background

Script N multipool support?

Hi all,

Does MacMiner work with Script N multi pools, like TradeMyBit? Or is MacMiner's Script N support limited to specific coins, like VTC?

Thanks in advance. 

Comments

  • The vertcoin backend is a specific vertcoin version of cgminer, if there's a better backed you'd rather I'd compile for that please do point it out to me!
  • Thanks for getting back to me. 

    I'm an experienced scrypt miner (mostly with ASICs) and new to scrypt N, so I'm afraid that I don't know of a better backend miner. Just getting my feet wet here...

    While waiting for a reply to my original post, I've been experimenting with the cpuminer-vert downloaded from your instructions here:


    I've got it pointed at a scrypt N multi pool, and it *seems* to be working. But again, I'm unfamiliar with scrypt N mining software, as well as being a noob with CPU mining.

    Here's some sample output that I'm seeing from vertminer:

    --snip--
    [2014-05-21 15:17:07] thread 14: 25416 hashes, 1.73 khash/s
    [2014-05-21 15:17:07] accepted: 1982/2929 (67.67%), 40.99 khash/s (yay!!!)
    [2014-05-21 15:17:11] Stratum detected new block
    [2014-05-21 15:17:11] thread 5: 31596 hashes, 1.73 khash/s
    --snip--

    Does that seem reasonable to you? I seem to get similar results regardless of what coin the pool is mining, although I have to admit that I haven't kept exhaustive records. At the time I captured the above log, I think that the pool was mining EXE.

    Thanks in advance for any comments as to whether I'm on the right track here or not.
  • What pool were you using? P2Pool nodes can give 'false positives' as they don't require a password - if it wasn't a P2Pool node it looks as though it was working fine :)
  • edited May 2014
    I'm currently on the TradeMyBit multi pool. It requires setting up workers with names and passwords.

    I am seeing coin accumulating in my TradeMyBit account, so I guess it's working. But I have no idea if it's working optimally.

    Scrypt mining with ASICs is quite a bit easier to figure out and manage, at least for a Mac/Linux guy like me. But I am having fun experimenting with Scrypt N.
  • There's really nothing you can do to optimise CPU mining, setting the number of threads is the only thing but it's max by default so I'd say you're running as best as you can at present :) Despite compiling the miner, I've only used it for VTC!
  • One more quick question. When I run vertminer from the command line it says it's running in "scrypt" mode. Am I supposed to somehow specify that the algorithm is scrypt 2048? Or is scrypt N automatically baked into this version?

    There was no mention on your site for doing this, and when I tried to set scrypt 2048 (various syntactical guesses) via the --algo option, I kept getting errors from vertminer.

    The reason why I ask is that I can't seem to get vertminer to run any better than 85% accepts. The high reject level makes me wonder if I've got a configuration problem. 

    I've tried a different pool and get similar results. Since I have a pathetic amount of hash, I've tried a lower difficulty (8, for the moment) and still I see rejects in the 15 - 20% range.

    Thanks in advance for any suggestions.
  • I took a look at the github site, and from there I found the corresponding bitcointalk forum thread. According to this post, there's a problem when building with optimization levels of -O2 or higher under the Lion version of clang that could result in high reject rates:


    I tried the latest build from that thread, but it gives me 100% rejects. Maybe that version only supports bitcoin and scrypt?

    Could you possibly look at rebuilding your Scrypt-N capable version with optimizations set lower? Or maybe the optimization problem has been fixed in the latest Xcode 5.1.1 clang?

    Note that one nice thing about the other version (other than the fact that it got 100% rejects) was that it was completely self contained. It's built such that the binary just worked on my Mountain Lion Mac without needing its own copies of the dylibs.
  • The cpuminer in that thread is the original pooler cpuminer which only supports SHA256 and original scrypt, but I'll look at recompiling the relevant miner with gcc. I suspect though the problem is to do with relatively low hash rate perhaps causing a high rate of stale shares.
  • Thanks in advance. I hope that the recompile is not a waste of time.

    I'm happy to test when you have a rebuilt version.

    Note that I tried to build this myself (I'm an experienced Mac developer), but ran into problems. My build environment probably has an older version of automake, because I got errors just running ./autogen.sh.

    At some point I should figure out some d'etant between my work build environment and my hobby build environment ;-)
  • I'll update here. 

    If you have a chance to bug check the source of MacMiner at any point that would be HUGELY appreciated as there are quite a lot of issues yet to be fixed!

    with regards to compiling some of these more obscure miners - if autogen.sh doesn't work try running autoreconf and go from there.
  • No problem. I'll review the code (which is I think what you're asking). I probably can't spare the time till this weekend though. I'll let you know if I spot anything.

    Thanks for the autoreconf suggestion. It got me further, but now I'm seeing a config error which I think is due to the fact that I didn't bother to install any of the prerequisites (like lib curl). I'll keep poking at this, time permitting. But I'd be just as happy to try your version after you've built it.
  • Please download this, decompress it and chuck it in MacMiner.app/Contents/Resources so you can then run MacMiner.app/Contents/Resources/sgminer/bin/sgminer

    Note please ignore the two warnings that show immediately on launch. I believe these are related to the lack of ADL monitoring on Mac OS X, but in any case, didn't affect mining in my test.

    http://downloads.fabulouspanda.co.uk/sgminer.zip

    I believe you need to use the flag
    --algorithm nscrypt

    for scrypt-n
  • Thanks!

    So you recommend this miner instead of the cpuminer-vert version from your blog?
  • this one is a GPU miner - you can use them together - not sure whether you can GPU mine though..
  • Ah, with the machines I'm on (that have crappy GPUs) I've been using the cpuminer you posted here:


    I don't suppose I can trouble you to build that one with -O1?

    Sorry to be a pain...
  • I went ahead and tried the sgminer that you provided. I'm getting similar KH/s out of the GPU as with cpuminer, but with no stales pointed at the same pool. 

    So it's looking more and more likely that the optimization level explains why cpuminer is performing so badly.
  • One thing to keep in mind is that one GPU = one thread. all the kh are working towards one goal. with the cpu you have multiple threads, each working on separate work more slowly. 

    It looks as though scrypt N is inches away from poolers cpuminer. I compiled it, but --algo=scrypt:N as is set out, and manifold variations are not working at present. will come back to it soon!
  • Ah, I get it. Even though my Mac Pro (5,1) has the same roughly 40 KH/s for the GPU and CPU, the CPU actually has 24 1.7 KH/s workers. And it's because these workers are slow that they can't keep up.

    Looking forward to pooler's cpuminer with script N support, although it sounds like it might not make any difference. We'll have to see...
  • FYI, I managed to rebuild the Bufius cpuminer with optimizations set to -O1 and the resulting binary has zero rejects. Looks like there is a clang optimization problem with cpuminer after all.
  • Thanks for the update. I've a bunch of rebuilds planned for the near future so expect an update soon :)
  • edited June 2014
    Thought I might share..I have been CPU mining Scrypt-n on Waffle-Pool. Currently at 69.71%accepted shares (last 12-15 hours) it has been as low as 45% before. 

    Reading all that technical stuff you guys wrote is like reading chinese to me…so I will be glad when we can mine S-n and x11.

    Thanks for all the hard work on Macminer…it is appreciated

    Edit: I am using version 1.5.20b3
  • The included Bufius cpuminer for BTC has been recompiled with -O1 - please let me know if this solves the issue!

Leave a Comment

Drop image/file