I wuold like to know if the cpu ryzen 5 3600 6x 3.6 and 16gb ram ddr4 2400 is good enought to multiple plots.
It should have 12 thread so how many plots it could make.
Thank you for helping I am trying to collect information about chia
The rule since chia 1.0.4 is 1 thread / 4.5GiB RAM / 258GiB fast temp, how much fast temp ya got??
Thank you for the answer.
So with that processor if I had necessary ram and disk space I could run 12 plots at the same time.
That is correct?
yes, however if you want to maintain optimal per plot performance (lowest time to completion) I’d suggest not going over 9 in parallel. For the 3900x and 5900x 18 seems to be the sweet spot, when I try to push to 21 each plot slows.
It is correct the suggestion to use ssd dc interface u.2 and not sata?
I am looking for on ebay but they are expensive.
What is your suggestion?
I’ve had great success with software (mdadm) raid-0 across multiple smaller m.2’s If you go over 2, you’ll need a special add-on card. I have an Asus hyper m.2 v4 (4x 2TB inland platinum’s)
what about the problem of tbw of the ssd m.2?
Thank you for your help
A good rule of thumb is absolutely no more than 1 DEFAULT (2c/4gb) plot for each core and thread, e.g.
|Cores||Threads||Max Parallel Plots|
|2c||0t||2 + 0 = 2 / 2||1|
|4c||0t||4 + 0 = 4 / 2||2|
|4c||8t||4 + 8 = 12 / 2||6|
|6c||12t||6 + 12 = 18 / 2||9|
|16c||32t||16 + 32 = 48 / 2||24|
Remember, these are absolute maximums, since 90% of plotting is CPU perf in my experience. Going higher than this is a very bad idea IMO!
This is also assuming you have the required disk space (~300gb) and memory (~4gb) per plot.
I don’t run this many plots on my machines typically… for example on my Ryzen (16c/32t) I do 12 parallel plots of 4c/6gb. I do, however, run 5 parallel 2c/8gb plots on a Hades Canyon nuc which is 4c/8t.
wait wait I am little bit confused, i am new here. lol
ryzen 5 3600 6x 3.6 it is a good start with chia?
Follow the formula in the chart based on how many cores and threads your CPU has … that’s the ABSOLUTE MAX.
so in my case 6 corse 12 Threads= 9 max parallel always assuming I have the required disk space (~300gb) and memory (~4gb) per plot.
r9 5900x (12c/24th) with 32gb ram (3600mhz) but look stuck a 1060mhz not sure need to check in bios
I did in parallele default setting:
-3390MB ram / 2 thread / 128 buckets
12 in parallel (6 on one NVME then 45minutes after 6 on other NVME)
7 hours later only 35% on 6 and other lower %
sure i did shit some where but where ^^
Looks like you might be hitting the RAM limits. You said you have 32 GB but running 12 plots. When you say you run the default setting, do you add -b XXXX for the memory or not? If you didn’t add -b XXXX for how many MB of memory you want, it would default to ~4.5GB of memory per plot. But even if you specified the -b 3390, that’s again more than 32GB of memory.
Staggering them like that with 45 mins should help on the CPU (IIRC stage 1 uses parallel CPU) but the RAM would be needed more or less throughout the whole plotting process. With that much RAM I think you should go 6 or 7 plots MAX with default ram of ~4.5 GB, lowering would slow down the plotting. Good luck
Look i am on win gui so i do’nt where I add -b for the memory. But yes I thinks you are right need upgrade my ram so for next time i will test with less plot in parallele
The GUI does actually default to 3389MB per plot.
The memory option under the advanced options in the plot creation screen is the GUI equivalent of the CLI’s
ok good infos in this forum !
I saw some prob too my ram is at 2133MHZ on win and it’s some 3600mhz
need check later on bios but now i am at 45% with 12 plots in 8 hours loool
will not stop it wait this night
That’s because the industry-wide standard for DDR4 RAM doesn’t specify 3600 Mhz as an available option, even though many RAM manufacturers and both AMD and Intel individually choose to support it, see: What is XMP? | Crucial.com
In most applications, RAM speed isn’t actually important past a certain point and I’d bet it doesn’t matter for plotting either. Don’t stop plotting to change this.
OK will sleep more smart this night for sure !! CHIA community power!!!
Bug chia stopped working so i changed the ram to 3600MHZ windows accept it.
So i restarted 3plot in parallèle on 1xNVME and 3 plot on other NVME
I am new to farming and I feel like I’m messing up big time.
I have a Ryzen 5800X, 64GB RAM.
I am currently plotting 12 in parallel with 12 threads. My max RAM usage is 6750 MiB, and I did not put a delay before the next plot starts.
I know I am doing something wrong because my plotting speeds are way too slow (all 12 plots are in 7% after about 8 hours).
Do you have any suggestions to improve this time. Also, how important is having a delay before the next plot starts.
I am open to any suggestions, thank you.