WildRig: new miner for Wild Keccak algo

[andrucrypt]

I have one PC x 1080 ti, 1 PC x 1070, 1 PC x (rx480+rx580), no rig to test, sorry...

Anyway, thanks for response, because of it I clarified a bit statement for hashrate

[nordmann666]

tried intensity 256, 1024 with no changes in hasrate

2x vega56

2,2KH´s

whats wrong?

wildrig.exe --opencl-platform=0 --opencl-devices=0,1 --opencl-launch=0x256 --url=bbr.miningclub.info:7777 --user=xxxxxxxxxxxxxxxxxxxxxxxxxx --pass=x --scratchpad-file=D:\wildrig\scratchpad.bin --scratchpad-url=http://bbr.miningclub.info/scratchpad.bin

[andrucrypt]

tried intensity 256, 1024 with no changes in hasrate

2x vega56

2,2KH´s

whats wrong?

wildrig.exe --opencl-platform=0 --opencl-devices=0,1 --opencl-launch=0x256 --url=bbr.miningclub.info:7777 --user=xxxxxxxxxxxxxxxxxxxxxxxxxx --pass=x --scratchpad-file=D:\wildrig\scratchpad.bin --scratchpad-url=http://bbr.miningclub.info/scratchpad.bin

If you expect 5.5+ MH/s per Vega56 when mining Boolberry, I will disappoint you - because of bigger scratchpad size than in Purk, hashrate here is much less. Otherwise - you use --opencl-launch wrong, try --opencl-launch=256x0 or --opencl-launch=1024x0. First parameter there is intensity, second one - worksize(better to set it by default to zero).

[igotek]

is there any information about power ?

cryptonight or wild keccak has the lower power usage ?

[minerjed]

tried intensity 256, 1024 with no changes in hasrate

2x vega56

2,2KH´s

whats wrong?

wildrig.exe --opencl-platform=0 --opencl-devices=0,1 --opencl-launch=0x256 --url=bbr.miningclub.info:7777 --user=xxxxxxxxxxxxxxxxxxxxxxxxxx --pass=x --scratchpad-file=D:\wildrig\scratchpad.bin --scratchpad-url=http://bbr.miningclub.info/scratchpad.bin

first, you're mining bbr, so you aren't going to get a big hashrate bcause the scratchpad size, but try using

Code:

--opencl-launch=1912x256

and use a -20% power limit, -25% cpu clock an put memory clock at 945MHz, and see if the hashrate goes any better

could you mine BBR with your current settings and post hash rate , im testing with vega FE's but curious on the  56's hashrate on BBR
thanks

[minerjed]

ok  will do

[minerjed]

tried intensity 256, 1024 with no changes in hasrate

2x vega56

2,2KH´s

whats wrong?

wildrig.exe --opencl-platform=0 --opencl-devices=0,1 --opencl-launch=0x256 --url=bbr.miningclub.info:7777 --user=xxxxxxxxxxxxxxxxxxxxxxxxxx --pass=x --scratchpad-file=D:\wildrig\scratchpad.bin --scratchpad-url=http://bbr.miningclub.info/scratchpad.bin

first, you're mining bbr, so you aren't going to get a big hashrate bcause the scratchpad size, but try using

Code:

--opencl-launch=1912x256

and use a -20% power limit, -25% cpu clock an put memory clock at 945MHz, and see if the hashrate goes any better

could you mine BBR with your current settings and post hash rate , im testing with vega FE's but curious on the  56's hashrate on BBR
thanks

2350Kh/s 1 x MSI Vega 56 Retail (Founder Ed)

shoot me an email i cant send any more pm's today lol got a couple things for you to try

[andrucrypt]

I just got this error, first time getting it  don't know if it's a bug from the miner or something that can happen sometimes

this is poolside problem with scratchpad %) In new build miner will try reconnect first and only then redownload scratchpad if it's still a problem.

[andrucrypt]

Released 0.9.10 beta:
- now videocards sorted by BusID
- added luck calculation(need testing, experimental feature)
- fixed crash when logging and --print-full is used
- few more checks for scratchpad fails(also in some cases miner will reconnect if it failed and only then redownload it if this not solve the problem, need testing)

[andrucrypt]

Downloaded, 1rst, what you added is not luck, but effort, luck never is more than 99.9% and can be negative
2nd, I noticed a hash drop on the first seconds with this version...

Checked again. No, this is not an effort Effort higher than 100% means you need to do a lot of hashes to find a share, but if manually increase this number in miner - I get value way below 100%, so this is more like "efficiency", but I like "luck" more Will normalize this value to be more like "luck", when 0% is just normal, above zero - you are lucky and below zero - you are not.

ps: about hashrate drop - how big it is? It can jump a bit for fast cards. Also Purk scratchpad is growing and hasharte will drop more and more, and Vega's advantage will disappear.

[andrucrypt]

well, modified a little the intensity and workload and it works again at 6100 KH/s, but it had dropped to 5600 KH/s 

Hmm, strange, no one else get so high hashdrop. Will PM you later with some reverted changes of kernel to see is this a reason.

the meaning of effort is (hashes for this share) / (hashes expected for this share) * 100

My formula is a bit different, same as in minerd, and if we talk about 1 share it looks like: ( 1 share ) / ( ( hashes for this 1 share ) / ( hashes expected for this 1 share ) ) * 100 I read some articles, and there we can see luck value equal to 100% "in perfect world" ) So will leave everything as is. Just need to improve calculation when dev-fee is finished. Or maybe will replace it with effort as more understandable parameter. Anyway, thanks for some food for thoughts

[andrucrypt]

Released 0.9.11 beta
- new parameter --print-level
- more accurate luck calculation
- dynamic size for scratchpad buffer(no need in recompilation in future, less RAM used for PURK coin to mine )

[andrucrypt]

lol, luck = 100% would be getting a share without doing any hashes, so it's impossible

nah, I don't agree

Today after donation mining I got some shares rejected because "Duplicate Share":
 

This one is bad, will check my code for thread safeness more. How long it was, till new job or just his three shares?

[5BTC]

Anyone get this running on multi Vega rigs? Trying with ROCM on Ubuntu and fails anytime I enable more than 1 Vega

[andrucrypt]

Anyone get this running on multi Vega rigs? Trying with ROCM on Ubuntu and fails anytime I enable more than 1 Vega

Driver crash or miner crash(at what point?)? Or just an error(which one?)?

[Callanta787]

Anyway to make this work on old GPUs?

[andrucrypt]

Anyway to make this work on old GPUs?

Which one? GTX750Ti's works fine. On Radeon 7870 I solved the problem, but need time to add something like kernel option and test everything again.

[andrucrypt]

I got those 3 duplied ones, and then more than 1 minute without shares, so I just restarted the app

I got the same problem, but as soon as it got max-rejects of duplicated shares it automatically reconnected and everything went fine. Anyway, I found a real reason of that problem, will release new version soon with fix for that.

[andrucrypt]

Released 0.9.12 beta with IMPORTANT UPDATE
- fixed "Duplicate share" error after donation mining finished, same was possible to reproduce with pause/resume commands

[andrucrypt]

Released 0.9.13 beta
- reworked strategy a bit, now nvidias should support 0 and 1 too. Changed default strategy to that used in other miners, should be better with vardiff
- fixed processing error "invalid miner height, re-download scratchpad", now miner definitely will redownload it