GPU Mining is out, come and let us know of your bench scores!

1568101147

Comments

  • LogicaluserLogicaluser Member Posts: 214 ✭✭
    edited October 2015
    My best thus far is ~106mh for 4x 7950, but I need to be clocked at 1020/1500mhz to do so
    (using the ethpool.org qtminer, 200ms per batch, win7 x64 & 15.7.1)
    Stilt bios gave me maybe an extra 0.1-0.2 mh per card

    Can get my 7970/280x up to 27-28mh each at 1050-1100mhz, havent tried stilt bioses for these yet

    Did some experimenting with various Catalyst versions on Win7, was unable to find any hashrate gains

    However... I did discover that 14.4 and previous don't actually hash correctly, they show proper hashrates but submit few if any shares. (pool reported hash fell +90%)
    Post edited by Logicaluser on
  • EastwindEastwind Member Posts: 107
    Does anybody find that ethminer works faster under Win7 than under Win8.1?

    @Logicaluser What is your qtminer setting?
  • LogicaluserLogicaluser Member Posts: 214 ✭✭
    Eastwind said:

    Does anybody find that ethminer works faster under Win7 than under Win8.1?

    @Logicaluser What is your qtminer setting?

    currently using just --ms-per-batch 200

    I think it improves things vs default, but I haven't given it a lot of tinkering, it's likely more fine grained worksize adjustment would be better, will give it a try at some point

    oh, and --no-precompute on my single core sempron rigs, they don't like background computing next DAG

    My Win8.1 rig seems okay (108mh for pairs of 280x and 7950), altho not much faster than my 4x7950 rig, might try a win7 image to see if theres any gains
  • luffyluffy Member Posts: 8
    does anyone have success with AMD 6990 card?
    i keep getting the error
    "
    Creating one big buffer for the DAG
    Allocating/mapping single buffer failed with: clCreateBuffer(-61). GPU can't all
    ocate the DAG in a single chunk. Bailing.
    "
    i have a rig with 8GB ram
    thanks
  • ordoeordoe tehranMember Posts: 132 ✭✭
    luffy said:

    does anyone have success with AMD 6990 card?
    i keep getting the error
    "
    Creating one big buffer for the DAG
    Allocating/mapping single buffer failed with: clCreateBuffer(-61). GPU can't all
    ocate the DAG in a single chunk. Bailing.
    "
    i have a rig with 8GB ram
    thanks

    Your GPU needs at least 2GB ram. Not the system ram, but the on-GPU-ram. I guess the 6990 does not have enough.
  • luffyluffy Member Posts: 8
    edited November 2015
    yes, it has 4GB. 2 for every gpu core ;)
  • nassnass Member Posts: 1
    Anyone tried with Radeon R9 380 2GB DDR5. Is it as good as the R9 (280, 290) X series. Have one that's idle with a friend and considering borrowing it. Just to see if its worth the MH.
  • luffyluffy Member Posts: 8
    nass said:

    Anyone tried with Radeon R9 380 2GB DDR5. Is it as good as the R9 (280, 290) X series. Have one that's idle with a friend and considering borrowing it. Just to see if its worth the MH.

    check this thread. they report that 380 is around 18-19 MH/s
    not as good as 280-290
    the best cards price-wise are 7970-280
  • mplunkett5mplunkett5 Member Posts: 1
    Hey All,

    Just installed my new 390 MSI and got these figures:

    Windows / Windows 7 / GPU / 0.9.41
    min / mean / max: 25672996/ 25791835/ 26030960
    Inner mean: 17.2609 MH

    When pool mining what MH figure should you use? Is it the Max value or the Inner Mean Value?

    Thanks Guys.

    mark
  • GreaterNinjaGreaterNinja Member Posts: 60
    edited November 2015
    1 x Saphire 7990; i7 5820k; 16gb ddr4 ram; 2 GPU. I'm guessing the card pulls about 380-390 watts/hour at max load.

    At 100% load the entire system pulls 554-561 watts/hour. At 17cents/kwh, there is negative ROI. However, I'm doing this for fun.

    Using platform: AMD Accelerated Parallel Processing
    Using device: Tahiti(OpenCL 1.2 AMD-APP (1800.8))
    Using device: Tahiti(OpenCL 1.2 AMD-APP (1800.8))
    Creating one big buffer for the DAG
    Loading single big chunk kernels
    Mapping one big chunk.
    Creating buffer for header.
    Creating mining buffer 0
    Creating mining buffer 1
    Trial 1... 46923776
    Trial 2... 47972352
    Trial 3... 46661632
    Trial 4... 48147114
    Trial 5... 47535445
    min/mean/max: 46661632/47448063/48147114 H/s
    inner mean: 47477191 H/s
    47.477MH/s
    Post edited by GreaterNinja on
  • ZUIJftyihuYTZUIJftyihuYT Member Posts: 3
    edited November 2015
    Ubuntu 14.04 / GeForce GTX 780 / 0.9.21
    min / mean / max: 1.34/1.37/1.38 MH/s
    inner mean: 1.37 MH/s. www.casesam.co.uk
    Post edited by ZUIJftyihuYT on
  • GenoilGenoil 0xeb9310b185455f863f526dab3d245809f6854b4dMember Posts: 769 ✭✭✭
    @ZUIJftyihuYT you are probably CPU minig. GTX780 can do 18.2 MH/s with my CUDA miner.
  • crazyivancrazyivan Member Posts: 23
    So what s the consensus, what s currently the best etherum mining GPU? Has anyone been able to hit over 30 mhs per card with anything?
  • ordoeordoe tehranMember Posts: 132 ✭✭
    crazyivan said:

    So what s the consensus, what s currently the best etherum mining GPU? Has anyone been able to hit over 30 mhs per card with anything?

    The AMD FirePro S9150 should do and does only require 235W :-)
    The GSI Helmholtz Center in Darmstadt has 600 of them. Maybe we could ping them for a benchmark?
  • crazyivancrazyivan Member Posts: 23
    ordoe said:

    crazyivan said:

    So what s the consensus, what s currently the best etherum mining GPU? Has anyone been able to hit over 30 mhs per card with anything?

    The AMD FirePro S9150 should do and does only require 235W :-)
    The GSI Helmholtz Center in Darmstadt has 600 of them. Maybe we could ping them for a benchmark?
    Hah, that s 3000 euros GPU. For that money, you can get 12 280x in 3 rigs.
  • LogicaluserLogicaluser Member Posts: 214 ✭✭
    crazyivan said:

    So what s the consensus, what s currently the best etherum mining GPU? Has anyone been able to hit over 30 mhs per card with anything?

    7950/280 seem to be the sweet spot in terms of dollar per hashrate (25.5-26.5mh each), with considerably higher clocked 7970/280x's only offering maybe an extra megahash or so
  • o0ragman0oo0ragman0o Member, Moderator Posts: 1,291 mod
    7990's are probably still the bomb (until I managed to destroy mine) followed by 7950's. GTX 750 Ti's were right up there but have now succumb to the >1GB DAG bandwidth crunch. my 750Ti and 7950 rigs were doing great but recent increase in network hash and low markets have tipped over my break-even point. :(
  • GreaterNinjaGreaterNinja Member Posts: 60
    Saphire 7990 @ 46.9 MH/s; GPU 1 & 2 usage @ 99%@1000mhz; Temps inside case are 86C and 83C. I downvolted the cores from 1.2v to 1.120v; Fan 1 @ 65%; Fan 2% 56%; Total system power usage is at 493-500 watts to the wall, which is about 140 watts or less per GPU core. So about 6 watt/MH. If I downvolt them more and reduce GPU clock I can get it to about 5 watt/MH. The only downside is when I play starcraft 2, these settings usually require 1-2 reboots to revert/fix. I prefer convenience.

    I also have 3 x 7950 and 2 x 7870XT Myst (tahiti) collecting dust.
  • dont12dont12 Member Posts: 60
    7990 getting 52 Mh/s cores at 1.2V Fans at 100% temps about 85 to 90 Wattage at wall 550W
  • davidrentaodavidrentao Member Posts: 60
    don using high freq to mining, two reason you need careful, 1. repair cards faster, 2. high ele-cost on mine. so reduce the freq to mine better in this period, bcz coin price not very high, don need mine very faster and if you solo mine faster can got block between other people, hash around 8.5T, very high now.....
  • peoplespeoples Member Posts: 29 ✭✭
    Stats for R9 380X

    min/mean/max: 19835562/19957896/20010325 H/s
    inner mean: 13311089 H/s
  • Dr_VictorDr_Victor Member Posts: 4
    Could you tell me, what kind of changes should I do to increase hashrate for my 7970 from 24Mh/s?
  • chevrochevro Member Posts: 16
    I got an XFX R7 370 2Gb from Bestbuy yesterday. 16.3 Mh/s is what pool says
  • fublyfubly Member Posts: 1
    Sapphire AMD FirePro™ S10000

    what do you think about this card?
  • nikhilshekhawatnikhilshekhawat Member Posts: 1
    Can someone please help me debug this? This command does not work for me on Mac OSX for a GeForce 750 M with openCL 1.2

    ++)Ethereum
    [OPENCL]:Found suitable OpenCL device [Iris Pro] with 1610612736 bytes of GPU memory
    Benchmarking on platform: { "platform": "Apple", "device": "Iris Pro", "version": "OpenCL 1.2 " }
    Preparing DAG...
    Warming up...
    ℹ 10:26:48.446|gpuminer0 workLoop 0 #00000000… #00000000…
    ℹ 10:26:48.446|gpuminer1 workLoop 0 #00000000… #00000000…
    ℹ 10:26:48.447|gpuminer0 Initialising miner...
    ℹ 10:26:48.447|gpuminer1 Initialising miner...
    [OPENCL]:Using platform: Apple
    [OPENCL]:Using device: GeForce GT 750M(OpenCL 1.2 )
    [OPENCL]:Using platform: Apple
    [OPENCL]:Using device: Iris Pro(OpenCL 1.2 )
    [OPENCL]:Printing program log
    [OPENCL]:
    [OPENCL]:Creating one big buffer for the DAG
    [OPENCL]:Loading single big chunk kernels
    [OPENCL]:Mapping one big chunk.
    [OPENCL]:ptxas application ptx input, line 5641; error : Call has wrong number of parameters
    ptxas application ptx input, line 8472; error : State space mismatch between instruction and address in instruction 'st'
    ptxas application ptx input, line 8473; error : State space mismatch between instruction and address in instruction 'st'
    ptxas application ptx input, line 8474; error : State space mismatch between instruction and address in instruction 'st'
    ptxas application ptx input, line 8475; error : State space mismatch between instruction and address in instruction 'st'
    ptxas application ptx input, line 8525; error : Call has wrong number of parameters
    ptxas application ptx input, line 8642; error : Call has wrong number of parameters
    ptxas application ptx input, line 11426; error : State space mismatch between instruction and address in instruction 'st'
    ptxas application ptx input, line 11427; error : State space mismatch between instruction and address in instruction 'st'
    ptxas application ptx input, line 11428; error : State space mismatch between instruction and address in instruction 'st'
    ptxas application ptx input, line 11429; error : State space mismatch between instruction and address in instruction 'st'
    ptxas application ptx input, line 11491; error : Call has wrong number of parameters
    ptxas application ptx input, line 8472; error : Unknown symbol 'func_retval0'
    ptxas application ptx input, line 8472; error : Label expected for forward reference of 'func_retval0'
    ptxas application ptx input, line 8473; error : Unknown symbol 'func_retval0'
    ptxas application ptx input, line 8473; error : Label expected for forward reference of 'func_retval0'
    ptxas application ptx input, line 8474; error : Unknown symbol 'func_retval0'
    ptxas application ptx input, line 8474; error : Label expected for forward reference of 'func_retval0'
    ptxas application ptx input, line 8475; error : Unknown symbol 'func_retval0'
    ptxas application ptx input, line 8475; error : Label expected for forward reference of 'func_retval0'
    ptxas application ptx input, line 11426; error : Unknown symbol 'func_retval0'
    ptxas application ptx input, line 11426; error : Label expected for forward reference of 'func_retval0'
    ptxas application ptx input, line 11427; error : Unknown symbol 'func_retval0'
    ptxas application ptx input, line 11427; error : Label expected for forward reference of 'func_retval0'
    ptxas application ptx input, line 11428; error : Unknown symbol 'func_retval0'
    ptxas application ptx input, line 11428; error : Label expected for forward reference of 'func_retval0'
    ptxas application ptx input, line 11429; error : Unknown symbol 'func_retval0'
    ptxas application ptx input, line 11429; error : Label expected for forward reference of 'func_retval0'
    ptxas fatal : Ptx assembly aborted due to errors
    [OPENCL]:clEnqueueWriteBuffer(-38)
    Abort trap: 6
  • auburn_niceauburn_nice RussiaMember Posts: 22
    edited December 2015
    It's wrong place for such comments.
    Yoi don't have enough GPU memory.
    Next time please read the description first. =)

    Btw Sapphire HD 7970
    24 Mh/s on dwarf average. 1125/1350 T=66
  • dont12dont12 Member Posts: 60
    Power Color Devil 390X, no OC just stock, 27.5 Mhs at 56C. Got a hybrid closed loop cooler, sandwiched between two GPUs and running nice and cool at 27.5 Mhs power draw sub 200W
  • chafer99chafer99 Member Posts: 1
    Win-7 64bit / 3 x AMD Sapphire R9-380 2GB / ethminer 0.9.41
    Inner mean: 58877262 H/s

    AMD Gigabyte R7-370 2GB + Sapphire R7-260X 2GB
    Inner mean: 14624490 H/s
Sign In or Register to comment.