Quote from: Bare on January 26, 2018, 04:52:18 PM

Quote from: fanatic26 on January 25, 2018, 04:34:31 PM

I have a few thousand cards that run around 70c all day long. That is nowhere near the thermal limit and they will run for a long long time like that. I wouldnt worry about temps at all until they are breaking 80c

Quote from: Shnikes101 on January 25, 2018, 04:43:52 PM

Quote from: fanatic26 on January 25, 2018, 04:34:31 PM

I have a few thousand cards that run around 70c all day long. That is nowhere near the thermal limit and they will run for a long long time like that. I wouldnt worry about temps at all until they are breaking 80c

Agreed. Below 80c is fine for almost every card. However, we've found that Vega's will throttle when HBM gets too hot - even in the 70's can sometimes impact performance. Crazy, but it happens with these finicky cards was all I was trying to point out.

Good thing you said "almost" every card, because as you may or may not know every graphic card has it's own temp limits under a workload, older cards such as amd HD5000-7000 series, generally withstand higher temps limits better than newer ones because of the bigger fabrication process as opposed to today's newer cards which are made on much smaller fabrication process.

I know this because my first gpus I ever mined with were 2x hd5970, power hungry beasts back in the days, and guess what? 73-78°C core was nothing for them 24/7, mined for 2 whole years and later sold them in a fully working condition, bought one of my 390x some time later, died after 4 days of mining at those same temps.

few experts told me the same, it's the fab. process hence vegas and other new cards should be treated according to this input too...

Quote from: sumitc1989 on January 25, 2018, 07:04:20 PM

Quote from: Bare on January 25, 2018, 01:59:00 PM

70-75°C - 24/7??
and +15% plimit

I shed a tear when I read that ;( you're raping your cards...wearing them out hard...

I learned my lesson the hard way when one of my hawaiis died in the summer of 2016 while mining for 4 straight days around 70-75°C
thankfully only 1 card died... :/

Since then I'm taking card temps very seriously and precautions making sure my cards work in an adequate environment maybe even more than really needed just in case.

Don't ever run cards above at least 70°C core if you want them to live long and without any potential issues.
I usually run my fijis and hawaiis always around 63°C max.
right now in winter they are below 60°C at all times and that's with fans running at around 60-65% power
so yeah im obviously downclocking, -10% powlimit atm

you have no reason to be dissatisfied with the performance of your vegas especially in regard that you're running them 12 in a single rig

One of the rigs I have is a 6 card rig consisting of 4 fijis(nitro furys all) and 2 hawaiis, best performing and worst performing fijis differentiate about 4-5% in hashrates...it's pure silicon lottery man... there's nothing wrong with your cards hashing, if you apply 4-5% difference between your cards you get the same results you're complaining about.

on the other note, I noticed it's normal to have a little bit worse hashrate across all cards when running larger number of gpus in a pc than usual, although I can't confirm for sure by how much or at all... just my opinion.

so to answer your questions and few free tips:
1. Run your cards below 70°C since your clearly intend to have them working for 2+ years I suggest @ 63°C max. your cards will thank you in profits later as much as they live the longer the better

2. I don't own any vegas unfortunately but the general rule every serious miner applies is downclock/downvolt/tweak until you have stable temps, don't be driven by hashrates,
card temps are FIRST PRIORITY! then according to temps try to get as much hashes from your cards as possible!
Mining is a long-term game, remember that!

3. Probably, wouldn't know for sure, as far as I know less tweaks in wattman means more stable miner, set gpu settings in bat/config files.

Happy mining!

Thanx a ton for the advice bro, I was able to maintain my temps between 60-65 by cooling the ambience in ac. So basically are temps upto 65 degrees safe for the long run or yet you recommend to lower the temps by sacrificing the hashes (undervolting+underclocking) because i was able to achieve stable hashes of 2030+ with 65 degrees max under ac keeping the same old settings of +15% power.
However i am willing to bear the electricity charges for those tempting hashrates Tongue

keep the temps below 70°C(at low-mid 60's) at all times and you'll never have to worry about any potential problems, make sure your cards are properly ventilated(or a/c if you can but more electricity costs), make some more distance between them and lastly if needed sacrifice some hashrate (downclock/downvolt) whatever as long as you can maintain these "ok" temps!
what are few hashes more compared to the longer lifespan/health of your gpu?? is it really worth it? and to add, card scarcity is already bad as it is and you don't need yourself even more headache.

of course one more thing, never max your fan rpm, that's also dumb, you gain nothing setting your fans above 75% power(some people say 80%) than heavily wearing your fans out, so I suggest to set them max. 75% fan speed.

after you've made sure your cards are working in these conditions you won't have to worry about any component on your cards, occasionally clean that dust residue between the heatsink fins when you see it accumulating and pretty much that's it

thanx man that was really helpful

Will surely work my card temps and hashes accordingly as i too agree that life span is more precious then hashes considering the scarcity and gpu price rise.