imo straps seems to be fine, i dont have -CL value and i have -RTP 6 RRDL 5 but dont think issue is there…sorry but cant really say what is the issue here. Are u running aircon or how your temps are so low ?. Have u tried to switch risers ? …If u find solution, please post and let us know what was the issue. good luck
mem 1020 will set ur SOC to 1028, i would stay with that.
Usually some Samsung´s may handel higher SOC but effiency goes down quite a bit then.
If u are happy with wattage and gpu is stable as is, imho just let it run with those oc´s …if VDD 810 is min required. If u want to lower your wattage, go down with core to 1030…35 and u still should get just shy 55 mhs
nice, it makes no harm to try 800…805, if u get crash just pump it back up. Under 120W(hive) is decent for that MSI so its up to u if feel like playing with it some more, when u go closer to limits with VDD you may get crash due temp´s rising a bit etc, stability is the key anyways.
Trial and error and u will find the sweet spot if not there yet
one thing to watch out for with vega mining ETH is as you lower VDD you may be stable and it will stick, but suddenly your stales will go up which you can only see poolside. One more thing to check before accepting the new settings.
has anyone tested wall power of ergo mining vs eth mining on a single card? Nvidia cards actually use less power on Ergo vs Eth but this is reverse on Vega. Ive been guessing average 150W ETH and 175W ERGO and wonder if anyone has looked into this. Thanks in advance
The fact is that ergo needs more juice on the core while eth needs only high memory clock. So you can go down with core which gives you less power consumption. This concerns all non-LHR cards (all AMD plus non-LHR Nvidia cards).
On LHR cards miners must cheat chip and pretend that it is not mining eth. That’s why power usage fluctuates and real power usage is higher than on ergo where miners can just mine with full power.
Hello,
i dont know what i keep doing wrong. My two Vega 56 still getting invalid shares. I tried a lot different OC and straps. Maybe someone of you have an idea. I’m using Phoenixminer.
I have a problem with vega 64 and vega 56, I used some settings that I found here on the forum and I got 54 mhs but a valid share appeared and some of the cards were interrupted, now I will tell you which cards are in order:
card “1: vega 56 with vega 64 bios” radeon "
card “2: vega 64 with vega 64 bios” radeon "
card “3: vega 64 with vega 64 bios” asus strix "
card “4: vega 64 with vega 64 bios” radeon "
card “5: vega 56 with vega 56 bios” radeon "
I’ll send you a picture below of my cards and settings in “amdmemtweak”
Mitrabrate, read the initial post and read the thread, you’re doing the opposite of what is recommended vega 64 → vega 56 bios not the other way around. Leave you vega 56 cards with original bios. All your ref cards can be flashed from hiveos with this bios: 016.001.001.000.008771
also try coreclock:1050, memclock: 1020
Core Voltage, mV: 830 (maybe more maybe less needs testing, all my cards using 780-810)
For BIOS Im not really sure don’t wanna tell you something I’m not sure. AFAIK all reference models with Samsung memory use AMD Bios so it’s all the same, make sure you Subsystem ID is the same on your card as the new bios