Please post your aionminer CUDA benchmarks here!

Please indicate the following:

a quoted copy/paste of benchmark results

  • aionminer version
  • driver version
  • GPU make and model
  • any GPU overclock / underclock settings you may be using

Here are my results:

[12:40:00][0x00007fc4e1c76740] Using SSE2: YES
[12:40:00][0x00007fc4e1c76740] Using AVX: YES
[12:40:00][0x00007fc4e1c76740] Using AVX2: NO
[12:40:00][0x00007fc4e1c76740] Benchmarking CUDA worker (CUDA-TROMP) GeForce GTX 1080 Ti (#0) BLOCKS=196, THREADS=64
[12:40:01][0x00007fc4e1c76740] Benchmark starting… this may take several minutes, please wait…
[12:40:27][0x00007fc4e1c76740] Benchmark done!
[12:40:27][0x00007fc4e1c76740] Total time : 25663 ms
[12:40:27][0x00007fc4e1c76740] Total iterations: 200
[12:40:27][0x00007fc4e1c76740] Total solutions found: 408
[12:40:27][0x00007fc4e1c76740] Speed: 7.79332 I/s
[12:40:27][0x00007fc4e1c76740] Speed: 15.8984 Sols/s

  • aionminer v0.1.8
  • Driver: NVIDIA 390.25
  • GPU: PNY GeForce GTX 1080 Ti
  • GPUPowerLevel: 170
  • GPUGraphicsClockOffset: 135
  • GPUMemoryTransferRateOffset: 425

[17:18:18][0x00007fa34e7c2b80] Using SSE2: YES
[17:18:18][0x00007fa34e7c2b80] Using AVX: YES
[17:18:18][0x00007fa34e7c2b80] Using AVX2: YES
[17:18:18][0x00007fa34e7c2b80] Benchmarking CUDA worker (CUDA-TROMP) GeForce GTX 1080 Ti (#0) BLOCKS=64, THREADS=64
[17:18:19][0x00007fa34e7c2b80] Benchmark starting… this may take several minutes, please wait…
[17:18:50][0x00007fa34e7c2b80] Benchmark done!
[17:18:50][0x00007fa34e7c2b80] Total time : 31577 ms
[17:18:50][0x00007fa34e7c2b80] Total iterations: 200
[17:18:50][0x00007fa34e7c2b80] Total solutions found: 370
[17:18:50][0x00007fa34e7c2b80] Speed: 6.33372 I/s
[17:18:50][0x00007fa34e7c2b80] Speed: 11.7174 Sols/s

aionminer v0.1.8
Driver: NVIDIA 390.25
GPU: GeForce GTX 1080 Ti
GPUPowerLevel: 127W
GPU not overclocked
[10:03:30][0x00007fd01d089740] Using SSE2: YES
[10:03:30][0x00007fd01d089740] Using AVX: YES
[10:03:30][0x00007fd01d089740] Using AVX2: YES
[10:03:30][0x00007fd01d089740] Benchmarking CUDA worker (CUDA-TROMP) GeForce GTX 950 (#0) BLOCKS=64, THREADS=64
[10:03:31][0x00007fd01d089740] Benchmark starting... this may take several minutes, please wait...
[10:05:14][0x00007fd01d089740] Benchmark done!
[10:05:14][0x00007fd01d089740] Total time : 102517 ms
[10:05:14][0x00007fd01d089740] Total iterations: 200
[10:05:14][0x00007fd01d089740] Total solutions found: 367
[10:05:14][0x00007fd01d089740] Speed: 1.9509 I/s
[10:05:14][0x00007fd01d089740] Speed: 3.57989 Sols/s
aionminer v0.1.8
Driver: NVIDIA 390.12
GPU:  GeForce GTX 950
GPU not overclocked

This is just my normal PC, not built for mining at all.

[17:20:48][0x00007f299936c740] Using SSE2: YES
[17:20:48][0x00007f299936c740] Using AVX: YES
[17:20:48][0x00007f299936c740] Using AVX2: YES
[17:20:48][0x00007f299936c740] Benchmarking CUDA worker (CUDA-TROMP) GeForce GTX 1060 6GB (#0) BLOCKS=64, THREADS=64
[17:20:49][0x00007f299936c740] Benchmark starting… this may take several minutes, please wait…
[17:21:34][0x00007f299936c740] Benchmark done!
[17:21:34][0x00007f299936c740] Total time : 44999 ms
[17:21:34][0x00007f299936c740] Total iterations: 200
[17:21:34][0x00007f299936c740] Total solutions found: 348
[17:21:34][0x00007f299936c740] Speed: 4.44454 I/s
[17:21:34][0x00007f299936c740] Speed: 7.73351 Sols/s

aionminer: v0.1.8
Driver: NVIDIA 390.12
GPU: GeForce GTX 1060 6GB
GPU not overclocked

2:~/Aion$ ./aionminer -cd 1 -cv 1 -cb 64 -ct 64 -b

    ============================= aion reference miner======================
                    Equihash<210,9> CPU&GPU Miner for AION v0.1.8
                    Base on NiceHash equihash miner.
    ============================= aion reference miner======================

Setting log level to 2
[18:40:50][0x00007f754f848740] Using SSE2: YES
[18:40:50][0x00007f754f848740] Using AVX: NO
[18:40:50][0x00007f754f848740] Using AVX2: NO
[18:40:50][0x00007f754f848740] Benchmarking CUDA worker (CUDA-TROMP) GeForce GTX 1080 Ti (#1) BLOCKS=64, THREADS=64
[18:40:51][0x00007f754f848740] Benchmark starting… this may take several minutes, please wait…
[18:41:18][0x00007f754f848740] Benchmark done!
[18:41:18][0x00007f754f848740] Total time : 27441 ms
[18:41:18][0x00007f754f848740] Total iterations: 200
[18:41:18][0x00007f754f848740] Total solutions found: 444
[18:41:18][0x00007f754f848740] Speed: 7.28836 I/s
[18:41:18][0x00007f754f848740] Speed: 16.1802 Sols/s

  • aionminer v0.1.8
  • Driver: NVIDIA 390.25
  • GPU: MSI Geforce GTX 1080 Ti
  • GPUPowerLevel: 168
  • GPUGraphicsClockOffset: 180
  • GPUMemoryTransferRateOffset: 500

A noob here, what command lines do you type in to get this info. I’m still reading the material so if I find I will post my results, if somebody puts it in here I can follow the commands and post our numbers. Thanks

Are you asking about the benchmark itself? The command I ran is at the top

./aionminer -cd 1 -cv 1 -cb 64 -ct 64 -b

If you are asking about the rest of the info at the bottom, I got that from various places. The miner version is in the output of the benchmark, and the GPU, driver, and powerlevel are reported by

nvidia-smi

Lastly, the GPU and Memory offsets are what I set them to, so I knew their vavlues.

2 Likes

./aionminer -cd 0 -cv 1 -cb 64 -ct 64 -b

============================= aion reference miner======================
		Equihash<210,9> CPU&GPU Miner for AION v0.1.8
		Base on NiceHash equihash miner.
============================= aion reference miner======================

Setting log level to 2
[18:39:59][0x00007fdcd2176740] Using SSE2: YES
[18:39:59][0x00007fdcd2176740] Using AVX: YES
[18:39:59][0x00007fdcd2176740] Using AVX2: NO
[18:40:00][0x00007fdcd2176740] Benchmarking CUDA worker (CUDA-TROMP) GeForce GTX 1070 (#0) BLOCKS=64, THREADS=64
[18:40:01][0x00007fdcd2176740] Benchmark starting… this may take several minutes, please wait…
[18:40:46][0x00007fdcd2176740] Benchmark done!
[18:40:46][0x00007fdcd2176740] Total time : 45110 ms
[18:40:46][0x00007fdcd2176740] Total iterations: 200
[18:40:46][0x00007fdcd2176740] Total solutions found: 436
[18:40:46][0x00007fdcd2176740] Speed: 4.43361 I/s
[18:40:46][0x00007fdcd2176740] Speed: 9.66526 Sols/s

aionminer v0.1.8
Driver: NVIDIA 390.12
GPU: GeForce GTX 1070
GPU not overclocked

============================= aion reference miner======================
        Equihash<210,9> CPU&GPU Miner for AION v0.1.8
        Base on NiceHash equihash miner.
============================= aion reference miner======================

Setting log level to 2
[17:58:15][0x00007f9ee3f07740] Using SSE2: YES [17:58:15][0x00007f9ee3f07740] Using AVX: YES [17:58:15][0x00007f9ee3f07740] Using AVX2: YES [17:58:15][0x00007f9ee3f07740] Benchmarking CPU worker (CPU-TROMP-AVX) [17:58:16][0x00007f9ee3f07740] Benchmark starting… this may take several minutes, please wait…
[18:03:15][0x00007f9ee3f07740] Benchmark done!
[18:03:15][0x00007f9ee3f07740] Total time : 299178 ms [18:03:15][0x00007f9ee3f07740] Total iterations: 200 [18:03:15][0x00007f9ee3f07740] Total solutions found: 389 [18:03:15][0x00007f9ee3f07740] Speed: 0.668498 I/s [18:03:15][0x00007f9ee3f07740] Speed: 1.30023 Sols/s

	============================= aion reference miner======================
			Equihash<210,9> CPU&GPU Miner for AION v0.1.8
			Base on NiceHash equihash miner.
	============================= aion reference miner======================

Setting log level to 2
[22:42:55][0x00007effd5e94740] Using SSE2: YES
[22:42:55][0x00007effd5e94740] Using AVX: YES
[22:42:55][0x00007effd5e94740] Using AVX2: YES
[22:42:55][0x00007effd5e94740] Benchmarking CUDA worker (CUDA-TROMP) GeForce GTX 1080 (#0) BLOCKS=64, THREADS=64
[22:42:56][0x00007effd5e94740] Benchmark starting... this may take several minutes, please wait...
[22:43:31][0x00007effd5e94740] Benchmark done!
[22:43:31][0x00007effd5e94740] Total time : 35112 ms
[22:43:31][0x00007effd5e94740] Total iterations: 200
[22:43:31][0x00007effd5e94740] Total solutions found: 348
[22:43:31][0x00007effd5e94740] Speed: 5.69606 I/s
[22:43:31][0x00007effd5e94740] Speed: 9.91114 Sols/s

aionminer: v0.1.8
Driver: NVIDIA 390.30
GPU: GeForce GTX 1080
GPU not overclocked

4 posts were split to a new topic: Support for GPUs

Dear MrYukonC,

could you please provide us some details about overclocking the 1080ti? I’m considering whether I need to overclock the new 1080ti for mining aion, any advice?

I don’t know about overclocking, but I find if I increase my blocks from 64 to 140 in my miner command, this will get you a few more SOLS. 140 seemed like a sweet spot. Changing the processes didn’t seem to help me.

Did you try to increase the number of threads per block or just modify the number of blocks per device?

./aionminer -cd 0 -cv 1 -cb 140 -ct 64 ?

I only had luck adjusting the blocks value. Changing the threads per block just resulted in decreased performance for me.

Yes that command looks correct.

[16:52:35][0x00007f6499e7c740] Using SSE2: YES
[16:52:35][0x00007f6499e7c740] Using AVX: NO
[16:52:35][0x00007f6499e7c740] Using AVX2: NO
[16:52:35][0x00007f6499e7c740] Benchmarking CUDA worker (CUDA-TROMP) GeForce GTX 1080 Ti (#0) BLOCKS=64, THREADS=64
[16:52:35][0x00007f6499e7c740] Benchmarking CUDA worker (CUDA-TROMP) GeForce GTX 1080 Ti (#1) BLOCKS=64, THREADS=64
[16:52:36][0x00007f6499e7c740] Benchmark starting… this may take several minutes, please wait…
[16:52:51][0x00007f6499e7c740] Benchmark done!
[16:52:51][0x00007f6499e7c740] Total time : 15104 ms
[16:52:51][0x00007f6499e7c740] Total iterations: 200
[16:52:51][0x00007f6499e7c740] Total solutions found: 402
[16:52:51][0x00007f6499e7c740] Speed: 13.2415 I/s
[16:52:51][0x00007f6499e7c740] Speed: 26.6155 Sols/s

With unclocked version

./aionminer -cd 0 -cv 1 -cb 64 -ct 256 -b

    ============================= aion reference miner======================
                    Equihash<210,9> CPU&GPU Miner for AION v0.1.10
                    Base on NiceHash equihash miner.
    ============================= aion reference miner======================

Setting log level to 2
[07:05:53][0x00007f025e6b3740] Using SSE2: YES
[07:05:53][0x00007f025e6b3740] Using AVX: YES
[07:05:53][0x00007f025e6b3740] Using AVX2: YES
[07:05:53][0x00007f025e6b3740] Benchmarking CUDA worker (CUDA-TROMP) Tesla V100-SXM2-16GB (#0) BLOCKS=64, THREADS=256
[07:05:54][0x00007f025e6b3740] Benchmark starting… this may take several minutes, please wait…
[07:06:07][0x00007f025e6b3740] Benchmark done!
[07:06:07][0x00007f025e6b3740] Total time : 12575 ms
[07:06:07][0x00007f025e6b3740] Total iterations: 200
[07:06:07][0x00007f025e6b3740] Total solutions found: 389
[07:06:07][0x00007f025e6b3740] Speed: 15.9046 I/s
[07:06:07][0x00007f025e6b3740] Speed: 30.9344 Sols/s

2 Likes

Hi,

I need some feedback from you.
I run the benchmark on two different graphic cards.
GRID K520 provides 3571.43 I/s and 0 Sols/s.
Tesla M60 provides 3.53307 I/s and 7.22513 Sols/s.

	============================= aion reference miner======================
		Equihash<210,9> CPU&GPU Miner for AION v0.2.1
		Base on NiceHash equihash miner.
============================= aion reference miner======================

Setting log level to 2
[10:39:49][0x00007f3a0e771740] Using SSE2: YES
[10:39:49][0x00007f3a0e771740] Using AVX: YES
[10:39:49][0x00007f3a0e771740] Using AVX2: NO
[10:39:49][0x00007f3a0e771740] Benchmarking CUDA worker (CUDA-TROMP) GRID K520 (#0) BLOCKS=64, THREADS=64
[10:39:50][0x00007f3a0e771740] Benchmark starting… this may take several minutes, please wait…
[10:39:50][0x00007f3a0e771740] Benchmark done!
[10:39:50][0x00007f3a0e771740] Total time : 56 ms
[10:39:50][0x00007f3a0e771740] Total iterations: 200
[10:39:50][0x00007f3a0e771740] Total solutions found: 0
[10:39:50][0x00007f3a0e771740] Speed: 3571.43 I/s
[10:39:50][0x00007f3a0e771740] Speed: 0 Sols/s

	============================= aion reference miner======================
		Equihash<210,9> CPU&GPU Miner for AION v0.2.1
		Base on NiceHash equihash miner.
============================= aion reference miner======================

Setting log level to 2
[10:36:37][0x00007fef83812740] Using SSE2: YES
[10:36:37][0x00007fef83812740] Using AVX: YES
[10:36:37][0x00007fef83812740] Using AVX2: YES
[10:36:39][0x00007fef83812740] Benchmarking CUDA worker (CUDA-TROMP) Tesla M60 (#0) BLOCKS=64, THREADS=64
[10:36:40][0x00007fef83812740] Benchmark starting… this may take several minutes, please wait…
[10:37:37][0x00007fef83812740] Benchmark done!
[10:37:37][0x00007fef83812740] Total time : 56608 ms
[10:37:37][0x00007fef83812740] Total iterations: 200
[10:37:37][0x00007fef83812740] Total solutions found: 409
[10:37:37][0x00007fef83812740] Speed: 3.53307 I/s
[10:37:37][0x00007fef83812740] Speed: 7.22513 Sols/s

Can somebody tell me what it means when GRID K520 doesn’t calculate any Solutions. Should I even start mining with this graphicard? Does it make sense?
How should I tune the settings regarding the blocks and threads? How to find out the best settings?

I’m mining on mainnet now.
The CPU miner at least outputs some Sol/s:
Speed [15 sec]: 0.866667 I/s, 1.53333 Sols/s

But the GPU Miner is always like:
Speed [15 sec]: 2433.67 I/s, 0 Sols/s

So is my GPU miner working properly?
It can’t be that CPU is better than GPU.
Please help me!

Did you enter

./aionminer -cd 0 -cv 1 -cb 64 -ct 64 -u {rewards_address}

?