That’s said writing speed at 250M maximum. Usually should be more than 500M. That’s why the CPU always idle waiting SSD finish the job.
which ssd should i buy then ? any suggestion
Samsung 980 PRO SSD 2TB for speed
1 Like
i have the exact same problem, where can i talk to you? did you solve this?
Your settings are nonsensical. GUI is not ur problems. Setup and hardware are. With 2 x 1tb ssds (no raid) & if you are using 128 buckets, max memory set 4000 (more is not used), max threads set 4 (diminishing returns). No overclocking, this is not a game, u already have a nice fast CPU. With that AIO and 99degs, smthg wrong with cooler to cpu contact, maybe regrease/reseat carefully.
First, run 2 queues each ssd x how many reruns u want in queue. Wait until phase 2 finishes on the 1st sets started (several hours I’m guessing), repeat again for 2 more queues each ssd again (so 4 queue sets total, staggered). That should get u 8 done in no time, over and over. Even then, I’m saying ur not using ur CPU or memory or IO barely at all. Ur system is loafing along doing almost nothing, regardless of ur ssd brand.
So with that sweet setup, u need MORE ssds, perhaps another two for 4 total. Then with 4 ssds, do 1st run with 2 per each (8 going initially), 2nd 2 more per each after 1st setup P2 finishes (so 8 total queues to start, 8 later staggered after 1st ones end P2). It’ll work 4 u. If still strange, uv got bigger issues. Now you’ll have 8 plots every few hrs once it get going. Good luck!
PS. Use some heatsinks to keep those ssds cool.
Raid 0 is not work with plotting. Use separate drive and tasks. You will see huge improvements. And if your CPU is not using its own potential there is one conclusion - temp-drive bottleneck.
1 Like
RAID 0 works just fine for me and my two 2TB inlands on the 5900x
Im unable to see how you staggered those plots but I guess that whast hurting you. When all those plot get to Phase 2-4 and the trying to offload them to the final destination ( if any ) bottleneck.
Try and have 2 plots on each phase at a time and you will get 8 running, then if that goes with plots bellow 8hours do 60 min stagger between the sets.
- name: m.2
max_plots: 999
farmer_public_key:
pool_public_key:
temporary_directory: F:
temporary2_directory:
destination_directory:- D:\
- E:\
- H:
size: 32
bitfield: true
threads: 4
buckets: 128
memory_buffer: 4000
max_concurrent: 7
max_concurrent_with_start_early: 7
initial_delay_minutes: 0
stagger_minutes: 60
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: false
unix_process_priority: 10
windows_process_priority: 32
enable_cpu_affinity: false
cpu_affinity: [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23 ]
i am currently using raid 0. before plotting process i stress test all my components. they are working well. my cpu in cinebench with PBO max degree is 90 while plotting 60-80.
i tried different values but my last attemp above. after couple hours ill add some ss.
The issue with RAID is there is no TRIM on AMD. Therefore to TRIM, it’s necessary to break the RAID, make simple volumes, do TRIM, remake RAID again, and continue plotting. Painful and disrupts any plotting schedule with stagger.
Funny because my raid 0 array on my B550 board seems to trim in Windows 10 just fine?
It really depends on the controller and the os and the drives
seems 8h per plot finish time. nothing changed still low cpu usage and low writing rate on my raid. next step ill break raid and try 3 parallel same time dual ssd. i think it will not change anything bcuz a friend of mine has same ssd not using raid and plotting as hell one of his machine other 2 not as like his first system. as i said ALL SAME COMPONENTS has 3 system but only 1 working fine other 2 not good enough. i really dont get it how to fiz this issue. Also same cpu same MB with samsung 980 has low cpu rate maybe windows problem need to try linux or ubuntu too
I have my 5900x on a 360 aio all core overclocked to 4.7ghz. Chia loves fast single cores. That might help too. Your total plot times are not that far off from mine. My bottleneck is my SSDs as well. My 8086K does plots 45min faster but I can run as many in parallel, so the 5900x still gives me more TB/day.
What’s ur B550 MB, that’s nice if u can TRIM in RAID.