only while running windows with sub par memory specs (2gig ram)

you may possibly have run into a system compatibility issue? i would take your dmesg and any logged system errors to the amdgpu-pro mailing list.

possibly also consider using the pure open source driver stack individually instead of the entire amdgpu-pro package if you are up for that kind of thing.

good luck!

oh. you're probably right. my bad

yes, definitely sounds like a memory leak to me.

grab some paper towels and check your motherboard/ram slots particularly, hopefully its not so bad that you are leaking underneath the rig too.

once i had stacked a couple rigs, like bunk beds you know.. so i could utilize the same 24" box fan for 14 gpu's.. and the top rig memory leaked all over the rig underneath it.. huge mess.

who would have thought all that fluid fits in such small semiconductors?

man i wish ram just ran on smoke like cpu's do..

please post the temps and fanspeed with room ambient temperatures.. wonder if this gimmicky backplate does the aorus any good.

this is a SYSTEM ram issue as alluded to before

with 2g system ram you may mine for 24h or so but after things "settle" and programs are installed to run in memory, your mining software will eventually be unable to start and crash out with an ocl error.

upgrade to 4 gigabytes of system memory or switch to linux.

i like to think of it as being capable of printing legitimate apple stock for 30 cents on the dollar, back in 2000..

about some time after their second stock split.

run linux, it handles these things better.

aorus has the most elaborate hardware.. backplate with relief for the back of the compute unit.. but it's not clocked as high as some of the other cards that are the generic release.

i kind of expected more considering the aorus is a step up from the gaming.

guys if you are unable to mod and strap your own bios at least learn to search.

helping the community is one thing, enabling poor form is another.

also, never sell more than 2/3 at one time.. college smarts.

pm'd. good to hear it worked out. did you get all the coins you expected?

rep this guy followed through.

you are going to want to back up your wallet.dat and remove all traces of your old wallet.

then reinstall a completely new wallet and replace the saved wallet.dat before you fire it up for the first time.

some old pos implementations were really buggy and i have had to do this with many coins in the past.

sometimes even had to completely resync with a new/fresh wallet.dat and import keys from all addresses in the old wallet.

let me know if you need help getting all the old address privkeys, hopefully it wont come to that it's pretty time consuming.

anyone else getting payments delayed due to this code that is being exploited in the mined blocks?

https://github.com/zcash/zcash/issues/2333

if you have ip's in a file anyways i would go ahead and take action to notify the users.

maybe not the isp but see if a service is open on the ip's port and send a notification accordingly.

it would help discover if some unscrupulous devs are putting malicious hardware into software associated with crypto here.

getting more/less shares accepted is not indicative of better or worse performance (seeing more of one color than the other)

this is because each pool (for the most part) uses vardiff which adjusts the diff of a share needed to be solved for a successful relay to the stratum server and be counted toward your hash rate tally.

this is how things are generally balanced out where a single gpu owner submits shares at the same pace as someone that just got the latest asic hardware from bitmain.

you can specify independent ports to bypass this if your pool offers it but at your stage at mining i'd avoid it.

cooler temperatures are better so theres no issue there. sometimes at hotter temps a gpu throttles back and hashrate is sacrificed.

constantly reinstalling drivers has been known to be detrimental to performance when not properly removed each time.

i'd try a fresh format and reinstall of your OS and go from there.

also if you're going to stick to nicehash learn about how it works more.. if you're going to mine ethereum it might be better to mine it outright on like f2pool for a paid per share basis which is more consistent payout wise and then u can sell it for bitcoin when the price rises, possibly buy some back as the market dips.

once you get things working right i think you will enjoy it more.. good luck

none of the above. what is that garbage..

i think the last important altcoin development was in pascalcoin.. things have kind of gone south since in that camp tho.

in the future what we could use is something that can either prove a singular identity and distribute a share of the network accordingly, and/or prevent pool mining with the centralization and profiteering that comes along with it.

i know the terminology is muddy since nvidia coined the term "gpu" with their geforce 256 back in '99.. but these days it's generally agreed that the gpu is the entire card.

i have always specified the chip in question as the shading or compute unit and that is what the 7990 has two of (where the heatsink makes direct contact with to draw heat away) and is the issue in this case.

if you're comfortable with the possibility of losing the card altogether trying to fix the faulty compute unit you can try to reflow it in a toaster oven.. this will renew the solder connections and has been known to revive dead cards.

look up "bake gpu oven" for details, i also have a write up in a thread titled "repair old gpu's" or something where i go in to some detail.

i think one should avoid calling this die on the gpu a "gpu" in itself.. although its the heart of the video card and a "cpu" is correctly the equivalent in any computer system.. whenever i hear it referred to as a gpu i think of ignorant sales people referring to computer towers and cases as "hard drives"

good luck with your cards.

looks like you're not the only one.. zec one of the few coins actually thriving in this bear market.

with all its shortcoming, zec still seems to be much more viable than any of the competitors in this sector.

sweeping with a laser pointing ir gun is good practice now and again.. but i see and raise the concept with using an entire camera.



these solutions are pretty inexpensive as a stand alone product and can save your time in spades with just a short glance of the screen.

CAT came out with an inexpensive cell phone a year or so back that includes this option.. personally i like the seek thermal solution for android or iphone.

what were you clocking the card at?

I used to do a lot of grounding work and always had these large bus bars where soft aluminum connectors that I crimped myself were bolted to the copper surface.. it was *highly* recommended to use no-ox of some sort to prevent the buildup of any barrier that inhibits conductivity between the surfaces.



since I had an r9 290 plug melt down on me I picked up the habit of using this no-ox after taking a scotch brite pad the connections and have had zero issues since.

it's an exponentially more vital practice between two uncommon mating surfaces like aluminum to copper vs copper to copper.

as for the life of a connector clicking on/off cards, i would believe 20 cycles is a fair number.. but with some needle nose pliers and a bit of common sense you could reset that number and tighten up the clearances on the soft female connectors quite easily.