SILENTARMY v5: Zcash miner, 115 sol/s on R9 Nano, 70 sol/s on GTX 1070

[penambang]

 

single R9-290:


Great job!!

[1714972221]

1714972221

[1714972221]

1714972221

[1714972221]

1714972221

[1714972221]

1714972221

[1714972221]

1714972221

[nerdralph]

Marc, have you considered making NR_ROWS_LOG fixed at 20, and clearing cnt for each round?  Then each row would have only collisions at each round, avoiding the need to search for collisions.  Then NR_SLOTS/OVERHEAD could be significantly reduced.
To avoid time penalty for clearing the cnt values, instead of calling kernel_init_ht at each round, you could zero the count after checking the row.  When the DRAM page is already open to check the cnt value, the time cost of doing a write back to the same page (in fact, same 64-byte cache line) is minimal.

Yeah NR_ROWS_LOG is pretty much always compiled at 20. But I have to offer the other options because people want to mine with GPUs having very little memory. And I don't know if you checked the latest commits, but OVERHEAD has been lowered to 9 so 1 Equihash instance needs only 1.2 GB, and I recently made the exact change you suggested (clearing the counter after we are done using/reading it).

I think the vast majority (>95%) of miners have cards with at least 2GB RAM.  The kernel I designed (and only implemented a bit of prototype code) is quite similar to yours, but only supporting 2^20 sorting bins (or rows).  My design has 2 tables like yours (I consider it to be like double-buffering), but I hadn't thought of dividing the saved indexes between the two tables so they could fit in 28 bytes and leaving 4 bytes for the collision counter.
I did see the change of OVERHEAD from 13 to 9, but I hadn't noticed the counter reset at the end of the round.
https://github.com/mbevand/silentarmy/blob/master/input.cl#L568

Since you say you have to support options other than 2^20 rows, I'll probably fork your code and optimize it for 2^20.  I think it should give another 10% performance boost.  I also think even more performance can be attained by optimizing for the 256-byte stride size of the Polaris, Tonga, and Pitcairn GPUs.  Despite your comment in the code about odd values of OVERHEAD being best  for avoiding channel conflicts, I had found 12 was faster than 13 in my testing on Tonga and Pitcairn.

[elgi76]

i have installed silentarmy, no error with compilation but i have this result :

# ./silentarmy
Traceback (most recent call last):
  File "./silentarmy", line 542, in <module>
    main()
  File "./silentarmy", line 539, in main
    Silentarmy(opts).run()
  File "./silentarmy", line 249, in run
    asyncio.ensure_future(self.reconnect())
AttributeError: 'module' object has no attribute 'ensure_future'

[greaterninja]

I'm guessing you are using ubuntu 14.04 or a older version of python 3.4x.

remove the installation directory and all files

Code:

rm -rf silentarmy

then rebuild it.

Code:

sudo add-apt-repository ppa:fkrull/deadsnakes -y
sudo add-apt-repository ppa:chris-lea/libsodium -y
sudo apt-get update
sudo apt-get install -y python3.5 libsodium-dev
git clone https://github.com/mbevand/silentarmy
cd silentarmy
make
sed -i 's/python3/python3.5/' silentarmy

then see if it runs by typing

Code:

./silentarmy --list

OR

Code:

./silentarmy --use 0,1,2,3,4,5,6 -c stratum+tcp://us1-zcash.flypool.org:3333 -u twalletaddress.WorkerName

If you have 2gb card use

Code:

./silentarmy --instances 1 --use 0,1,2,3,4,5,6 -c stratum+tcp://us1-zcash.flypool.org:3333 -u twalletaddress.WorkerName

Credit mainly goes to n1koo for the fix/workaround.  His response is here https://gist.github.com/n1koo/7c0455b674d3c3a9d4cccf53837b270e

[nerdralph]

i have installed silentarmy, no error with compilation but i have this result :

# ./silentarmy
Traceback (most recent call last):
  File "./silentarmy", line 542, in <module>
    main()
  File "./silentarmy", line 539, in main
    Silentarmy(opts).run()
  File "./silentarmy", line 249, in run
    asyncio.ensure_future(self.reconnect())
AttributeError: 'module' object has no attribute 'ensure_future'

RTFM:
"(beware: the silentarmy miner makes use of Python's ensure_future() which requires Python 3.4.4, however Ubuntu 14.04 ships 3.4.3, therefore only the sa-solver tool is usable on Ubuntu 14.04.)"

[elgi76]

thanks for your response.
but it don't work

# ./silentarmy -v
Connecting to us1-zcash.flypool.org:3333
Solver 0.0: launching
Solver 0.1: launching
Successfully connected to us1-zcash.flypool.org:3333
Received target 0083126e978d4fdf3b645a1cac083126e978d4fdf3b645a1cac083126e978d4f
Received job "42d07077cb3c893956d5"
Stratum server sent us the first job
Mining on 1 device
Received job "28db6dff8fbd87818695"
Received job "62506bf236bebd0d23da"
Received job "7ce46b6bd1bb37050bc8"
Received job "8b82bf18169fc28e2ede"
Received job "f23fc5bf6dbcd3a4c879"
Received job "3f7131043545762a5e36"
Received job "25e96b0d6c35854577af"
Received job "6542d22209fed3f3fbd3"

[espressodelisi]

I am using ubuntu 14.04 on my 4GPU rig (3x radeon 280x and 1 270)

when I silenarmy it stucks, so I had to add config option to use only 1 instance. Now its running fine and it makes total 67 sols. I do not know if that is good.
When I look at gpu utilisation on atitweak, O see utilization is low. why?

here is the output of atitweak:

tim@rig0:~$ .local/bin/atitweak -s
0. AMD Radeon R9 200 Series      (:0.0)
    engine clock 1070MHz, memory clock 1600MHz, core voltage 1.2VDC, performance level 3, utilization 75%
    fan speed 23% (994 RPM) (default)
    temperature 56 C
    powertune 0%
1. AMD Radeon R9 200 Series                           (:0.1)
    engine clock 1150MHz, memory clock 1500MHz, core voltage 1.175VDC, performance level 3, utilization 73%
    fan speed 20% (1070 RPM) (default)
    temperature 52 C
    powertune 0%
2. AMD Radeon (TM) R9 200 Series (:0.2)
    engine clock 975MHz, memory clock 1400MHz, core voltage 1.215VDC, performance level 3, utilization 79%
    fan speed 34% (1101 RPM) (default)
    temperature 54 C
    powertune 0%
3. AMD Radeon R9 200 Series                           (:0.3)
    engine clock 1150MHz, memory clock 1500MHz, core voltage 1.175VDC, performance level 3, utilization 72%
    fan speed 21% (1024 RPM) (default)
    temperature 54 C
    powertune 0%


any suggestions?

[nerdralph]

Running nicely on a rig with 2x R9 380 4GB and 3x R7 370 2GB with a small tweak to the kernel:
Total 125.8 sol/s [dev0 33.7, dev1 23.4, dev2 26.9, dev3 25.0, dev4 23.2] 23 shares
Total 125.2 sol/s [dev0 31.8, dev1 23.6, dev2 25.5, dev3 25.4, dev4 23.4] 23 shares

This is on Ubuntu 14.04/fglrx with python3 and libsodium installed from PPA.

[willrain]

https://paste.ubuntu.com/23425857/
It's on a 7 GPU rig.

First run:

Code:

21724 rain      20   0  237m  66m  28m S  21,2  1,7   0:20.57 /mnt/rw/rain/sa-solver --mining --use 6                                                          
21717 rain      20   0  237m  67m  28m S  19,5  1,7   0:21.42 /mnt/rw/rain/sa-solver --mining --use 3                                                          
21721 rain      20   0  237m  67m  28m S  18,8  1,7   0:21.42 /mnt/rw/rain/sa-solver --mining --use 5                                                          
21713 rain      20   0  238m  67m  27m R  18,5  1,7   0:18.85 /mnt/rw/rain/sa-solver --mining --use 1                                                          
21715 rain      20   0  238m  67m  28m S  18,5  1,7   0:21.15 /mnt/rw/rain/sa-solver --mining --use 2                                                          
21714 rain      20   0  237m  66m  27m S  17,9  1,7   0:18.69 /mnt/rw/rain/sa-solver --mining --use 1                                                          
21719 rain      20   0  237m  66m  28m S  17,2  1,7   0:19.81 /mnt/rw/rain/sa-solver --mining --use 4                                                          
21711 rain      20   0  240m  68m  27m S  16,9  1,7   0:18.39 /mnt/rw/rain/sa-solver --mining --use 0                                                          
21712 rain      20   0  240m  68m  27m S  16,5  1,7   0:18.23 /mnt/rw/rain/sa-solver --mining --use 0

On second run GPU6 left without sa-solver thread, on third it was GPU5.

[dagarair]

/sa-solver' biinary; make sure to run 'make' to compile it.

How can i fix that?  I ran make and no file found.

[cryptomined]

https://paste.ubuntu.com/23425857/
It's on a 7 GPU rig.

First run:

Code:

21724 rain      20   0  237m  66m  28m S  21,2  1,7   0:20.57 /mnt/rw/rain/sa-solver --mining --use 6                                                          
21717 rain      20   0  237m  67m  28m S  19,5  1,7   0:21.42 /mnt/rw/rain/sa-solver --mining --use 3                                                          
21721 rain      20   0  237m  67m  28m S  18,8  1,7   0:21.42 /mnt/rw/rain/sa-solver --mining --use 5                                                          
21713 rain      20   0  238m  67m  27m R  18,5  1,7   0:18.85 /mnt/rw/rain/sa-solver --mining --use 1                                                          
21715 rain      20   0  238m  67m  28m S  18,5  1,7   0:21.15 /mnt/rw/rain/sa-solver --mining --use 2                                                          
21714 rain      20   0  237m  66m  27m S  17,9  1,7   0:18.69 /mnt/rw/rain/sa-solver --mining --use 1                                                          
21719 rain      20   0  237m  66m  28m S  17,2  1,7   0:19.81 /mnt/rw/rain/sa-solver --mining --use 4                                                          
21711 rain      20   0  240m  68m  27m S  16,9  1,7   0:18.39 /mnt/rw/rain/sa-solver --mining --use 0                                                          
21712 rain      20   0  240m  68m  27m S  16,5  1,7   0:18.23 /mnt/rw/rain/sa-solver --mining --use 0

On second run GPU6 left without sa-solver thread, on third it was GPU5.

my 7 GPU rig has been running solid for the last 5+ hours ever since I fired it up.  No dropped GPUs ... No crash 

[ZenFr]

I have no AMD GPU.
I can not wait for the next version is to provide support for NVidia cards :-).

[Arceny]

Hi, using ubuntu 14.04 + amd gpu pro 16.30 on 6-card-rig (stable on ethereum + claymore) rig I have 0.0 (zero) sols after few minutes of mining. Hardware is Asrock H97 anniversary + RX470 refs.

Code:

Total 197.1 sol/s [dev0 39.4, dev1 38.8, dev2 0.0, dev3 37.5, dev4 37.8, dev5 44.1] 35 shares
Total 197.3 sol/s [dev0 38.0, dev1 40.5, dev2 0.0, dev3 37.0, dev4 32.7, dev5 44.0] 35 shares
Total 194.5 sol/s [dev0 37.4, dev1 39.0, dev2 0.0, dev3 38.4, dev4 27.4, dev5 45.5] 35 shares
Total 192.3 sol/s [dev0 38.3, dev1 39.0, dev2 0.0, dev3 39.3, dev4 24.9, dev5 45.4] 35 shares
Total 190.8 sol/s [dev0 37.0, dev1 40.0, dev2 0.0, dev3 38.3, dev4 21.0, dev5 45.6] 35 shares
Total 191.1 sol/s [dev0 38.9, dev1 39.2, dev2 0.0, dev3 37.6, dev4 16.8, dev5 46.0] 35 shares
Total 189.3 sol/s [dev0 37.6, dev1 38.2, dev2 0.0, dev3 39.5, dev4 13.0, dev5 43.1] 35 shares
Total 187.1 sol/s [dev0 37.2, dev1 36.8, dev2 0.0, dev3 40.4, dev4 8.9, dev5 43.8] 35 shares
Total 187.2 sol/s [dev0 36.9, dev1 40.1, dev2 0.0, dev3 38.9, dev4 6.7, dev5 42.1] 35 shares
Total 186.2 sol/s [dev0 37.4, dev1 39.7, dev2 0.0, dev3 39.1, dev4 2.5, dev5 43.5] 35 shares
Total 184.4 sol/s [dev0 37.6, dev1 38.2, dev2 0.0, dev3 37.7, dev4 0.0, dev5 42.2] 35 shares
Total 183.4 sol/s [dev0 36.3, dev1 34.3, dev2 0.0, dev3 41.4, dev4 0.0, dev5 41.2] 35 shares

If stop and restart miner threads could not been initialized, so I need to reboot. I am using zec.nanopool.org

Here is dmesg output: http://pastebin.com/fg5ZEiai

[nerdralph]

Hi, using ubuntu 14.04 + amd gpu pro 16.30 on 6-card-rig (stable on ethereum + claymore) rig I have 0.0 (zero) sols after few minutes of mining. Hardware is Asrock H97 anniversary + RX470 refs.

Code:

Total 197.1 sol/s [dev0 39.4, dev1 38.8, dev2 0.0, dev3 37.5, dev4 37.8, dev5 44.1] 35 shares
Total 197.3 sol/s [dev0 38.0, dev1 40.5, dev2 0.0, dev3 37.0, dev4 32.7, dev5 44.0] 35 shares
Total 194.5 sol/s [dev0 37.4, dev1 39.0, dev2 0.0, dev3 38.4, dev4 27.4, dev5 45.5] 35 shares
Total 192.3 sol/s [dev0 38.3, dev1 39.0, dev2 0.0, dev3 39.3, dev4 24.9, dev5 45.4] 35 shares
Total 190.8 sol/s [dev0 37.0, dev1 40.0, dev2 0.0, dev3 38.3, dev4 21.0, dev5 45.6] 35 shares
Total 191.1 sol/s [dev0 38.9, dev1 39.2, dev2 0.0, dev3 37.6, dev4 16.8, dev5 46.0] 35 shares
Total 189.3 sol/s [dev0 37.6, dev1 38.2, dev2 0.0, dev3 39.5, dev4 13.0, dev5 43.1] 35 shares
Total 187.1 sol/s [dev0 37.2, dev1 36.8, dev2 0.0, dev3 40.4, dev4 8.9, dev5 43.8] 35 shares
Total 187.2 sol/s [dev0 36.9, dev1 40.1, dev2 0.0, dev3 38.9, dev4 6.7, dev5 42.1] 35 shares
Total 186.2 sol/s [dev0 37.4, dev1 39.7, dev2 0.0, dev3 39.1, dev4 2.5, dev5 43.5] 35 shares
Total 184.4 sol/s [dev0 37.6, dev1 38.2, dev2 0.0, dev3 37.7, dev4 0.0, dev5 42.2] 35 shares
Total 183.4 sol/s [dev0 36.3, dev1 34.3, dev2 0.0, dev3 41.4, dev4 0.0, dev5 41.2] 35 shares

If stop and restart miner threads could not been initialized, so I need to reboot. I am using zec.nanopool.org

Here is dmesg output: http://pastebin.com/fg5ZEiai

You're probably overclocking your cards and/or have flashed a custom BIOS.  zcash mining is a lot different than eth mining.
https://bitcointalk.org/index.php?topic=1670132.0

[uray]

compiled binary in case some one need it
compiled with CFLAGS="-O3 -march=x86-64" on ubuntu 16.04

http://files.etherlink.co/silentarmy-v3-x86_64.tar.xz

[elgi76]

after a reboot it work fine

Total 253.5 sol/s [dev0 47.7, dev1 36.6, dev2 36.8, dev3 41.8, dev4 41.2, dev5 46.1] 495 shares
Total 254.4 sol/s [dev0 47.7, dev1 37.9, dev2 36.1, dev3 40.8, dev4 40.1, dev5 44.0] 495 shares
Total 253.5 sol/s [dev0 47.4, dev1 39.2, dev2 35.8, dev3 40.0, dev4 39.7, dev5 44.4] 495 shares
Total 252.8 sol/s [dev0 45.9, dev1 39.6, dev2 36.7, dev3 38.4, dev4 41.4, dev5 43.0] 495 shares
Total 252.8 sol/s [dev0 46.7, dev1 38.6, dev2 38.4, dev3 37.8, dev4 41.3, dev5 45.5] 496 shares
Total 253.3 sol/s [dev0 47.2, dev1 39.0, dev2 38.4, dev3 38.8, dev4 42.4, dev5 44.2] 497 shares
Total 253.6 sol/s [dev0 44.4, dev1 40.1, dev2 38.7, dev3 38.4, dev4 40.7, dev5 43.6] 497 shares
Total 253.0 sol/s [dev0 44.7, dev1 40.1, dev2 37.9, dev3 37.9, dev4 42.4, dev5 42.0] 498 shares

[number222222222222]

Been running it since release. Ubuntu 15.04. Six 280X show 202 sol/s flypool shows 185 sol/s
2 380s 390X and 4 280X does 185 sol/s flypool shows 185 sol/s. Old 7990 and 7970 does 70 sol/s

[Arceny]

You're probably overclocking your cards and/or have flashed a custom BIOS.  zcash mining is a lot different than eth mining.
https://bitcointalk.org/index.php?topic=1670132.0

No, I have stock bioses/clocks....

[nerdralph]

You're probably overclocking your cards and/or have flashed a custom BIOS.  zcash mining is a lot different than eth mining.
https://bitcointalk.org/index.php?topic=1670132.0

No, I have stock bioses/clocks....

Then you probably have shitty risers.  I bet any cards plugged directly into a 16x slot work fine.

[toptek]

cool got all three working im at

Code:

Total 129.6 sol/s [dev0 47.3, dev1 40.4, dev2 39.9] 247 shares
Total 129.2 sol/s [dev0 45.4, dev1 41.4, dev2 40.2] 247 shares
Total 129.5 sol/s [dev0 46.1, dev1 41.9, dev2 40.0] 247 shares
Total 128.7 sol/s [dev0 43.9, dev1 41.2, dev2 38.5] 248 shares
Total 129.2 sol/s [dev0 43.7, dev1 41.3, dev2 39.4] 248 shares
Total 128.0 sol/s [dev0 42.5, dev1 40.9, dev2 40.4] 250 shares
Total 127.4 sol/s [dev0 41.4, dev1 41.0, dev2 39.3] 250 shares
Total 127.1 sol/s [dev0 40.7, dev1 40.2, dev2 38.8] 250 shares
Total 126.1 sol/s [dev0 40.8, dev1 40.3, dev2 39.6] 252 shares
Total 126.3 sol/s [dev0 40.3, dev1 42.2, dev2 40.4] 252 shares
Total 125.5 sol/s [dev0 39.6, dev1 42.4, dev2 40.5] 252 shares
Total 124.7 sol/s [dev0 39.5, dev1 41.3, dev2 40.2] 253 shares
Total 124.2 sol/s [dev0 40.2, dev1 39.1, dev2 40.9] 254 shares

with 3 RX470 about 41 sol each is about right , now im at the right speed per card.

thanks  top