CUDA miner

1282931333469

Comments

  • furionlovefurionlove Member Posts: 10
    @Genoil where can one download the 1.0.4b3 for Windows?
  • ImAMiner?ImAMiner? ✭✭ Member Posts: 208 ✭✭
  • jakobshadowjakobshadow Member Posts: 5
    I am getting 18mhs with gtx 970 only using ethminer -U , any tips on what to tweak to get it to 20mhs?
  • Mo35Mo35 Member Posts: 43
    @jakobshadow use either NV Inspector or MSI afterburner and rise memory clock
  • safi182safi182 Member Posts: 14
    msi afterburner can do 20 mhs cuse you need to set up p0 power state
    http://cryptomining-blog.com/
  • jakobshadowjakobshadow Member Posts: 5
  • oddlyxooddlyxo Member Posts: 6
    hello,

    i have been using
    ethminer -F http://eth-us.dwarfpool.com/xxx/miner-05 -U --gpu-batch-size 20

    the following [minor] success:



    Using Windows 10 and a single GTX 750TI.

    I have consumed a lot on the subject and have successfully mined. However, from other reports using a 750TI
    seem to suggest I should be somewhere around 6-8MH/S.

    Suggestions?
  • oddlyxooddlyxo Member Posts: 6
    Also, thank you @Genoil, I was unable to mine successfully until I found your CUDA miner. (Which may be a part of the problem.)

    I feel like my GPU isn't working as hard as it could. The fan speed never goes above 42% and the GPU temp is steady under 40 C while mining.
  • voodoospiritvoodoospirit Member Posts: 3
    with -M i get 19-20Mhs with my 980 gtx but on when i mine in a pool , i get only 8Mhs in the cmd box....have no clue why i have so differences...
  • DavBEDavBE Member Posts: 26
    @voodoospirit this has been discussed many many times, read a few pages back and you will find an answer.
  • voodoospiritvoodoospirit Member Posts: 3
    must have missed a bunch of pages....thx :)
  • oddlyxooddlyxo Member Posts: 6
    I must have been misinterpreting the data.

    I changed the --gpu-batch-size 20 to --gpu-batch-size 18 and it seemed to run much smoother.

    I am happy with my 5-8mh/s on my gaming rig.
  • MilotMehmetiMilotMehmeti Member Posts: 1
    I have difficulties installing cuda miner on Ubuntu 15.10

    example:

    [email protected]:~/cpp-ethereum# make -j4
    [ 13%] Built target scrypt
    [ 16%] Built target secp256k1
    [ 45%] Built target devcore
    [ 52%] Built target ethash
    [ 56%] Building NVCC (Device) object libethash-cuda/CMakeFiles/ethash-cuda.dir/ethash-cuda_generated_ethash_cuda_miner_kernel.cu.o
    [ 65%] Built target devcrypto
    [ 69%] Built target ethash-cl
    nvcc fatal : redefinition of argument 'std'
    CMake Error at ethash-cuda_generated_ethash_cuda_miner_kernel.cu.o.cmake:207 (message):
    Error generating
    /home/gega/cpp-ethereum/libethash-cuda/CMakeFiles/ethash-cuda.dir//./ethash-cuda_generated_ethash_cuda_miner_kernel.cu.o


    libethash-cuda/CMakeFiles/ethash-cuda.dir/build.make:63: recipe for target 'libethash-cuda/CMakeFiles/ethash-cuda.dir/ethash-cuda_generated_ethash_cuda_miner_kernel.cu.o' failed
    make[2]: *** [libethash-cuda/CMakeFiles/ethash-cuda.dir/ethash-cuda_generated_ethash_cuda_miner_kernel.cu.o] Error 1
    CMakeFiles/Makefile2:487: recipe for target 'libethash-cuda/CMakeFiles/ethash-cuda.dir/all' failed
    make[1]: *** [libethash-cuda/CMakeFiles/ethash-cuda.dir/all] Error 2
    make[1]: *** Waiting for unfinished jobs....
    ETH_FATDB: 0
    [ 69%] Built target BuildInfo.h
    Makefile:127: recipe for target 'all' failed
    make: *** [all] Error 2


    Please help
  • Marvell9Marvell9 ✭✭✭ Member Posts: 593 ✭✭✭
    Anyone have some idea why this EVGA 970gxt says NA and not supported when I try to change the clocks through SMI ?

    Right now I'm only get 16.5 Mhs using the newest binaries win 7 with the older driver suggested

    any ideas?
  • safi182safi182 Member Posts: 14
    Marvell9 said:

    Anyone have some idea why this EVGA 970gxt says NA and not supported when I try to change the clocks through SMI ?

    Right now I'm only get 16.5 Mhs using the newest binaries win 7 with the older driver suggested

    any ideas?

    did you chek what clock is your card support with this nvidia-smi -q -d SUPPORTED_CLOCKS | more

  • trotoltrotol Member Posts: 102
    Genoil, it seems that 1.0.4b3 miner doesn`t work with coinotron-proxy
  • Marvell9Marvell9 ✭✭✭ Member Posts: 593 ✭✭✭
    safi182 said:

    Marvell9 said:

    Anyone have some idea why this EVGA 970gxt says NA and not supported when I try to change the clocks through SMI ?

    Right now I'm only get 16.5 Mhs using the newest binaries win 7 with the older driver suggested

    any ideas?

    did you chek what clock is your card support with this nvidia-smi -q -d SUPPORTED_CLOCKS | more

    @safi182 yeah checked that command it still says NA on the card maybe the clocks are locked ?
    this is the card
    http://www.newegg.com/Product/Product.aspx?Item=N82E16814127832

    I have two other EVGA cards
    http://www.newegg.com/Product/Product.aspx?Item=N82E16814487207

    all three on one motherboard I see around 54mh/s that seems very low with this new ethminer build right ?

    basically 17 mhs/s power use is really low though around 300 watts for the whole machine (seems lower than what it should be)

    also I notice that the machine is really slow when mining cant do anything else unlike my AMD boxes which are quite usable in windows 10 when the cards are mining. All my miners use the dual core penntiums so are these Nvida cards more CPU dependant than AMD or something ?

    I should be able to get closer to 20mhs per card right 17 seems really low even withoute the SMI fix.

  • safi182safi182 Member Posts: 14
    without smi fix 17 is not low thats the max hashrate with the new build (if you working in pool) hm how many card connecting to the motherboard?
    cuse i have msi card and i have no problem with it. ihave one gpu. 21 mhs
  • Marvell9Marvell9 ✭✭✭ Member Posts: 593 ✭✭✭
    @safi182 I have 3 cards connected to the board , but even with 1 card it was slow
    I'll try that Nvidia inspector tool when I get home from work I guess sucks though power usesage is great which is why I got these I have high cost eelectric around 18cents per kwh so I cant afford to run too many AMD cards.

  • AmphAmph Member Posts: 106
    how much one 980ti can do?
  • Bensam123Bensam123 Member Posts: 62
    Stratum is currently incompatible with Dwarfpools stratum?
  • trotoltrotol Member Posts: 102
    edited March 2016
    Bensam123 said:

    Stratum is currently incompatible with Dwarfpools stratum?

    The point is to remove support of getwork-protocol from genoil`s miner or to make a special key to deactivate it B)
    declared compatibility with coinotron are not proved - at least on my win7 x64 rig
  • gpfransengpfransen Member Posts: 1
    edited March 2016
    Hi all,

    I've been playing around with the options for hours but keep running into the following error when -U for CUDA is used, with -G it runs but far from optimal.

    CUDA error in func 'init' at line 220 : out of memory.
    Cuda error in func 'set_header' at line 66 : out of memory.

    Default example tried;

    ./ethminer -F http://eth-mine.suprnova.cc:3000/$USER.$INSTANCE/10 -U --cuda-extragpu-mem 0 --cuda-block-size 128 --cuda-grid-size 8192 --cuda-schedule auto

    Only thing I could think of is that it didn't accept my GPU version on build and had to resort to lowest accepted value of 53.

    CUDA toolkit shows Kepler as 30 or 32, so tried compiling in both but without further results.

    Also adding 65 to CMakeLists.txt gives error on build.

    Looks like I have a new challenge to compile this with the 6.5 toolkit :)

    Does anybody have any other ideas?
    Post edited by gpfransen on
  • xenomorph113xenomorph113 Member Posts: 7
    @Genoil miner is working great! I was having such a difficult time mining till I found your miner. got ~36MHs on my 2x 970's windows 10 rig (the secret to windows 10 and 970s seems to be the 347.52 drivers as @mo35 suggested, with anything newer the hashrate drops to 16-17MHhs)
  • Bensam123Bensam123 Member Posts: 62
    trotol said:

    Bensam123 said:

    Stratum is currently incompatible with Dwarfpools stratum?

    The point is to remove support of getwork-protocol from genoil`s miner or to make a special key to deactivate it B)
    declared compatibility with coinotron are not proved - at least on my win7 x64 rig
    Not sure what that means...
  • Mo35Mo35 Member Posts: 43

    @Genoil miner is working great! I was having such a difficult time mining till I found your miner. got ~36MHs on my 2x 970's windows 10 rig (the secret to windows 10 and 970s seems to be the 347.52 drivers as @mo35 suggested, with anything newer the hashrate drops to 16-17MHhs)

    if you got 36Mhz witch is first step , play around with NV inspector and try to tweak P0 state clocks , depends on individual card and luck (how your cards OCs) , but 2x970 should easely get you to ~40Mhz. i managed to OC my 2x950 and 1x970 setup from 36 to ~41Mhz
  • Marvell9Marvell9 ✭✭✭ Member Posts: 593 ✭✭✭
    Marvell9 said:

    @safi182 I have 3 cards connected to the board , but even with 1 card it was slow
    I'll try that Nvidia inspector tool when I get home from work I guess sucks though power usesage is great which is why I got these I have high cost eelectric around 18cents per kwh so I cant afford to run too many AMD cards.

    Solved my SMI problem on Windows 7. It seems the old driver thats recommend here does not allow SMI access on Windows 7, I updraft do the windowx7 box to the current driver for win 7 (602) or something and SMI state changing was possible.

    I Also had to turn off Windows 7 aero on my dual core machine or the miner stutters like crazy and cpu usage was 100%

    Now on the pool I see 50mhs to 63mhs max for my 3 970s around 20 to 21mhs per card
  • Marvell9Marvell9 ✭✭✭ Member Posts: 593 ✭✭✭
    edited March 2016
    @gpfransen I use -U not G sorry
    Post edited by Marvell9 on
  • GenoilGenoil ✭✭✭ 0xeb9310b185455f863f526dab3d245809f6854b4dMember Posts: 769 ✭✭✭
    Still out skiing. Will answer questions and DMs next week. Happy mining!
  • coldminecoldmine Member Posts: 2
    I have a really strange problem with my GTX 660Ti. When I'm using the latest version (1.0.4b3) I get an "Cuda error in func set_constants" error.



    Then I tried it again with an older version and stumbled across a new problem. It now hashes properly without any error but it doesn't send any shares whatsoever. I ran it for over one hour without a single share beeing transmitted. It doesn't work with eth-proxy nor with the inline pool option. What makes it even stranger is that it does send shares when I use OpenCL (-G). As soon as I use the -U command I get much more Mh/s but it doesn't transmit anything.

    I use the following command: ethminer --farm-recheck 200 -U --cuda-block-size 128 --cuda-schedule auto -F http://127.0.0.1:8080/rig1

    Anyone with a similar problem?
Sign In or Register to comment.