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

[abg00]

I added a full copy paste example on how to use algo switching on Monero Ocean.
Check it out on the first page

moneroocean realy support heavy algo? reject auth on heavy

[1638784471]

1638784471

[1638784471]

1638784471

[1638784471]

1638784471

[N2DCRYPT]

Hi all,

I was wondering why some of my vegas have lower hashrate of up to 300Hs compare with others when mining fast algo?

[2018-08-16 11:38:54] Pool sent a new job (ID: LiMmZRHeK8LOjapQjqJ67poFTTn0) [fast]
[2018-08-16 11:38:54] GPU0 :    3742 H/s [T: 55c, RPM: 2079, CC: 1407 MHz, MC: 950 MHz][BUS:7]
[2018-08-16 11:38:54] GPU1 :    3553 H/s [T: 52c, RPM: 2993, CC: 1407 MHz, MC: 950 MHz][BUS:16]
[2018-08-16 11:38:54] GPU2 :    3868 H/s [T: 58c, RPM: 4273, CC: 1407 MHz, MC: 950 MHz][BUS:10]
[2018-08-16 11:38:54] GPU3 :    4062 H/s [T: 57c, RPM: 4081, CC: 1408 MHz, MC: 1100 MHz][BUS:13]
[2018-08-16 11:38:54] GPU4 :    3891 H/s [T: 58c, RPM: 3357, CC: 1407 MHz, MC: 950 MHz][BUS:3]
[2018-08-16 11:38:54] Total:    19116 H/s

When mining the other algos there is no significant different

I can verify that this happens for me as well, usually always 2 on an 8 card machine, all cards identical.

[Zazax21]

I added a full copy paste example on how to use algo switching on Monero Ocean.
Check it out on the first page

moneroocean realy support heavy algo? reject auth on heavy

It does man, i've tryed it myself, but you have to make 3 different configs hevan, heavy, bittubev2.

[doktor83]

I will check this fast algo thing.

You guys have enough size for pagefile?

[Amgine]

Dok,

I have a problem when using algo switch

here's the crashlog.

[2018-08-16 21:04:46] Miner version: 1.6.6
[2018-08-16 21:04:46] Windows version: 10.0 build 17134
[2018-08-16 21:04:47] Video driver version: 24.20.11021.1000
[2018-08-16 21:04:52] AMD Platform ID: 0
[2018-08-16 21:04:52] AMD platform FOUND
[2018-08-16 21:04:52] Found 4 AMD devices
[2018-08-16 21:04:52] CPU AES-NI: TRUE
[2018-08-16 21:04:52] GPU0: Radeon RX Vega [gfx900] [8176 MB][Intensity 112.0][W: 16][T: 1][K: 1][BUS: 34]
[2018-08-16 21:04:52] GPU1: Radeon RX Vega [gfx900] [8176 MB][Intensity 112.0][W: 16][T: 1][K: 1][BUS: 34]
[2018-08-16 21:04:52] GPU2: Radeon RX Vega [gfx900] [8176 MB][Intensity 77.0][W: 16][T: 1][K: 1][BUS: 37]
[2018-08-16 21:04:52] GPU3: Radeon RX Vega [gfx900] [8176 MB][Intensity 77.0][W: 16][T: 1][K: 1][BUS: 37]
[2018-08-16 21:04:52] GPU4: Radeon RX Vega [gfx900] [8176 MB][Intensity 112.0][W: 16][T: 1][K: 1][BUS: 41]
[2018-08-16 21:04:52] GPU5: Radeon RX Vega [gfx900] [8176 MB][Intensity 112.0][W: 16][T: 1][K: 1][BUS: 41]
[2018-08-16 21:04:52] GPU6: Radeon RX Vega [gfx900] [8176 MB][Intensity 112.0][W: 16][T: 1][K: 1][BUS: 45]
[2018-08-16 21:04:52] GPU7: Radeon RX Vega [gfx900] [8176 MB][Intensity 112.0][W: 16][T: 1][K: 1][BUS: 45]
[2018-08-16 21:04:52] ADL is enabled
[2018-08-16 21:04:52] CryptonightV7 mode enabled
[2018-08-16 21:04:52] DevFee pool SET
[2018-08-16 21:04:52] DevFee address SET
[2018-08-16 21:04:52] Starting init of mining threads
[2018-08-16 21:04:53] Loading kernel for DEVICE BUS ID 34 ...
[2018-08-16 21:04:54] Stopping miner process
[2018-08-16 21:04:54] Loading kernel for DEVICE BUS ID 34 ...

If I run each algo seperately they run fine after compiling kernel. but if I use algoswitch, the miner closes itself.

[doktor83]

Dok,

I have a problem when using algo switch

here's the crashlog.

[2018-08-16 21:04:46] Miner version: 1.6.6
[2018-08-16 21:04:46] Windows version: 10.0 build 17134
[2018-08-16 21:04:47] Video driver version: 24.20.11021.1000
[2018-08-16 21:04:52] AMD Platform ID: 0
[2018-08-16 21:04:52] AMD platform FOUND
[2018-08-16 21:04:52] Found 4 AMD devices
[2018-08-16 21:04:52] CPU AES-NI: TRUE
[2018-08-16 21:04:52] GPU0: Radeon RX Vega [gfx900] [8176 MB][Intensity 112.0][W: 16][T: 1][K: 1][BUS: 34]
[2018-08-16 21:04:52] GPU1: Radeon RX Vega [gfx900] [8176 MB][Intensity 112.0][W: 16][T: 1][K: 1][BUS: 34]
[2018-08-16 21:04:52] GPU2: Radeon RX Vega [gfx900] [8176 MB][Intensity 77.0][W: 16][T: 1][K: 1][BUS: 37]
[2018-08-16 21:04:52] GPU3: Radeon RX Vega [gfx900] [8176 MB][Intensity 77.0][W: 16][T: 1][K: 1][BUS: 37]
[2018-08-16 21:04:52] GPU4: Radeon RX Vega [gfx900] [8176 MB][Intensity 112.0][W: 16][T: 1][K: 1][BUS: 41]
[2018-08-16 21:04:52] GPU5: Radeon RX Vega [gfx900] [8176 MB][Intensity 112.0][W: 16][T: 1][K: 1][BUS: 41]
[2018-08-16 21:04:52] GPU6: Radeon RX Vega [gfx900] [8176 MB][Intensity 112.0][W: 16][T: 1][K: 1][BUS: 45]
[2018-08-16 21:04:52] GPU7: Radeon RX Vega [gfx900] [8176 MB][Intensity 112.0][W: 16][T: 1][K: 1][BUS: 45]
[2018-08-16 21:04:52] ADL is enabled
[2018-08-16 21:04:52] CryptonightV7 mode enabled
[2018-08-16 21:04:52] DevFee pool SET
[2018-08-16 21:04:52] DevFee address SET
[2018-08-16 21:04:52] Starting init of mining threads
[2018-08-16 21:04:53] Loading kernel for DEVICE BUS ID 34 ...
[2018-08-16 21:04:54] Stopping miner process
[2018-08-16 21:04:54] Loading kernel for DEVICE BUS ID 34 ...

If I run each algo seperately they run fine after compiling kernel. but if I use algoswitch, the miner closes itself.

as all your intensities are same, drop the duplicateid stuff from cmd and put 2 threads in config. Algo switching doesnt like cmd parameters, i will try to fix that

[dragonmike]

I will check this fast algo thing.

You guys have enough size for pagefile?

Same issue here. Plenty of size for the pagefile too.

[Amgine]

as all your intensities are same, drop the duplicateid stuff from cmd and put 2 threads in config. Algo switching doesnt like cmd parameters, i will try to fix that

I removed --enableduplicateguid from config but miner still close.

here's the crashlog

Code:

[2018-08-16 21:39:25] Miner version: 1.6.6
[2018-08-16 21:39:25] Windows version: 10.0 build 17134
[2018-08-16 21:39:25] Video driver version: 24.20.11021.1000

only 3 lines now

here's the start.bat

Code:

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 --sendallstales --config config-normalv7.txt --pools pools-algos.txt --logfile crashlog.txt --usealgoswitching
pause

i remove duplicate line from the config too

Code:

{
"cryptonight_type" : "normalv7",
"intensity" : 0,
"double_threads" : true,
"min_rig_speed" : 7500,
"min_rig_speed_duration" : 60,
"gpu_conf" :
[
{ "id" : 0, "intensity" : 112, "worksize" : 16, "threads" : 2},
{ "id" : 1, "intensity" : 77, "worksize" : 16, "threads" : 2},
{ "id" : 2, "intensity" : 112, "worksize" : 16, "threads" : 2},
{ "id" : 3, "intensity" : 112, "worksize" : 16, "threads" : 2},
]

/* THERE ARE A LOT OF OPTIONS/PARAMETERS YOU CAN USE HERE */
/* YOU CAN FIND THEM IN THE README FILE */
}

If I remove the --usealgoswitching it will run

[doktor83]

I will check this fast algo thing.

You guys have enough size for pagefile?

Same issue here. Plenty of size for the pagefile too.

Is this happening only on vegas?

[bucah]

I will check this fast algo thing.

You guys have enough size for pagefile?

I set 60GB - 90GB on virtual memory

And I have only Vega and just 1 RX550, sorry

[doktor83]

as all your intensities are same, drop the duplicateid stuff from cmd and put 2 threads in config. Algo switching doesnt like cmd parameters, i will try to fix that

I removed --enableduplicateguid from config but miner still close.

here's the crashlog

Code:

[2018-08-16 21:39:25] Miner version: 1.6.6
[2018-08-16 21:39:25] Windows version: 10.0 build 17134
[2018-08-16 21:39:25] Video driver version: 24.20.11021.1000

only 3 lines now

here's the start.bat

Code:

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 --sendallstales --config config-normalv7.txt --pools pools-algos.txt --logfile crashlog.txt --usealgoswitching
pause

i remove duplicate line from the config too

Code:

{
"cryptonight_type" : "normalv7",
"intensity" : 0,
"double_threads" : true,
"min_rig_speed" : 7500,
"min_rig_speed_duration" : 60,
"gpu_conf" :
[
{ "id" : 0, "intensity" : 112, "worksize" : 16, "threads" : 2},
{ "id" : 1, "intensity" : 77, "worksize" : 16, "threads" : 2},
{ "id" : 2, "intensity" : 112, "worksize" : 16, "threads" : 2},
{ "id" : 3, "intensity" : 112, "worksize" : 16, "threads" : 2},
]

/* THERE ARE A LOT OF OPTIONS/PARAMETERS YOU CAN USE HERE */
/* YOU CAN FIND THEM IN THE README FILE */
}

If I remove the --usealgoswitching it will run

This is very interesting, cause everything looks ok. Good config, good driver (18.6.1), really strange i don't know what to think.

[wgestickley]

Hey,

i got my first RX Vega Cards and have Problems to set them up, Driver is installed as on Page 1. Als settings are done but the Hashrate is for CN_Heavy only 700Hash by expected 1550...

What is to do here?

Thank you

[zhmonzh]

Hi,
I have 4 vega frontier and install amd 18.Q3 drivers and 18.4.1 gaming drivers. With SRBminer1.6.5 for crypto heavy, i can set intensity only 58, but with blockchains drivers i can set intensity 112, someone has same problem? or solution? For cryptonightV7, i can set intensity 124.

[XiXiMios]

You have to be sure you use the correct algo set in config file!

[doktor83]

Hey,

i got my first RX Vega Cards and have Problems to set them up, Driver is installed as on Page 1. Als settings are done but the Hashrate is for CN_Heavy only 700Hash by expected 1550...

What is to do here?

Thank you

Use soft pp table, set compute mode and use 2 threads.

[wgestickley]

Hey,

i got my first RX Vega Cards and have Problems to set them up, Driver is installed as on Page 1. Als settings are done but the Hashrate is for CN_Heavy only 700Hash by expected 1550...

What is to do here?

Thank you

Use soft pp table, set compute mode and use 2 threads.

Is there a tutorial how to do this soft pp table?

Sorry, i am a newbie on Vegas cards

[ojcrisanto]

This miner needs a lot of pagefile memory but other than that, this miner is really stable in hashes.

[matt7659]

I will check this fast algo thing.

You guys have enough size for pagefile?

Hello! I have the same issue on all my rigs. How much pagefile is needed per gpu?
Thank you!

[hesido]

I will check this fast algo thing.

You guys have enough size for pagefile?

Hello! I have the same issue on all my rigs. How much pagefile is needed per gpu?
Thank you!

Depends on the intensity and algo, linearly related to how much mem GPU's are using. For 13 gpu's with 7MB mem usage each on cn/heavy, I needed to have at least 100MB vmem for proper operation. You can check out your GPU mem usage for 1 card, and then multiply that with the number of your cards, and add 20% extra just in case.

The vmem is almost never accessed. Just sits there, it could be on the slowest drive you have, has no impact on performance.

[livada]

I will check this fast algo thing.

You guys have enough size for pagefile?

Hello! I have the same issue on all my rigs. How much pagefile is needed per gpu?
Thank you!

for vega? or ?

1vega 8GB or more
i use 60GB for 6 vega 56

read this:
http://vega.miningguides.com/#WinSettings