Topic: Megaminer 6.3 - Multi pool / Multi Algo launcher - page 13. (Read 70970 times)

you can just rename your current gpu groups, then it will have to benchmark them again.
For example I use a group with 3x1070 (devices 0,1,2), if I want to game on my card that has a display on it, i just copy a different config.txt where I have a group named 2x1070 with devices 1,2

EDIT:
that said, I was also wondering about what to do when I change the overclock on my cards, I sometimes run them at 64-75% power limit (when I sleep I want them quiter), and sometimes at 75-89% PL, so hashrate changes a little, so when downclocking it's more probable it will change algos more often, while when upclocking, it will change algos less.

I think a better solution would be to delete your current benchmarks and re-bench everything when you make a hardware change.. That way you would have the most up to date accurate info out there...

I see a problem with the recently added always updating "benchmark" results: when you add or remove GPUs after you've done the benchmarks the results get messed up, as the currently running algos will change in values, whereas others will not. especially when adding GPUs this will move the current top algo(s) up ahead above the others so they never run again and therefore also don't update their results.

maybe an option to disable the updating after benchmarks could be added?

@tutulfo: About the new feature you are developing.

I think that it will be great if MM can check and apply the best settings for a card. Currently i mine with less power and some overclocked settings (1080 Card), however some algorithms/coins can handle this better as others. Should be nice if MM can handle this fully automated, for me it is difficult to determine the best settings for all coins.

Awesome thanks! Saves me the trouble of comparing the pools.

Could i trouble you to drop hash refinery and compare ahashpool and nicehash? I'm sure a lot of people would love to know.

I tried compiling your poolapitocsv but i got errors (i must admit i can't code or use the computer to save my life  ) else i would have done it

BTW, Could megaminer implement this function (logging wallet total to csv)?

so this is the continuation of that graph (starting points are where the other one ended)

https://i.imgur.com/wDFRX0E.jpg

So after a little more than 24 hours, I can tell that ahashpool is still the most profitable, but the differences aren't as great as I thought.
We now see some small drops on ahash as well, I think those are caused by actual exchange rates not being as high as I initially predicted, and these should probably not apear when mining with no autoexchange.

About comparing estimated vs actual profit, that is a bit harder to do because it has to factor in your hashrate. While MegaMiner does this, it just shows instant estimated income, and it doesn't log anything.

For me to include it in the tool I've made, I'll need to interface with ccminer as well, and use its api, it's a lot more work.

Great work!...Can you add estimated profit?, I need to know difference between estimated and real profit.

From my experience time ago, ZPOOL was a lot less profitable than AHASPOOL, but this can change.

I've been mining with Megaminer on those pools too and made the same observations. Some time ago I made an excel file with json call every 5 minutes to get the estimates from the different pools and also have been mining on them. Despite of the inflated rate, ahash still gives the best results for me. Maybe you can find out more, looking forward to it. One idea I had but havn't started yet is to calculate the actual profit with the current nethash from the explorers of coins and an api from exchanges, circumventing the use of pool stats. Guess this could be really useful. For example, you could jump on the right coin before the pool actually writes the estimated profit, giving you more shares compared to the people who jump on the coin based on pool stats

dude215p, thanks for sharing. I'm trying to confirm your findings but it will take me a little more time.

I've been wondering myself what the actual rates are for different pools.

I was mining only on Ahash pool, and I realized that their estiamations are inflated. Especially when comparing with Hashrefinery, even for algorithms that are not 'hot', all estimations where 20% higher.

I have a small rig with 3x1070 and right now I'm running an instance of MegaMiner on each so I can compare between Ahash, Hashrefinery and Zpool.
I'm running into different problems comparing the pools though. For example it took a lot of time for Hashrefinery to give me some XVG for the x17 I minned because it's a smaller pool, and there where only about 100 other minners for x17.

Comparing actual vs estimated is even harder as the estimations vary wildly.

I've also made a small monitoring bot to access pools' API every 2 minutes and write results to CSV file. I can then open that in excel and make some graphs over real profit made.

For now, Ahash pool still seems the best, but this experiment will need to run for a lot longer to be sure.

Let me know if you guys are interested in this, or have any other suggestions, and maybe I'll post some numbers a few days from now, when more data accumulates.

EDIT:
here's the tool I made:
https://bitcointalksearch.org/topic/pool-api-to-csv-tool-2783884

EDIT2:

I have a 3x1070 Rig. I've run 3 instances of MegaMiner one for each card, pointing to Ahashpool, Zpool and HashRefinery

From this morning, this is what I have until now:
https://i.imgur.com/HK7k1XT.png

I find interesting that both hashrefinery and zpool sometimes the total earned ammount drops a little, while ahash only goes up, and is more consistent, you get many smaller increases, while zpool has some huge jumps (one of those might come later for hashrefinery too). We'll see in a couple of days if we find out anything else.

Right, you can now use MPH in 24h mode

A little confused by what you mean. So in Mining_Pool_Hub.ps1 I should change this line from this:
To this:
and this one:
to this:
Is that right?

MM (as its name dont indicates... ) is "only"  a launcher, when miner is running has no control over it, this means same parameters launch inside or outside MM you will get same results.

Alexis-miner uses high intensities but actuals are tested in 1060,1070,1070ti and 1080ti and works, check your overclock settings, also you can try lower intensities editing miner file or even delete miner file (I think all alexis algos are covered by other miners).

I'm trying your miner out and im having issues with it.
Several miners keep crashing for me.
I'm running Windows 10 Pro version 1709
6x GTX 1070s

CCminer gets force closed, to be more specific its the alexis78-ms2013-cuda7.5 version that gets forced closed, happens sometimes, and sometimes it starts up and runs fine.
I've tried updating to the latest alexis version with cuda 9 but keeps crashing at times, ill try stick around and see the error next time, because i cant find any logs for ccminer just for megaminer.

Suggestions what could cause it?

Its benchmarking at the moment.

It will be off by default, profit correction factors are required by some people also as independent feature.


If you mine always as much as possible, you can be losing money, example:

*Power Limit 120
----Power cost: 10$
----Revenue: 100$
----Profit: 90$

*Power Limit 90
----Power cost: 8$
----Revenue: 95$
----Profit: 87$

If you mine limited always you can be losing money, example

*Power Limit 120
----Power cost: 10$
----Revenue: 110$
----Profit: 100$

*Power Limit 90
----Power cost: 8$
----Revenue: 100$
----Profit: 92$

"Right" power limit is a combination of revenue, power usage and power cost.

No video, sorry, you only need to change "BASIC OPTIONS" section and run START.BAT or edit any of .BAT examples.


"Noob" questions are welcome, They are usefull to upgrade config.txt instructions.

Do you have any full video how to use this megaminer 5.2 only full batch script that i downloaded and edited the config file but the problem i don't know if what should i change inside the config..

I think the idea is to just mine as much as possible as long as its profitable (higher revenue than power cost).
Which would be a good feature to add, but the proposed feature, in my opinion, just adds unnecessary complexity.

clarity, if I do this you can think you are mining with 24h data but is no real.

Modify pool file to allow this is really easy, you only need set profit24h to same value than profit on final set and change $ActiveOnAutomatic24hMode=$true

I need help for next feature....

I can announce next feature I am working (I think is great), from a releases ago I am making changes to get this.

The feature is automatic power-limit calculation, you can define in config txt what points of Gpu power limits you can test (for example 110, 100,90 and 80) and MM will set Power limit to this points and benchmarks power and hashrate to get more profitable PL point based on actual electricity cost and renue per hash for each algo.

For example in my test my cards run at 110 at night (electricty is cheaper) and 90% at day for algo I´m testing.


This feature has 3 flaws:

- Requires at lest 10 minutes of benchmark for eac algo/miner/power limit combination
- For Nvidia, MM must run as administrator (I´m looking for easy way to do)
- Pools revenue estimations needs to be adjusted to reallity, here is where i need your help, I have defined pools revenue correction factor (on config.txt) but I usually mine on Ahashpool and I can´t estimate other pools factors.

For example Ahaspool estimations are 30% over real revenue and Nicehash are 2% over, What factors (based on your experience) must set for other pools?