Topic: [NemosMiner] multi algo profit switching NVIDIA/CPU miner - page 86. (Read 289464 times)

How can i add a custom mineur into nemosminer? I want to change bitcore miner to spmod2, if anyone knows...

EDIT : .... my bad not the same issue

Also decided to try with 2 cards, there seems to be the point is not that different cards behave differently on different algorithms, but the fact that every single map will process the various algorithms, which at launch was the most profitable. We run the first card (interval 15 minutes, not 24hr.bat), after 7.5 minutes, run the 2nd card with the same interval. Accordingly, at one point in time, 2 cards will process 2 different algorithms, since they are constantly changing... but, this is not accurate and horseradish knows what will come of it =))

In the file .ps1 does not need anything edit, just run the programs (.bat) from other folder, the client (miner) yourself determines the desired port.

This can happen when the miner "default" intensity is to high, look at the miner log for the command line arguments, open a command prompt window, cd to the bin folder where the miner is located (ie: Bin\NVIDIA-Alexis78Phi) paste the command line from the miner log and see what intensity value is used. (ie: -i 19, etc)  and observe in afterburner how long it takes for all card to rev up, if takes to long, lower the intensity 1 point at time ( 19 -> 18...) until you can find the right intensity for your system. you can then edit the .ps1 miner file and add this setting.

Hope this not too confusing.

edit start.bat change -Interval 60

default 60 = 60 seconds before checking miner hash rate you can change this to any number

eg:

-Interval 90
-Interval 120
-Interval 240

To illustrate the "very long miner warming up time" on 2.4
exemple, 3 to 7 minutes, switching on blake2s
https://reho.st/self/8c5d4c0610c44a985c139fb8e22cf175ff5df858.jpg

link from whatsapp group miner names, to discuss ideas.
 https://chat.whatsapp.com/A9jrZSkJ0aW0Qbl0NAIOWY

Does anyone have this working for Linux?

I also can report that ccminerKlaust is faster on Neoscrypt for me than PalginHSR.
6,75MH/s vs 6,48MH/s on 6x MSI 1070ti Armor with settings 70/+200/+650

Same problem here
I lost 3 to 4 minutes sometimes when the miner switch.
Only 1 or 2 GPU works, and after many minutes others comes alive

how is this calculating profit?
yesterday it had me mining timetravel for 1.5 hours 0 blocks found that whole time so no profit at all
and by the time the block was found my share ratio was nothing
and today the same thing on phi
on a long block coin finding a block means anyone mining up to that point is making more profit
but when u switch miners there after the block has been found and move them b4 another that is a huge problem
ive actually been finding i make more profit benchmarking than when profit switching is running
9/10 nist5 pays out more than nemos miners current most profitable
essentially when round robin beats profit switching something is wrong

actually i think im going to just remove all long block algos from the miner and try that
essentially phi timetravel xevan and prolly skien
skien runs a bunch all the time and makes dust even with 10x 6 card rigs

I also have the issue with "Very long miner warming up time" as reported in https://github.com/nemosminer/NemosMiner-v2.4.1/issues/94 .

Alexis78 miner in nemosminer-23, but also polytimos and Alexis78Phi in nemosminer-241, have often this behavior on algo X17, Skein, Phi, blake2s, nist5.
Only 1 or 2 miner threads/gpus are starting (out of 6) and after some long minutes (sometimes up to 10) the other gpus get full load. No errors are reported in miner log.

Thats a really big performance drop. You closed the issue on github, but it seems that the problem still exist.

Searching Google on gpu limit nemosminer brought me to this page: https://github.com/nemosminer/NemosMiner-v2.4.1/issues/165

Hope it helps

Equihash -> Skein -> Neoscrypt -> Lyra2v2

Is there a max limit on here?

Tried running benchmark but it just stops working after the first algo (11 gpu). My solution is usually to set one gpu, bench, then set the others, works but think it throws the estimates off

TIA
J

Hey Sol,

What is being suggested at the github page by another user is that the miner is probably using different GTX versions in the same RIG. With only the same cards is probably not really giving an advantage. Could explain why you didn't see any real gains while testing. Also to have a proper compare you should run similar rigs at the same time. But most of us dont have this luxury (yet)

where do I set only a gpu on the bench, in which file?

Hi everybody.
New to the forum here but long time lurker.
I would just like to thank everybody for their creative and experimental flair when it comes to mining. I have learned  a great deal from you all.  Thank you everybody that has been doing side by side testing and thank you nemo for your continued development of your miner.

I hadn't felt the need to register until a few moments ago. Being relatively new to mining I felt I wouldn't have much to contribute and I had also found that any questions I had were already answered some where on the net, so rather than to spam the network and fill up the forums with previously answered question I had decided rather to use my own initiative and search when I was in need of information.


Late last nite I came across the following "issue" on the github for nemos miner
https://github.com/nemosminer/NemosMiner-v2.4.1/issues/192
jjude80 commented 4 days ago
Through my experimenting, I have found that running nemosminer in multiple sessions (one for each card) staggered ,each with a 15 minute cycle time, is the most profitable way to run it. That way you get your share percentage up all the way per algo, and you don't loose out on another algo if it goes most profitable. I have doubled my daily usd by doing this. However it is irritating to edit each miners ps1 file each time a new version comes out. I propose being able to change the api ports used via the bat file to make it easier. that way someone unable to figure out how to edit the ps1 file properly can do this, and so that I dont have to have several full nemosminer folders configured.
https://user-images.githubusercontent.com/18168364/35174067-fce0f74e-fd3b-11e7-9ae6-4d79a5e2679e.png
The difference in price on that graph is quite staggering!


Having seen the potential increase in profits I attempted to replicate the results.
To begin with I bench marked a single gpu then I duplicated the folder 7 times and assigned a different gpu to each instance  and changed the interval to 900. Running for 12 hours I'm seeing an estimated 0.00416 btc/day which is actually less than i was recieving on ahashpool24. I have 7 1080ti FTW3s' MSI afterburner is set to core+50 mem+400 power80%.

My questions are:
Which ps1 files specifically is he editing and how is he editing them? (ps1 for each individual miner?)
Does anybody have a greater understanding how to "change the api ports" to achieve the same results?

Thanks Everybody for being Great
SOL

Yeah i understand, but how much can price drop affect amount of mined coins? Im mining still with same speed so amount should be same and only autoexchange can be affected by price i guess (or am i wrong?). Can it be that much difference?

Anyway today i received new batch of GPUs so i can do test with two rigs at same time.

Agree, you can't compare two different weeks of mining. Too much up an down currently, you can't compare two pools this way