WildRig: new miner for Wild Keccak algo

[andrucrypt]

Released 0.9.23 beta
- fixed crazy avg. effort after donation mining
- fixed "Duplicate share" sending in rare cases(was possible on fast rigs, when new job is not arriving too long)
- fixed crash when --opencl-devices was set not like 0,1,2,3 and so on

[1714892310]

1714892310

[1714892310]

1714892310

[1714892310]

1714892310

[1714892310]

1714892310

[1714892310]

1714892310

[pwhite11798]

I am having problems with this miner when started it says error reading scratchpad.bin: No error
downloading new scratch pad from scratch. then the miner does absolutely nothing. Am i missing something

[andrucrypt]

I am having problems with this miner when started it says error reading scratchpad.bin: No error
downloading new scratch pad from scratch. then the miner does absolutely nothing. Am i missing something

Is there rights for writing to folder where miner unpacked? What path you specified for --scratchpad-file?

[andrucrypt]

Released 0.9.24 beta
- fixed rare case when one of 2+ shares solved at once(that green "yay!" messages) can be ignored
- now miner shows difficulty sent from pool, instead of calculation based on target value
- reverted back validation of share(hash will be checked for "<= target", instead of "< target")

[pwhite11798]

I am having problems with this miner when started it says error reading scratchpad.bin: No error
downloading new scratch pad from scratch. then the miner does absolutely nothing. Am i missing something

Is there rights for writing to folder where miner unpacked? What path you specified for --scratchpad-file?

i figured it out. Thanks i am now hashing @38 MH/s

[pwhite11798]

the command opencl-launch is the intensity tuning, default is 256x0 what effect does it have on the miner if you increase the x0 part of the comand. Sorry im not a command line junky maybe some one will get a laugh out of this.

[andrucrypt]

the command opencl-launch is the intensity tuning, default is 256x0 what effect does it have on the miner if you increase the x0 part of the comand. Sorry im not a command line junky maybe some one will get a laugh out of this.

In most cases it doesn't help, and OpenCL decides better But when mining BBR coin, using 128x32 seems to be better option on Vega 56/64 for some miners(I saw how it boost hashrate from ~4000kH/s to ~4200kH/s).

[skygonga]

My 570-4G Samsung, BBR speed 1.65M, why is the speed not 2M? Memory speed 2000, core 1200

[andrucrypt]

Released 0.9.25 beta
- reworked share validation, now calculation are the same as on pool side
- after donation mining miner will reconnect to pool now
- implemented --send-stale parameter to send stale shares when new job arrived without scratchpad update
- removed restart miner when watchdog enabled and miner found not valid share on CPU(in 0.9.24 it was a mistake...)
- extended info for invalid share

NOTE: Next week I won't be available, so use older versions in case of instability/etc. with latest one.

[xscxsc]

Hello

I use "SAPPHIRE NITRRO+RADEON RX VEGA 56", GPU is 1407Mhz 950mV, Memory is 905 Mhz 905mV. Use AMD blockchain driver（Aug 23）. Run Purk only 5400+KH/s(--opencl-launch=256x0). How do you set it to 6000KH/s?

I use --opencl-launch=1912x256, Run Purk only 5580KH/s.

[tbearhere]

Now I got this. Is that ok? Using 0.9.25.
[21:21:56] GPU #0: yay! found 3 shares at once

[schriss]

Now I got this. Is that ok? Using 0.9.25.
[21:21:56] GPU #0: yay! found 3 shares at once

Released 0.9.24 beta
- fixed rare case when one of 2+ shares solved at once(that green "yay!" messages) can be ignored

[andrucrypt]

Now I got this. Is that ok? Using 0.9.25.
[21:21:56] GPU #0: yay! found 3 shares at once

Yes, this is fine, this means gpu found 3 shares(nonces for them) at once Probably because of low difficulty you set, or you just started to mine.

ps: but I remember one case with driver crash because of teamviwer, such messages were infinite, and this is not fine

[igotek]

I use HiveOS for my NVIDIA GPUs. I have over 150 GPU of NVIDIA.
I can switch to Windows if wildkeccak is more profitable.

is there any info of wildkeccak algorithm about some graphic cards ?
I need to learn power consumption and hashrate.

GTX1050TI 4GB
GTX1060 3GB
GTX1070 8GB

[mrmiroslav]

I use HiveOS for my NVIDIA GPUs. I have over 150 GPU of NVIDIA.
I can switch to Windows if wildkeccak is more profitable.

is there any info of wildkeccak algorithm about some graphic cards ?
I need to learn power consumption and hashrate.

GTX1050TI 4GB
GTX1060 3GB
GTX1070 8GB

1060 3gb - 1500kh

[dwarf0]

mrmiroslav

memory type and clock speed?

[jaryfaaad]

It is really a good try but i am still not able to run the 0.9.6 now.

[andrucrypt]

I'm back now. Released 0.10.0 beta:
- improved precision for share validation, now more shares should be solved again
- now miner can be started without specifying all devices(and --opencl-platform for AMD), but don't forget to specify --opencl-launch or --cuda-launch
- disabled OpenCL mode for NVIDIA

[Iamtutut]

How to find a list of minable coins with wild keccack please ? I'm very unfamiliar with this algo. Thanks.

[tbearhere]

Now I got this. Is that ok? Using 0.9.25.
[21:21:56] GPU #0: yay! found 3 shares at once

Yes, this is fine, this means gpu found 3 shares(nonces for them) at once Probably because of low difficulty you set, or you just started to mine.

ps: but I remember one case with driver crash because of teamviwer, such messages were infinite, and this is not fine

Teamviewer causing driver crashes .... hummm I guess the best thing to do is disabled teamviewer running in the background with resource monitor. At least on the main computer accessing the rig.