its hit or miss with the low end cards i have some that
have to be set to 1875 to be stable , best ram is samsung
Topic: Eth mining stability issue (Read 387 times)
RX nitro 4GB with hynix memory, is it possible that 1100 and 1900 is actually to much for some cards to handle? Last year I bought like 30 GPUs and every single one was doing better than this without any issues, rigs running for weeks without a single crash. In January bought some extra GPUs and can't get it running stable (crashes from a few minutes up to 6h top). It isn't undervolting issue because left it at +25mv (original). It isn't other hardware related issues (exchange every single part of rigs). Was playing for 2 weeks with bioses, but now used an original one and only changed strap and mem and gpu core, but even a stock bios with speed of 22,5 MHs is doing same thing.
What is strange I have found out that GPU run longer when they are connected to SATA power which is "closer to PSU", as soon as exchange it to the other end crashes are more often or even instant. This doesn't make much sense to me but I have done bizilions of tests and this look to be actually working.
Mining ETC on linux ubuntu, using same pool. All previous rigs running with exactly same settings, hardware without issues, all new GPUs give me a big headache. Anyone had similar problems?
What is strange I have found out that GPU run longer when they are connected to SATA power which is "closer to PSU", as soon as exchange it to the other end crashes are more often or even instant. This doesn't make much sense to me but I have done bizilions of tests and this look to be actually working.
Mining ETC on linux ubuntu, using same pool. All previous rigs running with exactly same settings, hardware without issues, all new GPUs give me a big headache. Anyone had similar problems?
Jump to: