Topic: [ANN] MM.HASH- Auto-Coin/Algorithm Hybrid Switching Miner (New Way To GPU Mine) - page 2. (Read 798 times)

When you say very many times, is it more than 3?

I recently changed benchmarking system to try to run algorithm for 3 switches before disabling. (Setting stat to 0). If is running longer than that, it may be a bug. I can test and address for next release, but this would be first report.

However, if you set the argument -No_Algo sib,keccak it should skip them.

Your oc seems fine, at least for a gtx 10 series. If you want to try to mine them, then adjust the intensity as I noted below. If not, then us -No_Algo and turn them off.

Let me know about the bechmarking, if its not stopping after third time of building database, or any other issues you may have.

here are my settings OC
Core: -100
Mem: 800
PL: 80
(Memory - HYNIX)
OS - HiveOS
very many times doing "sib" and "keccak". and does not switch to another algo

1.3.8 was really buggy, because I'm changing a boatload to integrate AMD into miner. I fixed it all with 1.3.9.

I actually have set up an AMD rig, so I can start setting up AMD miners. I just am clueless about AMD mining, so it is a slow process...But I am a quick learner.

Config Folder- nvidia-algorithms.txt. I just changed how that works, so I haven't updated github. I will do so by next version. If you simply don't want to mine an algorithm, you can either remove it from nvidia-algorithms.txt....Or you can use the -No_Algo argument (this means you can change it remotely). The full list of arguments are inside a .txt folder in the help file folder.

Also I wrote wiki on how to use hive commands.

I don't know how you can't pass benchmark though. I have a miner with 1050ti's, and they benchmark fine. It may be OC settings.

If you want to change intensities/add arguments to the miners- You simply edit the miner file for example:

"16r" = "-i 19"
"x16s" = "-21"

etc.

When you perform a remote update for a new release- Miner will transfer all your miner file settings, algorithm settings, and your initial benchmarks (backup folder).

t is impossible to pass a benchmark (sib, kessak ....). How can I edit the list of algorithms?
1060 3gb
V. 1.3.9

Looks great. Any ETA on the AMD, would love to take my Vegas for a spin on your miner 

We talked on discord.

You didn't screw up anything. I botched last release. I fixed it.

It's difficult managing two different platforms, but because you said something, I was able to find and fix...So thank you.

Upon startup...it got caught at the initial benchmarking. I was stuck benchmarking one algo (lyra2v2) for 35 mins before I called it quits. What did I screw up?

This seems cool. I think I'll check it out sometime soon.

Updated [ANN] to add the arguments needed for use and fix typos.

Updated [ANN] to reflect new version and system.

I updated original post for new beta release.

ChangeLog

1.2.9:

(ALL VERSIONS)
-Added New Zergpool Equihash Algos


(HIVEOS)
-Added 'active' command
 Type 'active' to view active miners in any terminal
-Add 'stats' command
 Type 'stats to view MM.Hash miner stats in any terminal
-Created Help Files Folder
-Log Integration with HiveOS
 Logs are now scheduled to HiveOS logging.
-Added Pools To Main Mining Screen
-Added Previous HashRates To Main Mining Screen
-Cleaned up Main Mining Screen
-Repaired Random Failures Of EWBF Hive Stats

-Added -Benchmark Parameter. Expiremental:

  -Benchmark [seconds]

   If miner hashrate equals $null, -interval time will
change to -Benchmark time. This only works for NVIDIA1.
Other device groups will still mine as usual, but will
have same interval as -Benchmark.

   I did not test fully. Report if it is working correctly
I will update wiki.



1.3.0:

(Windows & HiveOS)

-Zergpool https:/. Better API calls.
-Added Equihash Algorithms - zergpool and mph

(HiveOS)
-Ensured -Benchmark parameter working
-Fixed -Timeouts parameter
-Log rollover every 12 hours. Miner divides logs into 1 hour periods.
-PID read/write for better process control
-Discovered actual Powershell Core bug in (where).count -eq 0. It worked when it wanted to.
Reported to powershell, and then built workaround.
-Added Cuda-9-2 folder, will start filling with miner files for HiveOS cuda 9.2.
You must setup hiveOS to be able to use Cuda9.2 (Not reccommended)
-Added 10 second wait to give Miners time to spool up. Was noticing occasional lockup becuase
of too high resource use.
-Reduced hashrate display time to every 7 seconds instead of 5 for smoother display.
-Fixed Claymore API. However, noticed Claymore itself was displaying only 1 fan and 1 temp
during my testing occasionally. In future I will attempt to force defaults from Hive when this happens.
-Timeout Folder to notify when miners timeout. However, Timeouts no longer set all hashrates to zero. This
way miner can continue to try to mine that algo/coin. I simply reduces overall daily hashrate. The Timeout
folder will give you an idea of specific stats you may want to delete and then re-benchmark for fine tuning.
-Resolved Status flag issues. When typing "active", you should receive correct data.
-Resolved previous hashrate issues on screen
-Attempted to display better countdown times. NOTE: Internal countdown time varies by processor to processor.
However- Run time displayed for Miners is 100% correct.

1.3.2:

Truncate log stats for T-Rex, so user will know when miner isn't hashing in case MM.Hash fails
Attempted to fix multiple LogData screens
Migrated total t-rex log hash into try/catch argument to prevent on-screen errors (which is normal if hashrate isn't there).
Updated Miners
Migrated into a version setup, with remote upgrading capability- Details on full operation coming soon. (expiremental)
-Force paramter to clear T-Rex log.
updated Z-Enemy
updated T-Rex
Removed -Algorithms, as it was depreciated. Algorithms can now be changed by editing "Algorithms.txt" file
miner will self-install powershell
Added Date to stats command
Updated Help Files
Ensured All Miners Display Pool

1.3.3:

zergpool api updated.
for future implementation- zergpool_coin added
for future implementation- zerpgool_algo added
zergpool removed
blockmasters added (was not tested- Trying to push fast to fix zergpool)
starpool added (tested)
New command- "logdata" lets you see the stats recording miners that work from logs. Debug aid.
New command- "pidinfo" was always running in background- brought it to a screen so you can see MM.Hash process tracking. Debug aid.
LogData screen opens with every miner, is put to sleep if not needed (to preveny zombie app errors)
PIDInfo screens opens with every miner- Has always been running, it is just now visual
PIDInfo is now tracking MM.Hash process- For future implementation.
Added "clear" background process, work the same as miners- Double close command in case of hangs.
Small implementation to reduce resources.
Status flags written to disk, to provide double checking of MM.Hash states and its miners.

1.3.4:
-HiveOS Added hive integrated "mine" command. Enter to bring up NVIDIA1 screen (mining window).
-Added -No_Algo argument. -No_Algo excludes aglorithm of your choice. Must match Algorithms.txt.
Case Sensitive! (All lower case characters)
-EWBF coins added back in. Updated equihash to match zergpool.
-Reduced resources on PID system/tracking, removed read/write status- It was redundant.
-Newer streamlined benchmarking system. (Both versions)
-Better chance of miner catching and recording timeouts.
-Recorded hashrate displayed at benchmarking. Recorded hashrate is pulled from file (accurate).
-Enemy 1.15a
-CryptoDredge 0.8.3
-Started staging of much faster -Auto_Coin version. Not implemented yet.
-Started staging for T-Rex API Rest method hashrate recording (when implemented).


1.3.7 HiveOS Silver Lambo Edition:
-IMPROVED COIN MINING SYSTEM! Highly recommend to try -Auto-Coin "Yes"
-MM.Hash will not sort or download entire database, miner first sorts by most profitable algorithms, if the best pool is pool that has the capability of selecting a coin other than BTC: IT then chooses the coins of the algorithms of that pool.
-Stats will show coin list of a particular algorithm if coin mining occurs during that period.
- -Auto Coin switching is now around a 20-30 second process: which is 200% better.
-Fixed Powershell Core update bugs with a better workarounds.
-MM.Hash in background checks stats for possible timeouts, for accurate timeouts.
-MM.Hash now closes itself and all background windows within a few seconds of it stopping.
-Added -threshold argument default (internally) is set to .1 - If profit day shows to be .1 btc: MM.Hash ignores algo.
-Added back & improved active times for miners.
-Staged miners.txt. Does not work. However, you can see the direction it is going.
-This is beta version of new system.
-CPU Miners will now compile to your system. I included an experimental one that can work for ARM cpus. Note- Initial benchmark period is usually long for cpuminers. I would turn up -interval when benchmarking them. After initial benchmarks, you can set -interval to a much lower setting.
-hex added to algorithms.txt and files.
-fixed blockmasters pool file.
-fixed starpool pool file.
-pools initial stats are now estimate current. 24 hours was causing irregular initial stats. Improved 'stats' command and display.
-Enemy 1.16 added
-Code Staging for AMD

1.3.8
Important: -GPU_Count1 argument was removed. Miner now detects GPU's. You will get errors if used.
-Upated Windows Version To New Coin Switching System (Windows Miner Down- Could Not Test. Converting it to AMD rig. Please Report Bugs)
-Added -Favor_Coins argument. Default is "Yes"
-Combined Windows and Linux Miners into one folder sharing same files.
-Added Config Folder- Placed algorithms files there.
-Added \Config\update.txt
-Due to miner files being combined, you will have to benchmark. However, future updates will no longer your miner files. They will be copied over.
-Ensured T-Rex accepted/rejected are shown correctly on HiveOS.
-Lots of Code-Staging/Prep for AMD mining. Did not add AMD miners (having to change a lot to make it happen).
-Updated Algorithms and their placements to most optimal settings. Should see better hashrates on miners.
-Removed Alexis-Miner. It's depreciated
-Remove ocminer. It's depreciated
-Fixed threshold issues (will test further to confirm).
-Worked on better miner launch sequence
-Fixed myr-gr
-Added algorithms in case pools add them

1.3.9
-Fixed GPU array from double loading GPU index
-Removed extra pool files from github CoinPools folder
-Ensure -Favor_Coins was working.
-Miner will overwrite 1.3.8 versions, but transfer your miner configurations.
-Re-wrote benchmark parameter to test bench before writing stat, will write
message if testbench fails
-Failed Benchmarking should now require 3 failed benches in a row order to write
stat to 0
-Fix T-Rex Hive monitoring: Hashrates will now display correctly for each gpu.
-Added Sonoa
-Started Migration of cmdlets from main source code. Added some notations. Not
finished yet.
-Removed erroneous files from github.
-Changed Initialization procedure to a more fluid transision.

1.4.0a
-Windows Version Now Matches HiveOS version in core design
-Ensure 100% benchmarking is setting to 0 after 3 failed attempts
-Fixed some HiveOS stat issues from 1.3.9
-Updated Z-Enemy to latest version
-Noticed Times Were inconsistent on "active" screen: Attempted fix
-Removed Erroneus Messages involving closing screens (now that I know it works consistently)
-pidinfo now starts at miner startup, MM.Hash will always close within 5 seconds of it being stopped.
-Started code staging for windows HiveOS API 2.0 (not completed) integration
-Started code staging for HiveOS new version 1.4.0b
-Started code staging for Ubuntu new version 1.4.0c

1.4.0b
HiveOS: Changes:
-Added AMD miners, and changed code to work for AMD in HiveOS (HiveOS only)
-AMD miners are not plug-and-play. You will likely have to edit settings in the miner files, AMD Unix is complicated, but you should be familiar with this.
-Will add more/fine AMD miners- This is simply beta test of core script. Please use and review any issues you see.
-SGminer is not working on blockmasters- It is a pool issue, not miner.
-Fixed stats and have background miner logs sent to "Logs" folder
-Reduced switch load time by a few seconds.
-New Remote command: get-screen [type]...Show last 100 lines of group1 miner on HiveOS site.
-Ensured Timeouts will timeout miners.
-Added Function Restart-Database. Miner will now switch off algorithm/coin if detects no hash after 1 min.
-Times on 'active' command should be more accurate.
-'active' now shows database countdown.
-code staging for Ubuntu version.
-added more algorithm text files that are case specific for naming convetions between pool and miner.
-Enemy 1.1.7
-T-Rex 0.6.2 (added balloon)
-Lots of migrations and code cleanup to make miner better/faster