name: micron
max_plots: 47
farmer_public_key:
pool_public_key:
temporary_directory: E:\Chia
temporary2_directory:
destination_directory: D:
size: 32
bitfield: true
threads: 4
buckets: 128
memory_buffer: 4700
max_concurrent: 4
max_concurrent_with_start_early: 6
initial_delay_minutes: 0
stagger_minutes: 120
max_for_phase_1: 2
concurrency_start_early_phase: 4
concurrency_start_early_phase_delay: 0
temporary2_destination_sync: false
exclude_final_directory: false
skip_full_destinations: true
unix_process_priority: 10
windows_process_priority: 32
enable_cpu_affinity: false
cpu_affinity: [ 0, 1, 2, 3, 4, 5 ]

2 Questions:

  1. WHy is my phase 1 plotting time increasing?
  2. Why is the utilized space on my SSD based on progress (135+209+158+153=665GiB) vs below it shows 1.46 TiB is used? I have been checking since morning and this mismatch exists.

Any optimization suggested please?