SRBMiner Cryptonight AMD GPU Miner V1.9.3 - native algo switching - page 338.

bobben2

full member

Activity: 279

Merit: 104

Some early test result with version 1.3.2
My 2 x RX 574 test rig now hashing at 913+914 H/s. And both GPUs temp/RPM are showing.
So far all is well. Thank you, dev Cheesy

Config:

GPUs running at 1200/2000 MSI with Hynix mem.

"gpu_conf" :
[
{ "id" : 0, "intensity" : 56, "worksize" : 8, "threads" : 2},
{ "id" : 1, "intensity" : 56, "worksize" : 8, "threads" : 2}
]

doktor83

hero member

Activity: 2548

Merit: 626

Quote from: Vers on April 11, 2018, 10:10:42 AM

I'm tested 1.3.2 update. For 8Gb Polaris hashrate remained at the same level - 1100 h/s. 4GB Polaris lost performance -725 h/s against 762 h/s in 1.3.1. 4Gb Hawaii productive same level - 670-690 h/s.

Nothing is lost, i just lowered auto setup values a little bit, you can set intensity by hand and you will be back again at the same speeds.

Vers

newbie

Activity: 10

Merit: 0

I'm tested 1.3.2 update. For 8Gb Polaris hashrate remained at the same level - 1100 h/s. 4GB Polaris lost performance -725 h/s against 762 h/s in 1.3.1. 4Gb Hawaii productive same level - 670-690 h/s.

doktor83

hero member

Activity: 2548

Merit: 626

V1.3.2
- Cryptonight UltraHeavy algo updated
- Miner restarting on crash should now work in most cases
- Added detailed device listing with --listdevices parameter
- Added detailed logging, use --logfile filename, log option removed from config.txt
- Hopefully fixed bug when in 'gpu_conf' you want to set only one GPU
- Fixed temperature and rpm display

So you guys still having problems can now enable detailed logging, you have to use parameter --logfile filename, in bat.
I removed log option from config file.

Also please report if you can use 1 gpu now in 'gpu_conf', i think i fixed that one.

CryptoPlay

jr. member

Activity: 202

Merit: 2

Is this miner able to control Vega fan rpm?

heavyarms1912

full member

Activity: 729

Merit: 114

Quote from: inv1s1bl3 on April 11, 2018, 08:39:00 AM

Okay, so in comparison with xmrig's speed 3,5kh, with this SBRMiner my RX580s are hashing 820h/s on cryptonight heavy!!! Incredible performance! I am currently with intensity 32, because if I get the intensity higher, I have noticed that after some time some cards start hashing around 550h/s(dropping by 300h) .
Running stable with 4150h/s on my 5 cards @ 550W from the wall. Great work man, hands down!

4gb or 8gb? I read 8gb go over 1k on CN-Heavy.

inv1s1bl3

newbie

Activity: 65

Merit: 0

Okay, so in comparison with xmrig's speed 3,5kh, with this SBRMiner my RX580s are hashing 820h/s on cryptonight heavy!!! Incredible performance! I am currently with intensity 32, because if I get the intensity higher, I have noticed that after some time some cards start hashing around 550h/s(dropping by 300h) .
Running stable with 4150h/s on my 5 cards @ 550W from the wall. Great work man, hands down!

Nassim

newbie

Activity: 5

Merit: 0

Quote from: theodrim on April 10, 2018, 07:37:31 AM

Quote from: Nassim on April 10, 2018, 07:15:05 AM

When setting the GPUs manually i get the fellowing error :

Error CL_INVALID_DEVICE when calling clCreateContext

I think that the integrated intel gpu hip that is causing this, something like "platform_index : 0" from xmr-stack can help.

Check which gpu is which by

Code:

SRBMiner-CN.exe --listdevices

GPU ID is correct.
As i mentioned the INTEL GPU chip is conflicting with miner initialising other GPUs

SRBMiner needs an option to deal with this like in xmr-stack with : "platform_index : 0"

Bullus77

newbie

Activity: 14

Merit: 0

Is possible in the future to change the intensity /worksize on the fly? Same as for example sgminer

doktor83

hero member

Activity: 2548

Merit: 626

Quote from: ALEX_RAA on April 11, 2018, 07:18:31 AM

is there known bug that every 2 gpus showing the same temperature and fan speed?

you may be right, i will check it

ALEX_RAA

newbie

Activity: 78

Merit: 0

is there known bug that every 2 gpus showing the same temperature and fan speed?

Kaezar

newbie

Activity: 73

Merit: 0

Quote from: Nss75 on April 10, 2018, 08:38:12 PM

Hello everyone!
Thanks to the dev for the great job.

I have some question about Haven mining (cn heavy).

I run a 6 gpu (Vega 64 8mb) rig with srbminer to mine haven.
It seems that my hashrate is not optimal comparing to others reporting 1500+ h/s using Vega 64 or 54 with xmr-stak.

I only get around 950h/s per gpu (max 1400 but drowning quickly)...
I also get some miner crash quite often.
I wonder if i do something wrong with the settings or if it is a miner/gpu limits...

The best hashrate i get comes with this settings.
-algo: heavy
-pool use ssl/tls: false
-target temperature: 0
-"intensity 55"
-i use cryptoknight pool (luckypool gave me the same result).

gpu config
-Intensity 150 for each gpu, double thread "true"

-Virtual memory on my computer 30000-60000
-On afterburner: coreclock 950, memory 1020

I get around 900-950 h/s per card with thoses settings (and was the best i could hit), wich seems to be low according to others

Does anyone have any idea of what i did wrong, or if it is srbminer limits?

Thanks a lot!

I use SRBMiner... in v56-64 plays with worksize 5-12.
mem 1105/1200.
try diff worksize with diff cards. ex..

Code:

{ "id" : 0, "intensity" : 61, "worksize" : 8, "threads" : 2},
{ "id" : 1, "intensity" : 61, "worksize" : 7, "threads" : 2},
{ "id" : 2, "intensity" : 61, "worksize" : 10, "threads" : 2},

Code:

[2018-04-11 14:39:24] GPU0: 1551.0 H/S [T: 56c, RPM: 4192]
[2018-04-11 14:39:24] GPU1: 1541.0 H/S [T: 56c, RPM: 4192]
[2018-04-11 14:39:24] GPU2: 1512.0 H/S [T: 55c, RPM: 4560]
[2018-04-11 14:39:24] Total: 4604.0 H/S

doktor83

hero member

Activity: 2548

Merit: 626

Quote from: inv1s1bl3 on April 11, 2018, 05:52:32 AM

Tried mining Sumocoin, but I receive the following error :
Error CL_MEM_OBJECT_ALLOCATION_FAILURE when calling clEnqueueNDRangeKernel for kernel 0 for DeviceID 4 (Thread Cool

a LOT!
I first started the miner with intensity 0 and saw at the beginning that it was automatically set @ 51, so I lowered it to 40, but I still receive a lot of those errors. Now I am currently trying it out with i30 and the miners seems to be kinda stable, will report after a couple of hours.

https://bitcointalksearch.org/topic/m.32748623

Read third tip

inv1s1bl3

newbie

Activity: 65

Merit: 0

Tried mining Sumocoin, but I receive the following error :
Error CL_MEM_OBJECT_ALLOCATION_FAILURE when calling clEnqueueNDRangeKernel for kernel 0 for DeviceID 4 (Thread Cool

a LOT!
I first started the miner with intensity 0 and saw at the beginning that it was automatically set @ 51, so I lowered it to 40, but I still receive a lot of those errors. Now I am currently trying it out with i30 and the miners seems to be kinda stable, will report after a couple of hours.

zeii

newbie

Activity: 41

Merit: 0

Quote from: ?? on ??

Hi

when i try to just one gpu:

"gpu_conf" :
[

{ "id" : 3, "intensity" : 120, "worksize" : 16, "threads" : 2},
]
}

i get an error this error:

Error CL_INVALID_DEVICE when calling clCreateContext

however if i run all 4..it works. can you please assist me on getting one gpu only to mine?

thanks

same here. If i disable 1 gpu it eror like this.
if run all 4 it work fine.

Pennywis3

full member

Activity: 327

Merit: 100

Quote from: gugs on April 11, 2018, 01:44:51 AM

What a wonderful miner, awesome job done doktore

6x RX470/4GB Elpida @ 1200/2000 - 976h/s per card @ Cryptonightv7, Msi Afterburner @ -200 undervolt, power draw total from wall 685 watts

I don't know if it is already proposed as feature but adding failover pools option will make it the best miner.

Keep up the good work!

Try using another tool for setting core mV, like OverdriveNTool or smth. and set those cards below 900mV, some even run at 820mV@1200Mhz, depends on the Asic %.

You can get a rig like that running below 600W@the wall.

Just a suggestion, no hard feelings Wink

GelatinousCube

newbie

Activity: 7

Merit: 0

Quote from: Nss75 on April 10, 2018, 08:38:12 PM

Hello everyone!
Thanks to the dev for the great job.

I have some question about Haven mining (cn heavy).

I run a 6 gpu (Vega 64 8mb) rig with srbminer to mine haven.
It seems that my hashrate is not optimal comparing to others reporting 1500+ h/s using Vega 64 or 54 with xmr-stak.

I only get around 950h/s per gpu (max 1400 but drowning quickly)...
I also get some miner crash quite often.
I wonder if i do something wrong with the settings or if it is a miner/gpu limits...

The best hashrate i get comes with this settings.
-algo: heavy
-pool use ssl/tls: false
-target temperature: 0
-"intensity 55"
-i use cryptoknight pool (luckypool gave me the same result).

gpu config
-Intensity 150 for each gpu, double thread "true"

-Virtual memory on my computer 30000-60000
-On afterburner: coreclock 950, memory 1020

I get around 900-950 h/s per card with thoses settings (and was the best i could hit), wich seems to be low according to others

Does anyone have any idea of what i did wrong, or if it is srbminer limits?

Thanks a lot!

hello to give you a idea with my single vega 64 p1100/950 I am mining cn-heavy for sumo after a crash I set to 0 and false which set a 56 intensity setting and an getting 1500h/s with a h button of 1460h/s only been on this miner for 2 days so I have a way to go to dial in but this looks to me like the go to miner for cn-heavy
on reading did you toggle the vega off and on in device manger
upon further digging I have a vega 64 in a rig and that one is getting 988h/s so I might be in the same boat will have to investigate further

tenmoi

newbie

Activity: 59

Merit: 0

Quote from: gugs on April 11, 2018, 01:44:51 AM

What a wonderful miner, awesome job done doktore

6x RX470/4GB Elpida @ 1200/2000 - 976h/s per card @ Cryptonightv7, Msi Afterburner @ -200 undervolt, power draw total from wall 685 watts

I don't know if it is already proposed as feature but adding failover pools option will make it the best miner.

Keep up the good work!

Could you please share your config.txt file?

gugs

newbie

Activity: 20

Merit: 0

What a wonderful miner, awesome job done doktore

6x RX470/4GB Elpida @ 1200/2000 - 976h/s per card @ Cryptonightv7, Msi Afterburner @ -200 undervolt, power draw total from wall 685 watts

I don't know if it is already proposed as feature but adding failover pools option will make it the best miner.

Keep up the good work!

Turkish88

full member

Activity: 714

Merit: 104

Quote from: vanthuspkt on April 11, 2018, 01:00:46 AM

Quote from: doktor83 on April 10, 2018, 11:49:42 PM

Quote from: vanthuspkt on April 10, 2018, 11:25:58 PM

I use SRBMiner Cryptonight AMD GPU Miner V1.3.1 for RX380
but It error CPU does not support ASE-NI.
I mining HAVEN COIN.

can You help me?

Its not an error, its just a notice.

But after 15s this Tool close. It can not start run.

There is not enough information we have to hep you
can you post with what error program close

Topic: SRBMiner Cryptonight AMD GPU Miner V1.9.3 - native algo switching - page 338. (Read 237263 times)