specific python version? or nvidia driver then? for me its very bad in performance
I have used this post Miner- SILENTARMY v5 - #1183 by sharpenedrose from yarnik and use the nvidia driver version 375.70. Works like a charm.
what windows are you using? 8? 10? 7? what drivers?
thanks
using windows 7 sp1, didn’t pay attention to driver, so old 372.70, gtx980 oc: core:1490,mem:4000, nicehash eqm with -as 1 -cb 64 -ct 512
1 Like
Yes, but for me with falling prices instant trading was good option
1 Like
ok thanks for the info, i remember them saying win 10 was no good for 900 series cards
i tried genoils silentarmy miner in win10 and again bad results compared to silentarmy in linux
Far better than nicechas 120Sol with 1070 and 80Sol with 1060 2’5% dev fee as claymore
zcminer released new version not long ago, it do on windows about 105 sol/s with my gtx980 card
Hello, anyone has tried this card?
I have tried it with Cuda 7.5 and gcc 4.9.1 on CentOS. Here’s my benchmark output from nheqminer commit e769c27 with COMPUTE=35:
$ ./nheqminer_cuda_tromp -cd 0 -b 200 -t 0
==================== www.nicehash.com ====================
Equihash CPU&GPU Miner for NiceHash v0.3a
Thanks to Zcash developers for providing base of the code.
Special thanks to tromp, xenoncat and eXtremal-ik7 for providing
optimized CPU, CUDA and AMD equihash solvers .
==================== www.nicehash.com ====================
Setting log level to 2
[13:12:16][0x00007fb200fe1720] Using SSE2: YES
[13:12:16][0x00007fb200fe1720] Using AVX: YES
[13:12:16][0x00007fb200fe1720] Using AVX2: YES
[13:12:16][0x00007fb200fe1720] Benchmarking CUDA worker (CUDA-TROMP) Tesla K80 (#0) BLOCKS=91, THREADS=64
[13:12:16][0x00007fb200fe1720] Benchmark starting... this may take several minutes, please wait...
[13:12:45][0x00007fb200fe1720] Benchmark done!
[13:12:45][0x00007fb200fe1720] Total time : 28930 ms
[13:12:45][0x00007fb200fe1720] Total iterations: 200
[13:12:45][0x00007fb200fe1720] Total solutions found: 413
[13:12:45][0x00007fb200fe1720] Speed: 6.91324 I/s
[13:12:45][0x00007fb200fe1720] Speed: 14.2758 Sols/s
Ever tried running like 16 Instances of the miner?
Hmm, not sure how I would run 16 instances of the miner with only one card. 
The code is already running 64 threads on the card by default.
If I knock it back to 32 threads (for example) with
$ ././nheqminer_cuda_tromp -cd 0 -ct 32 -b 200 -t 0
then the speed also gets cut back, to around 4.75 I/s , 9.5 Sols/s
Ah alright. There were no thread numbers in your last cmdline example and I didn’t know you could specify the number of threads with nheqminer. Doubt the kernel is in any way optimized for HPC cards.
Damn 
(insert 20 characters)
only 14 sol? From this super gpu card I was expecting at least 100/120 sol or more!
Eventually, the cpu are still the best!
Hey, could you please tell me what exactly is -b 200 for? I see -ct 32 are threads, how many should be used on a gtx 960? Thanks a lot!
Well, that’s effectively using only half the card (one physical K80 = two logical K40s if I understand correctly.) So if I use the whole physical card I get, as you’d hope, around double the throughput:
$ ./nheqminer_cuda_tromp -cd 0 1 -b 200 -t 0
==================== www.nicehash.com ====================
Equihash CPU&GPU Miner for NiceHash v0.3a
Thanks to Zcash developers for providing base of the code.
Special thanks to tromp, xenoncat and eXtremal-ik7 for providing
optimized CPU, CUDA and AMD equihash solvers .
==================== www.nicehash.com ====================
Setting log level to 2
[15:10:19][0x00007ffa39b78720] Using SSE2: YES
[15:10:19][0x00007ffa39b78720] Using AVX: YES
[15:10:19][0x00007ffa39b78720] Using AVX2: YES
[15:10:19][0x00007ffa39b78720] Benchmarking CUDA worker (CUDA-TROMP) Tesla K80 (#0) BLOCKS=91, THREADS=64
[15:10:20][0x00007ffa39b78720] Benchmarking CUDA worker (CUDA-TROMP) Tesla K80 (#1) BLOCKS=91, THREADS=64
[15:10:20][0x00007ffa39b78720] Benchmark starting... this may take several minutes, please wait...
[15:10:32][0x00007ffa39b78720] Benchmark done!
[15:10:32][0x00007ffa39b78720] Total time : 12487 ms
[15:10:32][0x00007ffa39b78720] Total iterations: 200
[15:10:32][0x00007ffa39b78720] Total solutions found: 363
[15:10:32][0x00007ffa39b78720] Speed: 16.0167 I/s
[15:10:32][0x00007ffa39b78720] Speed: 29.0702 Sols/s
How about running multiple instances of the miner?
acceptable as a value. I make 30/32 with an i7 6700K 
OP, if you are thinking about using the p2.16xlarge instance of AWS, that’s not cost effective. You’d better use the g2.8xlarge, and even then that’s still multiple times more expensive than using CPU only, even using spot price at low traffic times.