Anyone having issues with the last few versions of HiveOS?
I have a 4x vega 4x rx580 rig that has been stable for months. Since upgrade i keep getting dead cards, lots of restarts. Now applying straps commands dont work, either does putting them in the OC section.
What should i try? Nothing else has changed apart from the hive version. There is no roll back available now either, there used to be a list of all previous versions i could choose, not i get nothing but the latest to choose from.
When i apply straps same way i have been for months i now get this:
Fail exit code: Cannot find DRI instance for pci:0000:09:00.0
like allway with upgrade, i allso had massive amount of invalids on both rigs that have been running without any for months, it was next night after upgrade that i started to have issues…i usually dont upgrade a working machine, but on the latest was some interesting fix or featrure that got me to do it.
I managed to roll back to previous but there was no list of older versions to choose like have been before.
Last time i did upgrade, i allso started to have dead gpu´s…only thing that helped was roll back with the miner…im using TRM and if i recall right i went to vers 0.8.0 and got rid of constant dead gpu´s untill the bug was fixed…but again…imho…in new releases there is often some issues 
Thanks, how do you roll back versions when it doesnt give an option? Start again with a new write to HD or what ever you are using?
Hive used to give me the drop down menu with about 20 versions, now i just have the current and nothing else.
Just changed 1 version back on tem red miner and working better already, maybe it was just TRM.
Straps working now, seems like the whole problem was the new TRM version which must have come as part of the overall upgrade.
super noob question, I can not manage to put HIVE OS to work after apply straps, first time after a install it works, but as soon as I try to improve it stops working and I have to redownload HIVEOS to usb
I am sure it is my fault, any advice will be appreciated
set low oc and default straps for first boot so u get rig mining… then try tune for better hash. if your oc is too high, hive wont start, it stops to the point when it cheking and loading your oc.
Hello,
The temperatures are currently rising. What is the maximum safe HBM2 temperature. 
The max HBM temps are mentioned in the bios on techpowerup:
So 95°C seems the critical temp.
The higher you go in temperature, the more difficulties your HBM seems to have with it’s timings.
At 60°C my cards have close to 0 rejected shares per week. At 70°C they will throw a few per day. At 80°C even more.
Some say also that the longevity of the HBM degrades with higher temperatures. Some say to keep it below 80, some below 70, some at 60°C. Lots of different opinions, no facts.
I think it is something that you’ll need to evaluate for you personal. Do you want to invest a lot of money to get their temperatures down? Or are you happy with a few rejected shares at high temperatures? Or are you more happy with lowering the hash rate by lowering the wattage and the clocks?
I’ve also repasted a bunch of cards which helped a bit in their temperatures; some (not all) went down 5 to 10°C. Here outside temperatures are now also +30°C so the cards are sweating a lot and I can’t get them below 70°C anymore, some are running even around 80°C. I can’t imagine what temperatures those cards would have been if I hadn’t repasted them!
So Vega 56s not going over 1028 is apparently a myth… just pushed my 56 Asus Strix to 1060 and its running fine
Thanks for busting the myth!
How long stable at that hash already?
30 min so far haha but Ill let it run for a day, I expect invalids. Will report back if it survives the hot part of the day (next 6 hours)
amdmemtweak --CL 20 --RC 37 --RP 11 --WR 14 --CWL 8 --FAW 12 --RAS 20 --REF 65535 --RFC 248 --RTP 5 --RRDL 6 --RRDS 3 --WTRL 9 --WTRS 4 --RCDRD 13 --RCDWR 12
alright ERG diff just dropped so gotta switch but after 8 hours about 1 invalid per hour
one of my ref 56 is at 85…87deg with fan 100% changed thermalpaste the other day… this one just seems to run hotter than other same ref… allso Wattage 10w higher. i dropped Mem and Core quite a bit but still +85deg…
Maybe you did it, but in case you didn’t: when you change the thermal paste, use the opportunity to clean inside the cooler block, blow some air through it in the opposite direction. Some came from dirty houses and are full of dust and sometimes blocked. If they are blocked they cool less. Also the fan itself might be full of dust. Every bits help.
It seems there is no difference between flashed 64 to 56 and 56 samsung cards. The difference is caused by silicon lottery.
Did you see my post with modding 56 ref to core water cooling + air memory cooling? This solves all the cooling issues with spending only ~50eur.
Got 2 56 cards running and been getting very frustrated with the amount of “GPU dead” messages - about one every 15 minutes on average. I took off all OC setting overnight and the cards ran stable but only at about 32 MH.
Earlier today, I applied the settings in this thread on one and tested it for about 3 hours… not one single GPU dead incident. I’ve now rolled it out to my second 56 and early signs are running stable.
I’m very happy! Now these old 56’s are doing better than my 5700XT!
Does anyone know if there is a permanent way to get HiveOS to keep the straps?
I do have a third 56 sitting in the box that I might replace my 580 with on this rig.
Привет! Как у тебя получилось запустить ферму на 8 картах? У меня максимум 7 работают хорошо, с 8 картой ферма постоянно перезагружается
You can just put them in the OC window in the Tweakers field.
