Topic: S9 performance inconsistencies - firmware related? (Read 163 times)
Ok I'll try downgrading. Do you know which firmware corresponds to S9V2.23? The filesystem is dated to Jul 21. Or is there a way to dump it directly?
...Ditto for my 14Th units from Dec....they do just fine with the Nov firmware and worse (High temps/more amp draw/less hash if rolled back).
You got lucky. Ive had to roll back well over 1000 s9s away from that terrible November firmware.
If you have 1117 firmware on your november batch and you think it is causing issues id roll it back. I have found numerous problems with the 1117 firmware and have removed it completely from my environment.
Yes & no.....i had success rolling back the firmware for a unit from Aug 17....but
My Jan 18 units absolutely do not benefit....in fact ...flashing the April 17 firmware causes the temps to jump 10c while freq and hashrate is exactly same as Nov 17 firmware..
So I back to the Nov 17 firmware on all of my units from post sept 17
My best luck is with the July 17 firmware but can't find a copy of it anywhere..
...Ditto for my 14Th units from Dec....they do just fine with the Nov firmware and worse (High temps/more amp draw/less hash if rolled back).
Since bitmain removed the april firmware from their site and the aws link is so slow, here are some alternates:
https://www.dropbox.com/s/v5r5n46eajetme4/Antminer-S9-all-201704270135-autofreq-user-Update2UBI-NF.tar.gz?dl=0
https://www.mediafire.com/file/986jp6m02svz7ya/Antminer-S9-all-201704270135-autofreq-user-Update2UBI-NF.tar.gz
https://1drv.ms/u/s!AtkfLUFCwpoBaWvM7d0-19mvBYo
You should confirm that the sha1sum is 0515029015ad8dcedb185367fc5f829851dc65ed before flashing
https://www.dropbox.com/s/v5r5n46eajetme4/Antminer-S9-all-201704270135-autofreq-user-Update2UBI-NF.tar.gz?dl=0
https://www.mediafire.com/file/986jp6m02svz7ya/Antminer-S9-all-201704270135-autofreq-user-Update2UBI-NF.tar.gz
https://1drv.ms/u/s!AtkfLUFCwpoBaWvM7d0-19mvBYo
You should confirm that the sha1sum is 0515029015ad8dcedb185367fc5f829851dc65ed before flashing
If you have 1117 firmware on your november batch and you think it is causing issues id roll it back. I have found numerous problems with the 1117 firmware and have removed it completely from my environment.
I have two 13.5TH S9s from the November 2017 batch, lets call them worker1 and worker2.
Worker1 is running firmware version S9_V2.51: it averages 13.63TH, temps around 80, will occasionally switch to a lower priority pool (when worker2 has no trouble reaching priority 0), and has its chips clocked to an average of 626.74.
Worker2 is running firmware version S9V2.23: averages 13.92TH, temps around 75, and has its chips clocked to an average of 641.25.
Worker2 is performing faster, cooler, and running an older firmware version than worker1. Should I flash worker1 with S9V2.23? Or just accept this as random hardware variance.
Worker1 is running firmware version S9_V2.51: it averages 13.63TH, temps around 80, will occasionally switch to a lower priority pool (when worker2 has no trouble reaching priority 0), and has its chips clocked to an average of 626.74.
Worker2 is running firmware version S9V2.23: averages 13.92TH, temps around 75, and has its chips clocked to an average of 641.25.
Worker2 is performing faster, cooler, and running an older firmware version than worker1. Should I flash worker1 with S9V2.23? Or just accept this as random hardware variance.
Jump to: