SRBMiner Cryptonight AMD GPU Miner V1.9.3 - native algo switching

[gyIQ]

Nothing is wrong. Which driver are you using, and did you enable compute mode?

Used 19.2.2, and had 2000+Hs on 1.8.1 but 1100Hs with 1.8.2 then updted driver to 19.3.3. Same results, 1.8.2. hasing only the half of 1.8.1.

Compute mode is unavailable when you have got vega, you can change it with lower cards, like 580X. So stock driver, and HBCC off.

[1679384632]

1679384632

[1679384632]

1679384632

[1679384632]

1679384632

[1679384632]

1679384632

[1679384632]

1679384632

[optii]

doc, regarding loki-turtle , i have 2 gfx in one system, i checked your youtube video. How do i limit this mining to one GFX?

i tried changing specific config file for loki-turtle by adding:

"gpu_conf" :
[
   { "id" : 1, "intensity" : 60 , "worksize" : 8, "threads" : 2},
   
]

but miner instantly shuts down if i add this to cfg file

[doktor83]

doc, regarding loki-turtle , i have 2 gfx in one system, i checked your youtube video. How do i limit this mining to one GFX?

i tried changing specific config file for loki-turtle by adding:

"gpu_conf" :
[
   { "id" : 1, "intensity" : 60 , "worksize" : 8, "threads" : 2},
   
]

but miner instantly shuts down if i add this to cfg file

You probably miss a comma somewhere. It is easier to set it in cmd :

loki-turtle.bat:

Code:

SRBMiner-CN.exe --ccryptonighttype turtle --cgpuid 1 --cgpuintensity 60 --cgpuworksize 8 --cgputhreads 2 --cpool lokiturtle.herominers.com:10521 --cwallet your-loki-wallet --cpassword your-turtle-wallet

A note : if you set a non existent cgpuid it will ignore it, and instead try to load the default config.txt.

[doktor83]

I don't try another algorithm yet, sorry.

V4 has a known bug, that it runs out of memory, so make sure you have a big big virutal mem allocated.

[Iamtutut]

doc, regarding loki-turtle , i have 2 gfx in one system, i checked your youtube video. How do i limit this mining to one GFX?

i tried changing specific config file for loki-turtle by adding:

"gpu_conf" :
[
   { "id" : 1, "intensity" : 60 , "worksize" : 8, "threads" : 2},
   
]

but miner instantly shuts down if i add this to cfg file

You probably miss a comma somewhere. It is easier to set it in cmd :

loki-turtle.bat:

Code:

SRBMiner-CN.exe --ccryptonighttype turtle --cgpuid 1 --cgpuintensity 60 --cgpuworksize 8 --cgputhreads 2 --cpool lokiturtle.herominers.com:10521 --cwallet your-loki-wallet --cpassword your-turtle-wallet

A note : if you set a non existent cgpuid it will ignore it, and instead try to load the default config.txt.

Turtlecoin devs will implement Argon2ID algo in their next fork (height will be 1.8M), will you propose this algo with SRB ?

[doktor83]

doc, regarding loki-turtle , i have 2 gfx in one system, i checked your youtube video. How do i limit this mining to one GFX?

i tried changing specific config file for loki-turtle by adding:

"gpu_conf" :
[
   { "id" : 1, "intensity" : 60 , "worksize" : 8, "threads" : 2},
   
]

but miner instantly shuts down if i add this to cfg file

You probably miss a comma somewhere. It is easier to set it in cmd :

loki-turtle.bat:

Code:

SRBMiner-CN.exe --ccryptonighttype turtle --cgpuid 1 --cgpuintensity 60 --cgpuworksize 8 --cgputhreads 2 --cpool lokiturtle.herominers.com:10521 --cwallet your-loki-wallet --cpassword your-turtle-wallet

A note : if you set a non existent cgpuid it will ignore it, and instead try to load the default config.txt.

Turtlecoin devs will implement Argon2ID algo in their next fork (height will be 1.8M), will you propose this algo with SRB ?

Probably not because its not a CN algo.

[gyIQ]

Nothing is wrong. Which driver are you using, and did you enable compute mode?

Used 19.2.2, and had 2000+Hs on 1.8.1 but 1100Hs with 1.8.2 then updted driver to 19.3.3. Same results, 1.8.2. hasing only the half of 1.8.1.

Compute mode is unavailable when you have got vega, you can change it with lower cards, like 580X. So stock driver, and HBCC off.

Dok, same with 19.4.1.

[doktor83]

Nothing is wrong. Which driver are you using, and did you enable compute mode?

Used 19.2.2, and had 2000+Hs on 1.8.1 but 1100Hs with 1.8.2 then updted driver to 19.3.3. Same results, 1.8.2. hasing only the half of 1.8.1.

Compute mode is unavailable when you have got vega, you can change it with lower cards, like 580X. So stock driver, and HBCC off.

Dok, same with 19.4.1.

Ok, so Vega64, on V4 algo, right?
Are you using text config file or cmd ?

Please paste it here.

If there was such a bug in miner i guess others would also report

[gyIQ]

Ok, so Vega64, on V4 algo, right?
Are you using text config file or cmd ?

Please paste it here.

If there was such a bug in miner i guess others would also report

Correct. 64 on v4.

I've just changed my wallet at pools.txt, everything else is stock. What is weird, because this is working with 1.8.1. But for some reason 1.8.2 breings only 1100Hs, instead of 2000+.

[doktor83]

Ok, so Vega64, on V4 algo, right?
Are you using text config file or cmd ?

Please paste it here.

If there was such a bug in miner i guess others would also report

Correct. 64 on v4.

I've just changed my wallet at pools.txt, everything else is stock. What is weird, because this is working with 1.8.1. But for some reason 1.8.2 breings only 1100Hs, instead of 2000+.

Could you try to paste this in a .bat file and run :

Code:

setx GPU_FORCE_64BIT_PTR 1
setx GPU_MAX_HEAP_SIZE 100
setx GPU_MAX_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_MAX_SINGLE_ALLOC_PERCENT 100
@echo off
cd %~dp0
cls

SRBMiner-CN.exe --ccryptonighttype cryptonight_r --cgpuid 0 --cgpuintensity 120 --cgpuworksize 16 --cgputhreads 2 --cpool xmr-eu1.nanopool.org:14444 --cwallet 4A5hJyu2FvuM2azexYssHW2odrNCNWVqLLmzCowrA57xGJLNufXfzVgcMpAy3YWpzZSAPALhVH4Ed7xo6RZYyw2bUtbm12g --cpassword x
pause

[gyIQ]

sure, just around 19:00. will come back with results, thanks

[chimaeraa]

hi dok.
the new v work fine . almost 1 w with out any crash or ...

thank lot. if i can do something . plz tell me.

[gyIQ]

sure, just around 19:00. will come back with results, thanks

same, 1100Hs

[GKumaran]

sure, just around 19:00. will come back with results, thanks

same, 1100Hs

Code:

SRBMiner-CN.exe --setcomputemode

(sets AMD gpu's to compute mode & disables crossfire - run as admin)

[babkaVkedax]

Cryptonight R - After a minute of work, it gives errors - the wrong algorithm!
And so on all pools!

[alibabacool]

Hi, Doc!
I have a few questions that I would like to clarify:
1. If I use config file, where I specify gpu-only parameters (intensity, threads, worksize), then parameters rebootscriptgpuwatchdog and rebootscriptminrigspeed, which specified via cmd, not working. But if these parameters specify in config file, they are working. So, is it possible to give priority to cmd input over config file?
2. Is it possible to add support for variables for pools file? If you have many rigs you must specify worker name, in pools.txt, on each of them, and it's not comfortable. But worker name can be taken from variable %computername%.

[Amgine]

Dok, could you remove the 120 minutes cap of --maxnosharesent for next release?

Thank you

[doktor83]

Dok, could you remove the 120 minutes cap of --maxnosharesent for next release?

Thank you

I will make that parameter optional, so if it's not used there won't be any default value. Also will set the max to 24h.
This parameter can get 'crazy' when windows re-adjusts time if ntp is used, so it's better if its not on by default.

Hi, Doc!
I have a few questions that I would like to clarify:
1. If I use config file, where I specify gpu-only parameters (intensity, threads, worksize), then parameters rebootscriptgpuwatchdog and rebootscriptminrigspeed, which specified via cmd, not working. But if these parameters specify in config file, they are working. So, is it possible to give priority to cmd input over config file?
2. Is it possible to add support for variables for pools file? If you have many rigs you must specify worker name, in pools.txt, on each of them, and it's not comfortable. But worker name can be taken from variable %computername%.

1. You set everything either in config, or cmd. There is no more mixing of these two since 1.8.0 . So you can set --rebootscriptgpuwatchdog and --rebootscriptminrigspeed

2. Sorry but no. It would take too much time, and i don't think it's such a valuable option.

[gyIQ]

Code:

SRBMiner-CN.exe --setcomputemode

(sets AMD gpu's to compute mode & disables crossfire - run as admin)

Thanks, tried. Same sh!t, 1100Hs

[TheSloppyBeaver]

Hi Dok
   Great Program runs perfectly, got a question, how do you change the --maxnosharesent ? ive tried setting it for 10minutes. i tried -> ( --maxnosharesent 600 ) and then -> ( --maxnosharesent 10) tested both but it did not reset on either setting. ( not a big issue just wanted to know if i was setting it wrong.)

 I also have a request, could you add 1 or even 2 decimal places to avg share time ( 1.xx secs ) to provide more accuracy on hashrate submission for static diff. i would greatly appreciate it.