Mining ERGO with Vega 56 & 64

[ss-silver]

Switched my Vegas over to ERGO a couple days ago and it was hard to find good OCs so wanted to create this as place to collaborate and find the best OCs together.

Current Max hash Settings:

• 56 Samsung strap*:
  amdmemtweak --CL 20 --RC 36 --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 12 --RCDWR 12
  *GPU 5 is a flashed 64 running at RC 37 due to invalids at 36
• 56 Hynix strap:
  amdmemtweak --CL 20 --RC 36 --RP 11 --WR 14 --CWL 8 --FAW 12 --RAS 22 --REF 65535 --RFC 239 --RTP 5 --RRDL 6 --RRDS 3 --WTRL 9 --WTRS 4 --RCDRD 13 --RCDWR 12

Good hash, lower wattage settings
-SoC set to 860 in Over clocks

Getting Started

• OS version: 0.6-203@210519 or higher required

• Use TRM for the miner, v.0.8.3 required READ THE TUNING GUIDE FIRST!!!

• Ergo Wallet, I use Coinex because 40% of erg is currently traded on Coinex, so a little easier to sell higher quantities. You can use this referral link, when you sign up and get 5% of all transaction fees as kicks backs from Coinex.

Lets Talk Power, ERGO can use less power than mining ETH with very efficient settings (core 1225 / mem 847) and still preform at a good hashrate. However If you have plenty of room to increase wattage on your PSU(s) and breaker(s) you can get very high hash rates, comparable to the 3080 hash rate on ERGO with core 1400-1500 and mem 960-990.

ERGO mining observations

• Increasing core has the biggest impact on hash rate.

• Pushing core up to 1400+ seems to require about 900 vdd (give or take 25).
  This is alot of power, GPUs 1, 2 & 3 below are on a Server PSU with just those 3 cards and their risers, pulling 730w from the wall!

• If you get GPU dead or on boot the card does not increase wattage and start mining its one of two issues VDD is too low or core / mem setting is too high. I follow the TRM guide increasing up to 925 and if it crashes I lower Core.

• Hash rate moves around alot, my understanding is every time a block is solved and a new job is sent to the RIG hash rate drops temporarily while the GPUs build some table?? but then returns to the full hashrate.

• Hynix cards and 64s flashed to 56s will need about 40 more core for same hashrate, 64s can run at lower watts and lower vdd.

Questions? Better OCs?
Please provide a screenshot like the one above so we can all see accepted rate and uptime on new OCs. Also makes it quicker if you have issues or questions since OS version, Miner Version, HiveOS image, OCs & GPU brand/memory can all be seen from this page.

[Smining570]

Thanks for sharing, ill try ergo with my 7 vega´s rig later again, tried once but could´t get vegas to mine… What was expected earnings during the test with that hashrate / consumption with ergo ?

[Smining570]

so i managed to get one 580 to mine ergo on nanopool, cant apply such high core as u have on your settings. Happy to get it working anyway, have been trying for several days…for now with 1 5xx card, next the whole 5xx rig and later with vegas allso.I wouls like to have oc profiles and flightsheets saved for autoglygos 2 for both rigs for use later.
Cant get to my head those high core rates…so ergo is more core than mem…how high core u can go with vega 56 and how high mV it can take ? Same guestions about polaris cards…im gonne give it a try with mix of 590´s 80´s 70´s just need to find right oc for 570.
##EDIT… no luck with whole rig,must be that 570´s need some other values and i didnt managed to get rig mining.##
Good topic, hope we get more data to get this algo in control

[ss-silver]

So 7 vegas would do a little more hashrate than me and here is my 24 hour earnings with 1300 MH.
For the Vegas specifically its making a over $6/day each right now, so like 2x ETH.

Did you try higher VDD on the 580? mine always runs at 950 vdd no matter what i put there. Could also be a TRM thing that allows me to run that higher.

Ive noticed that average hash rate on the vegas is -4mh from reported. And funny only my 3 cards at 200+ ran at 100% for the past day

[Smining570]

managesd to get rig mining, didnt last for long…i cant afrord to play with ergo values now, too strong FOMO!! must mine eth while still can ill be following and maybe give a try with one card every now and then…interesting new area of tuning…like there is not enough´t with eth straps and oc´s

[bonezy]

Thx for this @ss-silver, nice to see @Smining570 here too, he sure knows his stuff

Lucky me to find this post, I just switched my 2 measly vega’s to Ergo 2 days ago.
On my first attempt, I was using my usual ETH timings and straps that gave me 54-55 and 52-53 respectively on ETH and I think I was getting something around 100-110.
I then came across a YT video from a turkish guy that had pretty similar timings to yours and I am now getting ±200 on each.
Mining to herominers now, they look quite organized and I send the ERGO to gate.io for the time being. I’m making one ergo per day with these 2 cards, so that’s around 12-13$. Might give woolypooly a shot though too.

Cards are drawing way more power than before and running hotter, probably due to the high core clocks.

I noticed the reset in speed too. Every minute or so, speed resets to 185, then picks back up again, didn’t know if it was a problem with the overclocks or something.

Also tried nanominer and srbminer multi but they seem to give me much lower hash rates. Nanominer’s rates are in the 70’s and SRBM’s are in the 100-110 with the same clocks.
I see people using NBminer but I can’t seem to set it up on Hive.

EDIT: Forgot to mention that my 3080’s on another rig give me around 210-220 MH/s each @ 230W. I tried it but left decided to leave them on Ezil for the time being.

Hope this post helps us sort out all this, looking forward to the response.

[ss-silver]

Nice those are great results! I know what you talking about with the 200 - 185 in hashrate. If you put a monitor on the rig and watch TRM, after each block is solved a new job is sent and the GPUs build a table, hashrate drops after building the table, then goes back up. Over the long run this results in a 2% loss on the average hashrate.

[Smining570]

I was playing with different values with one vega 56 / TRM /Woolypooly.
I does seem that its around 960-970 mem is good, didint made any big difference rising mem value, tried up to 1000, only seem´d to lower the hash.
Core 1440-1470 gave best results with vdd 940-970, didnt check from the miner for actual values, next time… but even higher vdd didnt produce more hash, as long there is enought mV for core its okay and it looks like 940 is enought, for this card anyway.
Tried few differen strap values allso, RC 36 made biggest difference, few hash more, have to try out some other starp combinations later.

[ss-silver]

@Smining570 Nice work that’s awesome!! just changed RC to 36 on the samsungs for an extra 2MH on two of them, the one set to 1400 core, hashrate didn’t change.

The flashed 64 hit 203 but threw 3 hardware errors in the first 90 minutes, so got to roll that back to 37.

Tried RC 35 but wound not mine with that setting… going to try out your OCs on the two that can handle RC 36 now.

[Smining570]

im sure we can push consumption down a bit with right straps and settings combo…lots of testing but imho its doable ! i noticed that there is a trshold in core 1400 / 1420-40, so little increase in core…i would like to know more about your findings about vdd not changing…could we force hive to use state 7 for core & 3 for mem with dpm 7 and mdpm 2 ?

edit…well now i saw from your pic that u had those on one of your vegas allready did it made any difference??

[ss-silver]

nope no difference, that one is still at DMP 4. But I dont think it really matters we’re already getting more than the high end for the TRM guide, and they are not crashing. Mem state is set to 3 on all the cards with mem 930-970.

Also got your settings to work on my Rog Strix 56, actually pushed that core up to 1480 and hit 207 MH tried dropping vdd 1 setting, to 931 but it crashed after 2 min, so think your right about pushing mem over 960 needs 940+ vdd.

Running this strap on that card:
amdmemtweak --CL 20 --RC 36 --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

The reference 56 hash rate dropped when I pushed mem over 1440, and in the miner saw that it reverts back to 1269 when I put any value over 1440.

[Smining570]

@ss-silver

Mem state is set to 3 on all the cards with mem 930-970.

okay, so that explains why i didnt got any benefits from higher mem…970 it is then.

@ss-silver

The reference 56 hash rate dropped when I pushed mem over 1440, and in the miner saw that it reverts back to 1269 when I put any value over 1440.

so thats it then…we are close to limits and optimal setings with mem 970 core 1470 vdd +940, i had it at 960 but im not so sure is that required, could be that a bit lower core and vdd will results better outcome…i havent payd attention too much to consumption yet…thats one variable that needs to be optimal level compared to max hash allso, right?
Good work with getting us all to right direction with these ergo values,i actually dont know how profitable mining this coin is, but i defo want to have my oc template for this algo ready to go when needed

[ss-silver]

ya we should consider wattage… eventually lol

Got the reference 64 (GPU 5) over 207 mh with some different settings

amdmemtweak --CL 20 --RC 37 --RP 11 --WR 14 --CWL 8 --FAW 12 --RAS 20 --REF 65535 --RFC 248 --RRDL 6 --RRDS 3 --WTRL 9 --WTRS 4 --RCDRD 12 --RCDWR 12

[bonezy]

So, is the constant table regeneration and drop in hash rate normal?
I didn’t seem to notice it with Nanominer and SRBMiner or maybe I’m wrong.

Tried your latest straps and they work just fine, thanks. I am thinking of lowering my core clock on both cards though so that I can drop my vdd even further. I can’t go lower than 900 without crashing and I don’t like riding my cards so hard.

Anyone know why there is such a huge disparity in the hash rate you can get on TRM, Nanominer and SRBMiner with the same clocks?

EDIT

I finally switched to lower OC on both cards. They are now both working under 130W and at temps I feel comfortable with. Anything over 1300 on the core will require more VDD and for some reason 847 on the ram works better on my strix than 960.

Another question that comes to mind is if you know if TRM has a B-mode for ergo mining? Can’t see it anywhere in their tuning guide.

[bontho]

Good day!

I’ve been reading a lot since the beginning of this thread and, first of all, I’d like to thank you all for the time you take to share your knowledge.

Now, I’m knowledgeable in IT but not that much with anything related to mining. I’m trying the same clock speed, mem, and other values like the ones you’ve provided and it turns out my GPU is crashing less than a minute after applying the values. I have a Vega 56 with Hynix and a Vega 64 with samsung. Both of those cards are running their original BIOSes.

Should I run the strap prior changing the clock, vdd and voltage?

Here are my settings. It has been running without crashing for 24 hours:

Thank you!

[ss-silver]

Hey that Asus Strix card with Hynix is a tough one, I used to have one and it was very picky about straps or would crash.
This is the strap i was using for that card which worked

amdmemtweak --CL 21 --RC 36 --RP 13 --WR 13 --CWL 7 --FAW 12 --RAS 23 --REF 65535 --RFC 248 --RTP 6 --RRDL 5 --RRDS 3 --WTRL 9 --WTRS 4 --RCDRD 20 --RCDWR 14
Just apply the straps in the over clocks so they apply with the OCs

For the Vega 64 it works best if you flash a 56 bios, that will reduce power and make it more stable. I use this bios for the reference 64. Once you flash the 56 bios the settings and strap I posted for GPU5 should work on that card.
This dude wrote a good guide for flashing the vega 64s

[ss-silver]

TRM is just better its the same with ETH, can only get the highest hash for vegas on TRM.

I’m not sure how normal the issue with table generation is, tried nanopool last night and it was same. Might be a TRM thing, would be cool to find a fix for that but I dont know where to start.
Block time has like doubled since difficulty went way up then everyone moved back to ETH so ERGO network hash dropped 4 Th in a day. So since less blocks are being solved this problem is not as bad right now.

[bonezy]

Thanks for the reply mate.

Well, I figured TRM is better, I just have never seen an algo that is SO much better hahahaha.
Why have peeps switched back to ETH? I haven’t noticed any big rise in profitability. I mean, I have my other cards on ETH still cause I want to hodl the ETH but when comparing hash rates on WTM, vega’s just seem to be more profitable on ERGO now.

Also, asking again in case any one knows. Is the a B-Mode for ERGO on TRM?

[ss-silver]

there is no mode for Autolykos2 in TRM
these are the only things you can put in miner cofig.

todxx/teamredminer/blob/master/USAGE.txt

Eth Profits have not changed much but ERG profits dropped quite a bit in the past 2 days since hashrate skyrocketed and difficulty went up accordingly. I was mining 4 ERG/day last week, cant even hit 3/day with that hash rate now. So last night Eth became more profitable to mine with a 3070, and probably several other cards. Vegas are still more profitable on ERG.

Vegas:
ETH 55Mh = $4.16/day right now
ERG 200Mh = $5.31/day right now

3070s:
ETH 62Mh= $4.69/day right now
ERG 170Mh = $4.52/day right now

[bontho]

Damn man! Just gained 20 Mh instantaneously! Thanks for sharing that strap! I’m now at 168! Hopefully it will be stable.

Just out of curiosity, how much hashrate did you get with the Asus Strix Vega 56?