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

[Csouza]

I have a little problem.
I'm using srb, for more than a month on my vega rigs and it works perfectly with auto settings.

I want to mine with my home pc which has an rx570 4gb in it. If I start it wit auto settings it quits instantly and says to set an intensity level.
If I set an intensity, thread, work etc.. it starts but 0 h/s. I have no idea what happens. Stock GPU setting at first.

Any idea?

Code:

[2019-04-13 09:46:22] Startup monitor attached
[2019-04-13 09:46:22] Miner version: 1.8.2
[2019-04-13 09:46:22] Windows version: 10.0 build 17763
[2019-04-13 09:46:23] Video driver version: 25.20.15027.5007
[2019-04-13 09:46:23] AMD Platform ID: 0
[2019-04-13 09:46:23] AMD platform FOUND
[2019-04-13 09:46:23] Found 1 AMD devices
[2019-04-13 09:46:23] CPU AES-NI: TRUE
[2019-04-13 09:46:23] PREPARE DATA: FALSE
[2019-04-13 09:46:23] GPU0: Radeon RX 570 Series [ellesmere] [4096 MB][I: 112.0][W: 8][T: 2][F: 8][BUS: 16]
[2019-04-13 09:46:23] ADL is enabled
[2019-04-13 09:46:23] Algorithm: Cryptonight V4 (R)
[2019-04-13 09:46:23] Algo switching: disabled
[2019-04-13 09:46:23] Startup monitor: enabled
[2019-04-13 09:46:23] Gpu watchdog: enabled [normal mode]
[2019-04-13 09:46:23] Coin forking: disabled
[2019-04-13 09:46:23] Starting init of gpu threads
[2019-04-13 09:46:23] Created OCL context
[2019-04-13 09:46:23] Using fragments 8 for DeviceID 0 (Thread 0)
[2019-04-13 09:46:23] Using heavy_mode 1 for DeviceID 0 (Thread 0)
[2019-04-13 09:46:23] Using thread_delay 319 for DeviceID 0 (Thread 0)
[2019-04-13 09:46:23] Created OCL input buffer for DeviceID 0 (Thread 0)
[2019-04-13 09:46:23] ctx->headers[0] created for DeviceID 0 (Thread 0)
[2019-04-13 09:46:23] ctx->Program created for DeviceID 0 (Thread 0)
[2019-04-13 09:46:23] Compiling [normalv4] kernel for DEVICE BUS_ID[16]
[2019-04-13 09:46:24] Binary kernel for DeviceID 0 (Thread 0) saved to disk
[2019-04-13 09:46:24] Using fragments 8 for DeviceID 0 (Thread 1)
[2019-04-13 09:46:24] Using heavy_mode 1 for DeviceID 0 (Thread 1)
[2019-04-13 09:46:24] Using thread_delay 319 for DeviceID 0 (Thread 1)
[2019-04-13 09:46:24] Created OCL input buffer for DeviceID 0 (Thread 1)
[2019-04-13 09:46:24] Loading [normalv4] kernel for DEVICE BUS_ID[16]
[2019-04-13 09:46:24] ctx->Program for DeviceID 0 (Thread 1) loaded
[2019-04-13 09:46:24] ADL initialised successfully!
[2019-04-13 09:46:24] Finished init of gpu threads
[2019-04-13 09:46:24] Startup monitor detached
[2019-04-13 09:46:24] Socket initialiasation done
[2019-04-13 09:46:24] Gpu watchdog started
[2019-04-13 09:46:24] API disabled

for 4gb cards use intensity beteween 57~59

[ingyenfrag]

Thx! It works, but ~730h/s seems a little bit low for me.
What are the optimum settings for RX570-s. Asus ROG STRIX 570 4GB with 1500+ memory timing bios mod.

Intensity set to 58.
1250/2000MHz GPU/MEM  @ 0.9V (0.8875V)  , MEM maxed out for these timings.

I rather want to see it around 900 H/s.

edit.: 19.3.2 Adrenaline, monero v4

[viljy]

Today I updated the driver to Adrenalin 2019 19.4.1 and nothing has changed.
The previous driver was Crimson. Strange, wasn't it supposed to speed up?
This is the uplexa mining.

[xmr8218]

hi, the miner always show this error "couldn't login to pool" when my internet drop for 1 or 2 second then reconnected with new IP(dynamic IP). I tried manually switch pool O/P option but same error still appear, but if i manually close and restart the miner then the miner can connect to pool without error.

My miner setup as follow:
1. miner IP are fixed by the router
2. miner connected to giga switch then only to router
3. miner uses SSL/TLS connection to pool with keep alive turn off. I tried turn on keep alive option but have the same issue.

any idea?

a log file would be really helpful

hi, attached is the log file https://drive.google.com/file/d/106A65dEHSQHhRxPkKcYd1VrAWlxF9dNd/view?usp=sharing
thanks alot

[mk111]

Dok - any plans to implement the AMD Memory tweak into SRB?
https://bitcointalk.org/index.php?topic=5123724.0

[doktor83]

hi, the miner always show this error "couldn't login to pool" when my internet drop for 1 or 2 second then reconnected with new IP(dynamic IP). I tried manually switch pool O/P option but same error still appear, but if i manually close and restart the miner then the miner can connect to pool without error.

My miner setup as follow:
1. miner IP are fixed by the router
2. miner connected to giga switch then only to router
3. miner uses SSL/TLS connection to pool with keep alive turn off. I tried turn on keep alive option but have the same issue.

any idea?

a log file would be really helpful

hi, attached is the log file https://drive.google.com/file/d/106A65dEHSQHhRxPkKcYd1VrAWlxF9dNd/view?usp=sharing
thanks alot

I see two things in the log : time shifting, and maybe a bug in the main pool reconnection mechanism.

If you can turn off automatic time setup, and set it manually that would be good. When windows re-adjusts time BACK that can mess up a lot of things, miner thinks a day passed..

The second, main pool reconnection, i will test and fix if needed. It was implemented a long time ago, and a lot of stuff changed since, so maybe it's not working as it should.

[doktor83]

Dok - any plans to implement the AMD Memory tweak into SRB?
https://bitcointalk.org/index.php?topic=5123724.0

I have something in plan, but you can already use Eliovp's tweaker with SRB, he pushed a windows binary  

[dulama]

Hi,

Please fix "Algorithm: Cryptonight V4 (R) - after 1 minute of work produces multiple errors!"

Thanks,
Vlad

[lncm]

Why remove webchain?

Webchain fork to Lyra2-webchain algorithm and not working with cryptonight-webchain. New webchain miner works on CPU only.

Didn't know that, thanks.

[xmr8218]

hi, the miner always show this error "couldn't login to pool" when my internet drop for 1 or 2 second then reconnected with new IP(dynamic IP). I tried manually switch pool O/P option but same error still appear, but if i manually close and restart the miner then the miner can connect to pool without error.

My miner setup as follow:
1. miner IP are fixed by the router
2. miner connected to giga switch then only to router
3. miner uses SSL/TLS connection to pool with keep alive turn off. I tried turn on keep alive option but have the same issue.

any idea?

a log file would be really helpful

hi, attached is the log file https://drive.google.com/file/d/106A65dEHSQHhRxPkKcYd1VrAWlxF9dNd/view?usp=sharing
thanks alot

I see two things in the log : time shifting, and maybe a bug in the main pool reconnection mechanism.

If you can turn off automatic time setup, and set it manually that would be good. When windows re-adjusts time BACK that can mess up a lot of things, miner thinks a day passed..

The second, main pool reconnection, i will test and fix if needed. It was implemented a long time ago, and a lot of stuff changed since, so maybe it's not working as it should.

thank you, let me try to turn off automatic time setup.

[dulama]

This is the error when I use Algorithm: Cryptonight V4 (R)

[2019-04-14 12:49:03] GPU0[BUS:1] result rejected for job[15] [low difficulty share]
[2019-04-14 12:49:33] Pool sent a new job, block height 416327 [ID: 5ac7b313-96fd-4601-94a0-9e51cecb4439]
[2019-04-14 12:49:44] Pool sent a new job, block height 416327 [ID: 5ac7b313-96fd-4601-94a0-9e51cecb4439]
[2019-04-14 12:49:44] GPU1[BUS:2] result rejected for job[17] [low difficulty share]

[MagicSmoker]

This is the error when I use Algorithm: Cryptonight V4 (R)

[2019-04-14 12:49:03] GPU0[BUS:1] result rejected for job[15] [low difficulty share]
[2019-04-14 12:49:33] Pool sent a new job, block height 416327 [ID: 5ac7b313-96fd-4601-94a0-9e51cecb4439]
[2019-04-14 12:49:44] Pool sent a new job, block height 416327 [ID: 5ac7b313-96fd-4601-94a0-9e51cecb4439]
[2019-04-14 12:49:44] GPU1[BUS:2] result rejected for job[17] [low difficulty share]

I'm getting something similar with 1.8.3 and 18.6.1 drivers on a mixed RX 560 / 570 rig. Will be going back to 1.8.1 here shortly if the "low difficulty share" errors persist.


One other bug I noticed is that the min_rig_speed parameter seems to only be looking at one of the GPU's hashrate, rather the sum total.

[doktor83]

This is the error when I use Algorithm: Cryptonight V4 (R)

[2019-04-14 12:49:03] GPU0[BUS:1] result rejected for job[15] [low difficulty share]
[2019-04-14 12:49:33] Pool sent a new job, block height 416327 [ID: 5ac7b313-96fd-4601-94a0-9e51cecb4439]
[2019-04-14 12:49:44] Pool sent a new job, block height 416327 [ID: 5ac7b313-96fd-4601-94a0-9e51cecb4439]
[2019-04-14 12:49:44] GPU1[BUS:2] result rejected for job[17] [low difficulty share]

I'm getting something similar with 1.8.3 and 18.6.1 drivers on a mixed RX 560 / 570 rig. Will be going back to 1.8.1 here shortly if the "low difficulty share" errors persist.


One other bug I noticed is that the min_rig_speed parameter seems to only be looking at one of the GPU's hashrate, rather the sum total.

Just clear the drivers with DDU from here and install 18.6.1
It must work. Also make sure you use normalv4 algo, because you can get 'low difficulty shares' if you use the wrong algorithm.

Also min_rig_speed checks the sum of all gpu threads, not 1 gpu.

edit:

tested for 27 minutes, works as expected. If it was looking at the hash of 1 gpu it would trigger a few times in these 27 minutes.

[Baikal miner]

Why i want this is error Error CL_BUILD_PROGRAM_FAILURE when calling clBuildProgram [ctx->Program] for DeviceID 1 (Thread 2)
Error while creating GPU threads.. Stopping miner process.
http://prntscr.com/nbutg3
http://prntscr.com/nbutn6

[doktor83]

Why i want this is error Error CL_BUILD_PROGRAM_FAILURE when calling clBuildProgram [ctx->Program] for DeviceID 1 (Thread 2)
Error while creating GPU threads.. Stopping miner process.

Because r9 380 is not ellesmere. Integrated gpu or driver is messing up there.

[ManDoneKTM]

Hello ! After i update it to 1 8 3 i have this issues . https://imgur.com/a/jwg9PeQ

[doktor83]

Hello ! After i update it to 1 8 3 i have this issues . https://imgur.com/a/jwg9PeQ

check the second tip on the first page here : https://bitcointalk.org/index.php?topic=3167363.msg32748623#msg32748623

[Sadopwnz]

I have 8 GPUs, but the 8th GPU crashes still mining after ~5-30 min with log:
https://c.radikal.ru/c42/1904/8b/c9e6fe0e2f72t.jpg
Switching GPUs, risers, stock settings did nothing. MB asus z370-p, 8xflashed vega56.
Any ideas?

[Baikal miner]

Why i want this is error Error CL_BUILD_PROGRAM_FAILURE when calling clBuildProgram [ctx->Program] for DeviceID 1 (Thread 2)
Error while creating GPU threads.. Stopping miner process.
http://prntscr.com/nbutg3
http://prntscr.com/nbutn6

Because r9 380 is not ellesmere. Integrated gpu or driver is messing up there.

i re-install driver not work

[Dubai-Industry]

nice relise