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

[ZaGunShow]

Its not new to 1.5... but can you explain the kernel setting? I'm getting the periodic hang compiling kernel on GPU 1 (of an 8 card system).

"Some additional parameters you can use in gpu_conf:
"kernel" : 0-4 , IF 0, MINER WILL SELECT MOST SUITABLE KERNEL, OTHERS ARE : 1-FOR GCN CARDS, 2-FOR PRE-GCN CARDS, 3-FOR PRE-GCN EXPERIMENTAL 1, 4-FOR PRE-GCN EXPERIMENTAL 2"

I'm running Vega 64's. I don't know what GCN is referring to. Do I use setting 1?

Thanks

[1632376956]

1632376956

[1632376956]

1632376956

[1632376956]

1632376956

[henri2018]

V1.5.0
- Added option to reset (disable/enable) AMD gpu devices with devcon before mining (admin rights needed, miner will ask)
- Now you can disable/enable up to 19 GPU's by pressing 0 - 9 (gpu0 - gpu9) and shift+0 - shift+9 (gpu10 - gpu19) on the keyboard
- Cryptonight Alloy new algo added (active from block 44.000)
- Added periodical GPU hashrate in log file
- No more nan displayed in hashrate view 

To explain a little bit :

GPU resetting is done with devcon.exe, the same that is used for resetting Vega GPU's with some scripts. It finds all AMD GPU's in the system and disable/enables them before mining starts.
I know there will be some issues with this, like if you do it with a remote connection, it may freeze or drop. But this is something that is not miner related.

Added possibility to disable/enable GPU's. To maximize efficiency, when a device is disabled it won't get disabled right at that moment, it will first finish the round it is working on, and then it will stop working.
This is good for you because if some shares are found, they will be sent to the pool, and only after that the device will get disabled. This can sometimes take even 15-30 sec.
For GPU id's from 0-9 just press the corresponding key, for 10, 11, 12 ..etc use SHIFT+NUMBER, so if you want to disable GPU13, press SHIFT+3 on the keyboard.

Cryptonight Alloy added, altough there is no test pool, so i could not try it first, just did everything by the informations the coin dev gave my. So at block 44.000 we will see is it working

Hi Dok, nice update. In fact, I just upgraded all my 9 rigs to version 1.4.9 .
Just to check, since you already added devcon feature to disable enable AMD card, what will happen if this feature is enabled, and then before we run SRBMiner-CN.exe in start.bat file, we run OverdriveNTool to apply overclock and undervolt? Seems that the setting will be reset as well?

[ZaGunShow]

in my experience, I don't believe enable/disable undoes the under or overvolting, although I do perfer to do it after just to be sure.

I personally have Devcon autostart with windows.. Then OverdriveNTool start after it.

This feature of Devcon to start with SRBMiner will only become useful to me if we also add a restart miner on a drop below hashrate to monitor if the cards need to be enable/renable mid run.. 

[belimedved]

Druze da li moze Awesome Miner da prati monitoring tvog minera?

[doktor83]

Its not new to 1.5... but can you explain the kernel setting? I'm getting the periodic hang compiling kernel on GPU 1 (of an 8 card system).

"Some additional parameters you can use in gpu_conf:
"kernel" : 0-4 , IF 0, MINER WILL SELECT MOST SUITABLE KERNEL, OTHERS ARE : 1-FOR GCN CARDS, 2-FOR PRE-GCN CARDS, 3-FOR PRE-GCN EXPERIMENTAL 1, 4-FOR PRE-GCN EXPERIMENTAL 2"

I'm running Vega 64's. I don't know what GCN is referring to. Do I use setting 1?

Thanks

yes for vegas use kernel 1, that is automatically selected if you omit the kernel parameter.

[ZaGunShow]

sweet.. This is currently the only problem in the autorestart chain

[Atomicc]

Ne zamaraj se . to rade samo zato sto imas dobar majner. oni u svojim opcijama nemaju  namjestavanje temperature sto je veliki plus tvojem mineru(ocito zele vidjet kako si to izveo da ubace u svoje minere). jos ako ubacis da mogu svakoj kartici dodat frekvenciju memorije ili gpu-a pa jos i voltazu onda bi bio savrsen jer  nebi vise nikad morao posezat u wattmine:)

Pa ne znam sto mi to rade al iskreno bas sam se iznervirao. Mada je to bilo za ocekivati, konkurencija je to, uzeo sam im deo kolaca.

Tvoj miner trenutno ima definitivno najbolje rezultate, i konstantno ga razvijaš, i unapredjuješ. Nastavi odličan posao, i ne zamaraj se konkurencijom.  Znaš onu staru: "Za dobrim konjem se prašina diže..."
Waiting for a new  options....

[nighthawk18]

Awesome, Cryptonight-Alloy support was added! I posted a link to your miner on our discord to spread the word.

[UnclWish]

Thanks for update, doc.

Important information: I didn't recommend to update AMD driver to version 18.4.1.
After update one of cards on rigs start to drop hashspeed to 0 after several seconds untill restart mining. Problem not only on my rig! Several peoples in neighbor themes wrotes about the same problem with 18.4.1 driver!

[jmigdlc99]

Switched over from CastXMR to SRBminer today and i love how it's more stable and it has a lower dev fee. Great job overall on the simplified console output as well.

However, my mining hashrates for my Vega56s are hovering around 1750-1800 compared to 1800-1850 when i was using CastXMR. I already have intensity set to 100 and dual thread turned on. Any tips on how to increase hashrate further?

EDIT: Found a problem with the auto restart on startup function: it disables/enables ALL GPUs in the system. I was hoping it would only restart the GPUs i specified in the config. It ended up restarting my other GPUs which were mining ETH.

[doktor83]

Thanks for update, doc.

Important information: I didn't recommend to update AMD driver to version 18.4.1.
After update one of cards on rigs start to drop hashspeed to 0 after several seconds untill restart mining. Problem not only on my rig! Several peoples in neighbor themes wrotes about the same problem with 18.4.1 driver!

Thanks for the info. I really only do testing on the blockchain drivers on Win7/Win10, on the few rigs i have, so this kind of info is welcome!

[doktor83]

Awesome, Cryptonight-Alloy support was added! I posted a link to your miner on our discord to spread the word.

Thanks

[doktor83]

EDIT: Found a problem with the auto restart on startup function: it disables/enables ALL GPUs in the system. I was hoping it would only restart the GPUs i specified in the config. It ended up restarting my other GPUs which were mining ETH.

Yes, this is the way it's working now, it restarts every AMD gpu it finds. I have in plan to move this restarting option to the gpu_conf, but it needs a lot of rework so it will take some time.
Anyways, i know Vega's need this kind of driver restart, but are there any other cards that benefit from this ?

i will add an option to run a script after the restart is done, so you can run overdriventool or something to set your clocks.

[Sevarchik]

Hello guys, i can set fan to constant speed with this miner or i need to use 3rd party software?

[citronick]

doktor83 - any plans to add remote monitoring for your miner?

[doktor83]

Hello guys, i can set fan to constant speed with this miner or i need to use 3rd party software?

You can try :

"gpu_conf" :
[
   { "id" : 0, "intensity" : 0, "worksize" : 8, "threads" : 2, "target_temperature" : 60, "target_fan_speed" : 2000}
]

The "target_fan_speed" is in RPM.

[eduardooaz]

Doctor my RX580 8GB hynix is always stable around 1050 when i manage to run srb since bigin with that values but my RX580 8GB micron is always jumping around 800-1000h/S, do you know any tips to improve this or a reason this should be happening? I increased voltage and still same results

[Mind163]

Hey. And is there an opportunity for an example to write in the first message the average speed, at least one card in all versions of algorithms?
For example, I can share the speed and frequency of several cards of two algorithms:

RX550 2gb heavy 320-370 ipbc 900-1000 V7 480-515
PX460 2gb heavy 360-390 ipbc 900-1000 V7 475-505
PX460 4gb heavy 550-615  ipbc 1000-1150
PX570 4GB heavy 700-800 ipbc 1800-1950 V7 890-910
PX580 4GB heavy 700-800 ippc 1850-2050 V7 950-985
PX580 8GB (Sams) heavy 1000-1100 ipbc 1900-2100
Vega 56 8 GB Heavy 1380-1550 Ipbc 3650

[rodyw]

Hi, maybe this question is asked before, but how should the fast switching ('p') work? As soon as I hit 'p' the miner tries to switch to the other pool, but I got errors 9 out of 10 times:
[2018-05-01 10:00:14] Switching to next pool from pools config
[2018-05-01 10:00:19] Connecting to sumo.pandapool.nl:5555
[2018-05-01 10:00:19] Connecting to sumo.pandapool.nl:5555
[2018-05-01 10:00:19] Connected!
[2018-05-01 10:00:20] Could not login user to pool
[2018-05-01 10:00:20] ERROR : A request to send or receive data was disallowed because the socket is not connected and (when sending on a datagram socket using a sendto call) no address was supplied.
[2018-05-01 10:00:20] Can't connect to sumo.pandapool.nl:5555
[2018-05-01 10:00:20] Switching to next pool from pools config

Sometimes it works once, but as soon I try to switch later on I got these errors again and have to restart the miner.

[xlm19877]

hello
 when i run it om my rigs

all time the program close

i try to underclock and do all things but all time the program close

what i can do for that ??

need much more info, if you want help.

hello

what you want info

my card is 470x4
580x2

i try to undervolt to stock and same problem