My 5950x been getting 1600 second which i am very happy with it BUT My 3970x on Madmax been getting 45 minute plus on both Windows 10 n Ubuntu (NOT Using Virtual Machine) I think that is something wrong with my MSI TRX40 Pro WIFI motherboard. It will be great if any 3970x owner with great results can share their motherboard spec.
I have try several NVME stick. Gen 4 n Gen 3 still getting same results. I set my Ramdisk to 112GB.
When i run on Windows 10. From task manager i notice my logical processor is only 50% running which is not under full load on stage 1. I have try many setup in bios but still cant run full load on stage 1.
I have the 64/128 pro and I cannot max out my cpu on windows…I currently have 4 running in paralell all using their own drive for temp/temp2…
On the flipside I have a 24/48 TR on ubuntu with 7 nvmes using Swar getting 79 plots per day…
Would love to try the 64/128 on ubuntu but I have to many USB storage drives on that rig(farmer)
You mean, you are running 4 Madmax parallels ? Your machine only can have max 256GB ram. So which mean you are running
1 NVME + ramdisk
2. NVME + Ramdisk
3. Nvme + nvme
4 Nvme + nvme
Yes i did try before running parallels But my 3970x , Job 1 took almost 1 hr to complete n Job2 took more then 1hr to complete. Job 1 Nvme + 110 Ramdisk n Job2 Nvme + Nvme.
i have try format my Windows 10 several time and the results is still the same. Madmax Nvme + Ramdisk 45 minute and SWAR MAX 62 plots in 24hr using 8pc NVME if more then 25 concurrent job then it will take more then 11hr for one task.
What motherboard spec u sing for 64/128?
My ram has not arrived so its all NVME. Each parallel process uses a separate NVME but the same NVME for both temp and temp2.
P1 - nvme1 - temp/temp2
P2 - nvme2 - temp/temp2
I have not tried Swar since installing the new CPU(was the 16/32 version). I am using the ASUS sage wrx80 Motherboard. Windows just plain sucks all around…
You can try this calculation , 8 core per Job which mean your 64C can run 8 concurrent job. I had a friend whom is 32 core had no problem running 4 concurrent plot. ONLY my stupid 3970x had a problem which i cant solve.
My 3970x plotting speed using 110G ramdisk is the same time i have on my i7 Nvme + Nvme 45 minute.
i had try replace ram n NVME but no luck so my last option its the motherboard.
I will have the same setup, but with the ASUS PRIME TRX40-PRO motherboard. My CPU will be coming in tomorrow, so I will let you know my results on Ubuntu.
Your feedback will be very helpful to me. As Asus Prime TRX40 Pro is the ONLY available option for me if i buy locally. Thanks in advance ya,.
Bro any feedback on your new setup
I’m seeing 20 minutes, before with chia-classic plotter I was seeing 5hr, but I staggered, so I was getting 16/day, now I’m getting 3*24 per day, so what
HDD storage space is the problem, the stuff is expensive now $60/TB, making plots was never a problem
Fast plotting has existed for GPU on burst & bhd for a 1,000 years, all of a sudden all these noobs to hdd mining think some thing is new, no its not, its just that team-moron @chia-dev made the decision at chia to give the community slow plotting sw that destroyed drives.
oh, 32 core amd, 64gb ram, 590 mobo, 3 nvme’s ( 2-mp600’s, 1-980 )
Sorry for the long delay. I couldn’t find the post again. I am average ~17 minutes a plot.