When is Jminer more efficient to use than Blago Miner?



  • Team,

    • I am able to run both Jminer-0.4.11-Snapshot and Blago V1.170820_AVX2 with no issues on either.
    • I currently have a plot capacity of 25.5TB and will be adding more over the next few days.
    • All my capacity is on Sata connections and all future capacity will be on Sata connections as well.
    • Platform is W10 64 bit, Enterprise

    Q1 - My questions are, when is Jminer more efficient to use over Blago AVX2?
    Q2 - Is there a breakpoint in capacity where it makes more sense to use Jminer over Blago Miner?
    Q3 - Do you earn or possibly earn more BURST using Jminer?

    Thanks in advance
    Digidigm



  • Jminer is a gpu miner. This can start to make sense if you have more drives than you have real cores in the machine. If it is small drive,s it will not make much of a difference if the drives are scanned fast with blago miner.

    The only thing to concider is wich one is fastest if you have alot of drives and the time to scan them exceeds blocktime. There is an extreemly small benefit in beeing superfast if you poolmine. That is that you might be able to submit an extra dl if the block is mined very fast. but it is on rare occations.



  • i have a 220TB unit running. First i was with jminer and a rx280.
    Speed was pretty ok, i was reading everything in about 45 sec. GPU at about 70% load.

    I'm now with a 7700K and a read time about 39sec. RX is mining zcash now.



  • I had a gpu that was running the Jminer way faster than the cpu Blago miner. Then the Jminer started running at 1/3 speed. That was right around the time I installed Nicehash and was playing around with it. The low speed was there whether the Nicehash was running or not. I think the Nicehash installation changed some settings on the GPU or it was a random breakdown of the card. So I CPU mine and the card runs on Nicehash for about $1 a day.



  • @rds Ok so maybe Nicehash did something to the card, who knows but in general is it better to mine with Jminer or Blago Miner? I just really don't want to burn through a GPU card, EVGA GTX1050ti card that I am using if I don't have to. And from what I can tell it is not even being taxed.

    I thought maybe after reaching a certain capacity it might make sense that is why I was asking.

    Digidigm



  • @digidigm , there is no one answer. You have to test each condition. I have a 32 core server that is way faster than the GPU card. I have another unit that is about the the same CPU and GPU same speed.



  • @rds RDS thanks, I know about your beast machines. But your right I guess I should just compare the read times on the Blago miner and GPU miner. I am plotting like nuts right now so I will do a test once my rig gets the new capacity I am adding and post my findings here. It will be a few days.



  • @rds Ok so I have connected my new capacity 47.3TB these are the numbers..

    From what I can tell the Jminer - GPU is reading all the capacity at 16s at an avg of 782 MB/s eff. 807 MB/s

    Blago miner - CPU 683.21 MB/s however I cannot tell from blago how long it takes.

    From what I can see is Jminer doing better?

    • All drives are connected to the MOBO
    • Some drives are connected to the MOBO via and LSI SAS to SATA PCIe card which gives great speeds.

    Comments?

    0_1512229442059_50TB.JPG 0_1512229451441_Blago.JPG



  • For JMiner, use the 100% for all drives read time at the very end of the block readout. For the Blago, you have to wait for a block that is longer than your read time. Looks like the two blocks you show were short.



  • @rds Hello Rds, thanks for getting back to me, what did you mean by "I was short" ? That is what my output look likes all the time with blago... Am I missing something here?

    See my new screen shot below with newer blocks. I also posted my mining config details.

    0_1512239475621_Blago2.JPG

    "Mode" : "pool",
    "Server" : "ninja.burstcoin.ml",
    "Port" : 8080,

    "UpdaterAddr" : "ninja.burstcoin.ml",
    "UpdaterPort" : 8080,

    "InfoAddr" : "ninja.burstcoin.ml",
    "InfoPort" : 8080,

    "EnableProxy" : false,
    "ProxyPort" : 8126,

    "Paths":
    [
    "D:\plot","E:\plot","F:\plot","G:\plot","H:\plot","I:\plot","L:\plot","M:\plot","N:\plot","O:\plot"
    ],
    "CacheSize" : 40000,

    "Debug" : false,
    "UseHDDWakeUp" : false,

    "TargetDeadline": 80000000,

    "SendInterval" : 100,
    "UpdateInterval" : 950,

    "UseLog" : false,
    "ShowWinner" : false,
    "UseBoost" : false,

    "WinSizeX" : 90,
    "WinSizeY" : 60
    }


  • admin

    @digidigm What version are you running? Mine shows the time to mine for each individual drive - not sure off the top of my head if thats default behavior or a setting.



  • @digidigm - miner-v1.170820_AVX2.exe

    When I used to mine on Party.burstcoin ..... I used to see times. When I switched over to "ninja.burstcoin.ml" it stopped displaying the times. I thought it was a function of the pool and what they sent back. I cannot see any of the config entries that controls this.



  • @digidigm , change debug to "true".



  • @rds BAM !!! That worked now we have times.

    So if I am reading this right it is taking between 18-19 sec to read my capacity?
    vs. my Jimer in the image above in the thread which was about 16 sec.

    0_1512272024316_Blago3.JPG



  • yes, so at this point either type miner is acceptable. As you add more drives recheck and adjust as necessary.



  • @rds Thanks for the exercise now I know what to compare to each other.



  • @digidigm try change value "CacheSize" : 4000,



  • @blago ,

    I moved my cache size from 100 to 10000000 in 1000 steps awhile ago. There was no appreciable diff for me at any level until it was really big then it bogged down the machine.



  • @blago @rds I don't know if anyone noticed but I have mine at 40,000 and it seems to be performing fine.

    Blago misunderstood what you were saying. I just changed it from 40,000 to 4,000 lets see how the reads go.

    Update - 3 hours later - and my read times are about 14-15 sec for my 47tb capacity. Down from 18-19 secs at a cache of 4,000. This is good stuff. I am going to let it run overnight see how it turns out.



  • @blago @rds

    Update 23 hours - Ok I have been running blago for almost 24 hours and the average read time for my 47.5 TB is about 14 sec... This works. Switching over to Jminer now since I have not been on Jminer since I added my new capacity.