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

[Topcd]

I've tested V1.9.1 on one of my Rigs with 6 RX VEGAs and it is stable after 1 day and a half of mining.
if it goes without any error for two or three days, then i will start to set it up on all of my Rigs.
proof: https://i.ibb.co/DW4XrYt/Monero-14-July.png

[1714958522]

1714958522

[1714958522]

1714958522

[1714958522]

1714958522

[netmebtc]

@doktor83, Hello, When will SRBminer support turtle's new algo "chukwa"   and loki's new algo "RandmXL"?

[doktor83]

@doktor83, Hello, When will SRBminer support turtle's new algo "chukwa"   and loki's new algo "RandmXL"?

I haven't decided yet what to do with the randomx algo family - it will suck on AMD gpus. Chukwa isn't in my plans atm.

[doktor83]

V1.9.2
- Added parameter '--disablerestartonrejectedshares' to disable auto restart of miner on too many rejected shares
- Added parameter '--maxrejectedshares' to set the number of allowed rejected shares on a pool connection ( min 1, max 100, def. 15)
- Optimisations in OpenCL code, so tiny hashrate increase could be seen on Vegas and Ellesmere cards.
- Minor bug fixes


+ Previous version added a feature that restarts miner when there are too many rejected shares. This version adds a new parameter to disable this feature, '--disablerestartonrejectedshares'

+ Also previous version had a hardcoded number of 5 for the max allowed rejected shares, this is now changed to 15, or you can set your own value with '--maxrejectedshares' value

+ Vegas and Ellesmere cards can see some small hashrate increase on major algorithms

[Topcd]

@doktor83, Hello, When will SRBminer support turtle's new algo "chukwa"   and loki's new algo "RandmXL"?

I haven't decided yet what to do with the randomx algo family - it will suck on AMD gpus. Chukwa isn't in my plans atm.

RandomX is the good news for CPU miners and bad news for GPU miners, specially for RX Vega Owners like me as i've heard that the hash rate will drop off around 50% for Rx Vega on this new algorithm, but they say they are doing some optimizations to make it more acceptable.
Let's wait and see what happens!

[GodOfStars]

V1.9.2
- Added parameter '--disablerestartonrejectedshares' to disable auto restart of miner on too many rejected shares
- Added parameter '--maxrejectedshares' to set the number of allowed rejected shares on a pool connection ( min 1, max 100, def. 15)
- Optimisations in OpenCL code, so tiny hashrate increase could be seen on Vegas and Ellesmere cards.
- Minor bug fixes


+ Previous version added a feature that restarts miner when there are too many rejected shares. This version adds a new parameter to disable this feature, '--disablerestartonrejectedshares'

+ Also previous version had a hardcoded number of 5 for the max allowed rejected shares, this is now changed to 15, or you can set your own value with '--maxrejectedshares' value

+ Vegas and Ellesmere cards can see some small hashrate increase on major algorithms

Abut the rejected shares parameter, where do we write this code, in the config file or in the start.bat?

[doktor83]

V1.9.2
- Added parameter '--disablerestartonrejectedshares' to disable auto restart of miner on too many rejected shares
- Added parameter '--maxrejectedshares' to set the number of allowed rejected shares on a pool connection ( min 1, max 100, def. 15)
- Optimisations in OpenCL code, so tiny hashrate increase could be seen on Vegas and Ellesmere cards.
- Minor bug fixes


+ Previous version added a feature that restarts miner when there are too many rejected shares. This version adds a new parameter to disable this feature, '--disablerestartonrejectedshares'

+ Also previous version had a hardcoded number of 5 for the max allowed rejected shares, this is now changed to 15, or you can set your own value with '--maxrejectedshares' value

+ Vegas and Ellesmere cards can see some small hashrate increase on major algorithms

Abut the rejected shares parameter, where do we write this code, in the config file or in the start.bat?

Every parameter that starts with -- is a cmd parameter and goes to start.bat:

Code:

SRBMiner-CN.exe --config Config\config-normalv4.txt --pools pools.txt --maxrejectedshares 20

[GodOfStars]

V1.9.2
- Added parameter '--disablerestartonrejectedshares' to disable auto restart of miner on too many rejected shares
- Added parameter '--maxrejectedshares' to set the number of allowed rejected shares on a pool connection ( min 1, max 100, def. 15)
- Optimisations in OpenCL code, so tiny hashrate increase could be seen on Vegas and Ellesmere cards.
- Minor bug fixes


+ Previous version added a feature that restarts miner when there are too many rejected shares. This version adds a new parameter to disable this feature, '--disablerestartonrejectedshares'

+ Also previous version had a hardcoded number of 5 for the max allowed rejected shares, this is now changed to 15, or you can set your own value with '--maxrejectedshares' value

+ Vegas and Ellesmere cards can see some small hashrate increase on major algorithms

Abut the rejected shares parameter, where do we write this code, in the config file or in the start.bat?

Every parameter that starts with -- is a cmd parameter and goes to start.bat:

Code:

SRBMiner-CN.exe --config Config\config-normalv4.txt --pools pools.txt --maxrejectedshares 20

Thank you Doc, for the update and you response. On another note perhaps an idea about the following update. Won't it be cool if let's say to automatically restart the miner every 2-3 days so at least the miner can insure a 24/7 undisturbed operation. For example if the GPUs are mining more than a week straight they start to become buggy at some point. Like, you won't able to close the miner whiteout a hard rig restart or power off, or one of the gpus will die and the miner wont be able to start it back again. So i was thinking if the miner automatically restarts every 2-3 days, of course optional input from the user, would simply amazing. Thaks ones again!

[sky2018miner]

Hello
I can not stop mining Rio
on Cryptonight Gpu
other algorithms work
I bring a log
[2019-07-16 23:26:22] Startup monitor attached
[2019-07-16 23:26:22] Miner version: 1.9.2
[2019-07-16 23:26:22] Windows version: 10.0 build 17763
[2019-07-16 23:26:22] Video driver version: 24.20.11021.1000
[2019-07-16 23:26:22] Please run miner as administrator to enable tweaking
[2019-07-16 23:26:32] AMD Platform ID: 0
[2019-07-16 23:26:32] AMD platform FOUND
[2019-07-16 23:26:32] Found 7 AMD devices
[2019-07-16 23:26:32] CPU AES-NI: FALSE
[2019-07-16 23:26:32] PREPARE DATA: FALSE
[2019-07-16 23:26:32] GPU0: AMD Radeon (TM) R7 370 Series [baffin] [2048 MB][CU: 14][I: 13.0][W: 8][T: 2][F: 0][BUS: 1]
[2019-07-16 23:26:32] GPU1: AMD Radeon (TM) R7 370 Series [baffin] [2048 MB][CU: 14][I: 13.6][W: 8][T: 2][F: 0][BUS: 3]
[2019-07-16 23:26:32] GPU2: AMD Radeon (TM) R7 370 Series [tonga] [2048 MB][CU: 28][I: 27.0][W: 8][T: 2][F: 4][BUS: 4]
[2019-07-16 23:26:32] GPU3: AMD Radeon (TM) R7 370 Series [baffin] [4096 MB][CU: 16][I: 15.5][W: 8][T: 2][F: 0][BUS: 8]
[2019-07-16 23:26:32] GPU4: AMD Radeon (TM) R7 370 Series [pitcairn] [2048 MB][CU: 16][I: 16.0][W: 8][T: 2][F: 0][BUS: 9]
[2019-07-16 23:26:32] GPU5: AMD Radeon (TM) R7 370 Series [pitcairn] [2048 MB][CU: 16][I: 16.0][W: 8][T: 2][F: 0][BUS: 10]
[2019-07-16 23:26:32] GPU6: AMD Radeon (TM) R7 370 Series [pitcairn] [2048 MB][CU: 16][I: 16.0][W: 8][T: 2][F: 0][BUS: 11]
[2019-07-16 23:26:32] ADL is enabled
[2019-07-16 23:26:32] Algorithm: Cryptonight Gpu
[2019-07-16 23:26:32] Algo switching: disabled
[2019-07-16 23:26:32] Coin forking: disabled
[2019-07-16 23:26:32] Gpu watchdog: enabled [normal mode]
[2019-07-16 23:26:32] Startup monitor: enabled
[2019-07-16 23:26:32] Tweaking: disabled
[2019-07-16 23:26:32] Socket initialiasation done
[2019-07-16 23:26:32] Starting init of gpu threads
[2019-07-16 23:26:32] Created OCL context
[2019-07-16 23:26:33] Created OCL input buffer for DeviceID 0 (Thread 0)
[2019-07-16 23:26:33] load_binary: Couldn't load default binary for DeviceID 0 (Thread 0)
[2019-07-16 23:26:33] ctx->headers[0] created for DeviceID 0 (Thread 0)
[2019-07-16 23:26:33] ctx->Program created for DeviceID 0 (Thread 0)
[2019-07-16 23:26:33] Compiling [gpu] kernel for DEVICE BUS_ID[1]
[2019-07-16 23:26:35] Binary kernel for DeviceID 0 (Thread 0) saved to disk
[2019-07-16 23:26:35] Created OCL input buffer for DeviceID 0 (Thread 1)
[2019-07-16 23:26:35] load_binary: Couldn't load default binary for DeviceID 0 (Thread 1)
[2019-07-16 23:26:35] Loading [gpu] kernel for DEVICE BUS_ID[1]
[2019-07-16 23:26:35] ctx->Program for DeviceID 0 (Thread 1) loaded
[2019-07-16 23:26:36] Created OCL input buffer for DeviceID 1 (Thread 2)
[2019-07-16 23:26:36] load_binary: Couldn't load default binary for DeviceID 1 (Thread 2)
[2019-07-16 23:26:36] Loading [gpu] kernel for DEVICE BUS_ID[3]
[2019-07-16 23:26:36] ctx->Program for DeviceID 1 (Thread 2) loaded
[2019-07-16 23:26:36] Created OCL input buffer for DeviceID 1 (Thread 3)
[2019-07-16 23:26:36] load_binary: Couldn't load default binary for DeviceID 1 (Thread 3)
[2019-07-16 23:26:36] Loading [gpu] kernel for DEVICE BUS_ID[3]
[2019-07-16 23:26:36] ctx->Program for DeviceID 1 (Thread 3) loaded
[2019-07-16 23:26:36] Created OCL input buffer for DeviceID 2 (Thread 4)
[2019-07-16 23:26:36] load_binary: Couldn't load default binary for DeviceID 2 (Thread 4)
[2019-07-16 23:26:36] ctx->headers[0] created for DeviceID 2 (Thread 4)
[2019-07-16 23:26:36] Error CL_COMPILE_PROGRAM_FAILURE when calling clCompileProgram [ctx->Program] for DeviceID 2 (Thread 4)
[2019-07-16 23:26:36] Error while creating GPU threads.. Stopping miner process.
[2019-07-16 23:26:36] Socket cleanup done
[2019-07-16 23:26:36] Shutting down miner...


please fix it

[Turkish88]

@doktor83, Hello, When will SRBminer support turtle's new algo "chukwa"   and loki's new algo "RandmXL"?

I haven't decided yet what to do with the randomx algo family - it will suck on AMD gpus. Chukwa isn't in my plans atm.

RandomX is the good news for CPU miners and bad news for GPU miners, specially for RX Vega Owners like me as i've heard that the hash rate will drop off around 50% for Rx Vega on this new algorithm, but they say they are doing some optimizations to make it more acceptable.
Let's wait and see what happens!

But its better then switching to asic friendly algo, because on this scenario RX Vega hashrates drops more.
Lets hope to the succes of randomX

[cas333]

One of the most stable versions. Thanks Dok!  

[just4u]

 It's been awhile since I've used SRB Miner.  Mining XMR  using 2x Strix Vega 56s with Hynix Memory (and latest drivers as I don't only mine with the system) Can't really go above tweak profile 2 without artifices, same goes if I raise mem clocks to high (930+)  Is there anything else I can try based on the information I am giving here?  I think it's fairly ok.. not sure what else I can do. Thanks for any advice.

GPU0 :    1780 H/s [T: 55c, RPM: 1998, CC: 1307 MHz, MC: 900 MHz][BUS:29]
GPU1 :    1796 H/s [T: 65c, RPM: 2446, CC: 1307 MHz, MC: 900 MHz][BUS:32]
Watts around 160 per card (GPUZ)

Batfile:

setx GPU_MAX_HEAP_SIZE 100   
setx GPU_MAX_USE_SYNC_OBJECTS 1   
setx GPU_SINGLE_ALLOC_PERCENT 100   
setx GPU_MAX_ALLOC_PERCENT 100   
setx GPU_MAX_SINGLE_ALLOC_PERCENT 100   
@echo off   
cd %~dp0   
   
cls   
SRBMiner-CN.exe --config Config\config-xmr.txt --pools Pools\pools-xmr.txt


Config file is fairly standard,

{   
"cryptonight_type" : "normalv4",   
"intensity" : 0,   
"double_threads" : true,
"persistent_memory" : false,
"gpu_conf" :
[
   {
     "id" : 0,
     "intensity" : 112,
     "worksize" : 16,
     "threads" : 2,
     "tweak_profile" : 2
   },
   {
     "id" : 1,
     "intensity" : 112,
     "worksize" : 16,
     "threads" : 2,
     "tweak_profile" : 2
   }
]
   
}

[Topcd]

It's been awhile since I've used SRB Miner.  Mining XMR  using 2x Strix Vega 56s with Hynix Memory (and latest drivers as I don't only mine with the system) Can't really go above tweak profile 2 without artifices, same goes if I raise mem clocks to high (930+)  Is there anything else I can try based on the information I am giving here?  I think it's fairly ok.. not sure what else I can do. Thanks for any advice.

GPU0 :    1780 H/s [T: 55c, RPM: 1998, CC: 1307 MHz, MC: 900 MHz][BUS:29]
GPU1 :    1796 H/s [T: 65c, RPM: 2446, CC: 1307 MHz, MC: 900 MHz][BUS:32]
Watts around 160 per card (GPUZ)

Batfile:

setx GPU_MAX_HEAP_SIZE 100   
setx GPU_MAX_USE_SYNC_OBJECTS 1   
setx GPU_SINGLE_ALLOC_PERCENT 100   
setx GPU_MAX_ALLOC_PERCENT 100   
setx GPU_MAX_SINGLE_ALLOC_PERCENT 100   
@echo off   
cd %~dp0   
   
cls   
SRBMiner-CN.exe --config Config\config-xmr.txt --pools Pools\pools-xmr.txt


Config file is fairly standard,

{   
"cryptonight_type" : "normalv4",   
"intensity" : 0,   
"double_threads" : true,
"persistent_memory" : false,
"gpu_conf" :
[
   {
     "id" : 0,
     "intensity" : 112,
     "worksize" : 16,
     "threads" : 2,
     "tweak_profile" : 2
   },
   {
     "id" : 1,
     "intensity" : 112,
     "worksize" : 16,
     "threads" : 2,
     "tweak_profile" : 2
   }
]
   
}

I recommend you to flash Vega 64 bios on your Vega 56s, so you will be able to get at least 1900 h/s per card in normal condition.

[livada]

It's been awhile since I've used SRB Miner.  Mining XMR  using 2x Strix Vega 56s with Hynix Memory (and latest drivers as I don't only mine with the system) Can't really go above tweak profile 2 without artifices, same goes if I raise mem clocks to high (930+)  Is there anything else I can try based on the information I am giving here?  I think it's fairly ok.. not sure what else I can do. Thanks for any advice.

GPU0 :    1780 H/s [T: 55c, RPM: 1998, CC: 1307 MHz, MC: 900 MHz][BUS:29]
GPU1 :    1796 H/s [T: 65c, RPM: 2446, CC: 1307 MHz, MC: 900 MHz][BUS:32]
Watts around 160 per card (GPUZ)

Batfile:

setx GPU_MAX_HEAP_SIZE 100   
setx GPU_MAX_USE_SYNC_OBJECTS 1   
setx GPU_SINGLE_ALLOC_PERCENT 100   
setx GPU_MAX_ALLOC_PERCENT 100   
setx GPU_MAX_SINGLE_ALLOC_PERCENT 100   
@echo off   
cd %~dp0   
   
cls   
SRBMiner-CN.exe --config Config\config-xmr.txt --pools Pools\pools-xmr.txt


Config file is fairly standard,

{   
"cryptonight_type" : "normalv4",   
"intensity" : 0,   
"double_threads" : true,
"persistent_memory" : false,
"gpu_conf" :
[
   {
     "id" : 0,
     "intensity" : 112,
     "worksize" : 16,
     "threads" : 2,
     "tweak_profile" : 2
   },
   {
     "id" : 1,
     "intensity" : 112,
     "worksize" : 16,
     "threads" : 2,
     "tweak_profile" : 2
   }
]
   
}

I recommend you to flash Vega 64 bios on your Vega 56s, so you will be able to get at least 1900 h/s per card in normal condition.

hynix not good with 64 bios. Just use
 TRM and you have 2100 with hynix or 2350Hr with samsung vega 56

[just4u]

It's been awhile since I've used SRB Miner.  Mining XMR  using 2x Strix Vega 56s with Hynix Memory (and latest drivers as I don't only mine with the system) Can't really go above tweak profile 2 without artifices, same goes if I raise mem clocks to high (930+)  Is there anything else I can try based on the information I am giving here?  I think it's fairly ok.. not sure what else I can do. Thanks for any advice.

GPU0 :    1780 H/s [T: 55c, RPM: 1998, CC: 1307 MHz, MC: 900 MHz][BUS:29]
GPU1 :    1796 H/s [T: 65c, RPM: 2446, CC: 1307 MHz, MC: 900 MHz][BUS:32]
Watts around 160 per card (GPUZ)

I recommend you to flash Vega 64 bios on your Vega 56s, so you will be able to get at least 1900 h/s per card in normal condition.

hynix not good with 64 bios. Just use
 TRM and you have 2100 with hynix or 2350Hr with samsung vega 56

I never really thought about flashing these cards (had them over a year now..) reports were it was "iffy" on Asus Strix cards that don't use Samsung mem.   I tend to switch between Xmrig, Stak, SRB, if mining xmr.. plus dok has low fees and works on his a lot.

[doktor83]

I never really thought about flashing these cards (had them over a year now..) reports were it was "iffy" on Asus Strix cards that don't use Samsung mem.   I tend to switch between Xmrig, Stak, SRB, if mining xmr.. plus dok has low fees and works on his a lot.

Hi, with SRB your best drivers would be 18.6.1 with CN/R, because on every new block from the pool, the kernel is recompiled and my tests show that 18.6.1 compiles the best code still.
I know there are users here who have Hynix mem Vegas and were able to get nice results, maybe they can help you out. I have only 1 Vega56 with Hynix mem but it is a really good piece, it goes on profile 10 without problems and 2100+ hash.

112/16/2 are the best settings you can get for CN/R. Maybe try out the low profiles (add L after the profile number)
If using config file, you must surround the value with quotation marks if you want to use the L profile. Example : "5L"

Thanks for the kind words, and for using SRB.

[doktor83]

Anyone knows are there some new projects with some custom CN variant algo ? 

[Elder III]

Considering that the Radeon VII has all of that juicy 16 GB of HBM memory, is it feasible for it to run 4 threads instead of 2?

[abg00]

Anyone knows are there some new projects with some custom CN variant algo ? 

randomX wownero

[livada]

Anyone knows are there some new projects with some custom CN variant algo ?  

randomX wownero

random X  algo is GPU kiler algo. v8/v4 kil ASIC now  random X kill ASIC and GPU. CPU rules

vega 64 have 1200HR
i3 have 550hr
i7 have 1700
Ryzen7 4100
i9 5800

XMR ism CPU mining algo now