Pages:
Author

Topic: SRBMiner-MULTI GPU & CPU Miner 0.9.4 - page 28. (Read 44750 times)

legendary
Activity: 2576
Merit: 1073
September 12, 2021, 08:19:19 AM
Is SRBMiner-MULTI latest version supposed to be working on good old R9 280x cards? I see they are in the supported devices list, however any attempt to mine 'autolycos2' crashes the system badly, sometimes to the point where only hardware reset helps.

Tried to run it on HiveOS, both on a single card, and a bunch of them - no way, the miner works for some time, with extremely low hashrate on some cards, and with zero hash on others, then system freezes. The 'lolminer' works however (more or less stably), so its not rig issue... I just read somewhere SRBMiner should be better for those older cards, especially given the fact they are in supported device list...


Have you tried to manually tune the --gpu-intensity parameter? Or anything else beside just starting the miner with the default settings?

Yep, sure, I did. Tried to run it with "--gpu-intensity 0", "--gpu-intensity 10", "--gpu-intensity 5", "--gpu-auto-tune 0", "--gpu-auto-tune 1"... No luck. First of all, it doesn't seem to pick up the intensity value I give it, while also doesn't completely ignore the parameter: if I don't set --gpu-intensity, the log shows intensity value of 32 takes effect; while when I set "--gpu-intensity 0", instead of 0 some other (higher than zero) parameter is used; when I set it to 5, I can see it is actually set as 17. Weird.
As to "--gpu-auto-tune", it looks like when normal tuning procedure is started, it never ends - it works for some minutes, until GPU freezes.

Though I did make some experiments, but didn't do too much of them (probably up to 10-15 attempts with different parameter combinations), because the rigs are remote, almost every such experiment brings it to hangup state, while my remote hardware reset procedure is somewhat complicated, so I just did experiments until my patience was exhausted... Then I came here to ask Smiley
hero member
Activity: 2548
Merit: 626
September 11, 2021, 11:20:41 PM
Is SRBMiner-MULTI latest version supposed to be working on good old R9 280x cards? I see they are in the supported devices list, however any attempt to mine 'autolycos2' crashes the system badly, sometimes to the point where only hardware reset helps.

Tried to run it on HiveOS, both on a single card, and a bunch of them - no way, the miner works for some time, with extremely low hashrate on some cards, and with zero hash on others, then system freezes. The 'lolminer' works however (more or less stably), so its not rig issue... I just read somewhere SRBMiner should be better for those older cards, especially given the fact they are in supported device list...


Have you tried to manually tune the --gpu-intensity parameter? Or anything else beside just starting the miner with the default settings?
legendary
Activity: 2576
Merit: 1073
September 11, 2021, 04:15:14 PM
Is SRBMiner-MULTI latest version supposed to be working on good old R9 280x cards? I see they are in the supported devices list, however any attempt to mine 'autolycos2' crashes the system badly, sometimes to the point where only hardware reset helps.

Tried to run it on HiveOS, both on a single card, and a bunch of them - no way, the miner works for some time, with extremely low hashrate on some cards, and with zero hash on others, then system freezes. The 'lolminer' works however (more or less stably), so its not rig issue... I just read somewhere SRBMiner should be better for those older cards, especially given the fact they are in supported device list...
member
Activity: 325
Merit: 42
September 10, 2021, 05:02:39 AM
Xmrig-cpu was just xmrig renamed, lol

I'm using Pop!_OS 21.04.

I'm not running with sudo.

Take it you have done the following tweaks allready.

GRUB_CMDLINE_LINUX="msr.allow_writes=on" in /etc/default/grub and update-grub.
As normally msr writes are not allowed. Also need to have msr-tools installed as well.

To boost xmrig get the following script from the xmrig/scripts directory on github:
randomx_boost.sh and and maybe enable_1gb_pages.sh run them as root or w/sudo.
In the config.json file change "rdmsr": true, and "wrmsr": true, both to false

and then you can run xmrig as user otherwise you won't benefit as most tweaks you need to run xmrig under root w/sudo.

Same actually counts for SRBMiner-MULTI and not only those miners but practically all miners.
hero member
Activity: 677
Merit: 500
September 09, 2021, 04:38:50 PM
Xmrig-cpu was just xmrig renamed, lol

I'm using Pop!_OS 21.04.

I'm not running with sudo.
member
Activity: 325
Merit: 42
September 08, 2021, 06:07:18 PM
Does this fix the miner from completely exiting out of a bash script?  I'm still having this issue.  Been using a different miner until this is fixed.

Can you try it in a script yourself to verify the behavior and debug to find out why this is happening though?  I am not a developer so I have no idea how to find out why this happens.

Im not sure i can fix anything for you. Miner on SIGINT or SIGBREAK does a cleanup, shutdown and exits to shell, it's up to you what you do in your script after that.
Mining distros like Hiveos, Mmpos, etc have no issues with the shutdown procedure...

And you will prob. now say but other miners work normally [this is what people usually say].

I have no idea why it would be different for different miners (or any application). The application has no control over the script
that called it, all it can do is exit and let bash take over from there.

Are there any differences in the script or the bash environment. Are there any other differences in behaviour like console messages?

I don't know why it behaves this way either.  It doesn't happen with other miners such as t-rex, ccminer, or even your own cpuminer-opt.

I don't know much about Linux as I'm just starting to learn it.  I'm carrying a lot of command line experience from my DOS days.

Do you run the script as a user or with sudo?
In your script put set -x under $!/bin/bash start it from the terminal it shows you what happens with the script. Also instead of using xmrig-cpu (which is not actively maintained anymore) use xmrig https://github.com/xmrig/xmrig as that is the latest official one.

Forgot to ask which distro?
hero member
Activity: 677
Merit: 500
September 08, 2021, 12:13:38 AM
Does this fix the miner from completely exiting out of a bash script?  I'm still having this issue.  Been using a different miner until this is fixed.

Can you try it in a script yourself to verify the behavior and debug to find out why this is happening though?  I am not a developer so I have no idea how to find out why this happens.

Im not sure i can fix anything for you. Miner on SIGINT or SIGBREAK does a cleanup, shutdown and exits to shell, it's up to you what you do in your script after that.
Mining distros like Hiveos, Mmpos, etc have no issues with the shutdown procedure...

And you will prob. now say but other miners work normally [this is what people usually say].

I have no idea why it would be different for different miners (or any application). The application has no control over the script
that called it, all it can do is exit and let bash take over from there.

Are there any differences in the script or the bash environment. Are there any other differences in behaviour like console messages?

I don't know why it behaves this way either.  It doesn't happen with other miners such as t-rex, ccminer, or even your own cpuminer-opt.

I don't know much about Linux as I'm just starting to learn it.  I'm carrying a lot of command line experience from my DOS days.
full member
Activity: 1424
Merit: 225
September 05, 2021, 11:22:23 AM
Does this fix the miner from completely exiting out of a bash script?  I'm still having this issue.  Been using a different miner until this is fixed.

Can you try it in a script yourself to verify the behavior and debug to find out why this is happening though?  I am not a developer so I have no idea how to find out why this happens.

Im not sure i can fix anything for you. Miner on SIGINT or SIGBREAK does a cleanup, shutdown and exits to shell, it's up to you what you do in your script after that.
Mining distros like Hiveos, Mmpos, etc have no issues with the shutdown procedure...

And you will prob. now say but other miners work normally [this is what people usually say].

I have no idea why it would be different for different miners (or any application). The application has no control over the script
that called it, all it can do is exit and let bash take over from there.

Are there any differences in the script or the bash environment. Are there any other differences in behaviour like console messages?
hero member
Activity: 677
Merit: 500
September 05, 2021, 09:34:51 AM
Does this fix the miner from completely exiting out of a bash script?  I'm still having this issue.  Been using a different miner until this is fixed.

Can you try it in a script yourself to verify the behavior and debug to find out why this is happening though?  I am not a developer so I have no idea how to find out why this happens.

Im not sure i can fix anything for you. Miner on SIGINT or SIGBREAK does a cleanup, shutdown and exits to shell, it's up to you what you do in your script after that.
Mining distros like Hiveos, Mmpos, etc have no issues with the shutdown procedure...

And you will prob. now say but other miners work normally [this is what people usually say].
hero member
Activity: 2548
Merit: 626
September 03, 2021, 12:32:52 AM
Does this fix the miner from completely exiting out of a bash script?  I'm still having this issue.  Been using a different miner until this is fixed.

Im not sure i can fix anything for you. Miner on SIGINT or SIGBREAK does a cleanup, shutdown and exits to shell, it's up to you what you do in your script after that.
Mining distros like Hiveos, Mmpos, etc have no issues with the shutdown procedure...

And you will prob. now say but other miners work normally [this is what people usually say].
hero member
Activity: 677
Merit: 500
September 02, 2021, 02:52:22 PM
V0.8.0
+ Added algorithm 'randomgrft' (GRFT - Graft coin) for CPU mining, fee 0.85%
+ Performance increase on 'ethash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'etchash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'ubqhash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'verthash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'heavyhash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'autolykos2' algorithm on ZEN2+ AMD CPU's (up to ~90%) (*)
+ Fixed 'cryptonight_gpu' algorithm for RDNA/RDNA2 GPU's
+ Fixed 'cryptonight_xhv' algorithm for RDNA/RDNA2 GPU's
+ Added GPU power consumption details to stats & api
+ Decreased time needed for algorithm switch when using --multi-algorithm-job-mode 3 with cpu mining
+ Minor bug fixes

* - might apply to other AVX2 CPU's too - not tested

Does this fix the miner from completely exiting out of a bash script?  I'm still having this issue.  Been using a different miner until this is fixed.
hero member
Activity: 2548
Merit: 626
August 25, 2021, 11:07:58 PM
V0.8.0
+ Added algorithm 'randomgrft' (GRFT - Graft coin) for CPU mining, fee 0.85%
+ Performance increase on 'ethash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'etchash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'ubqhash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'verthash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'heavyhash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'autolykos2' algorithm on ZEN2+ AMD CPU's (up to ~90%) (*)
+ Fixed 'cryptonight_gpu' algorithm for RDNA/RDNA2 GPU's
+ Fixed 'cryptonight_xhv' algorithm for RDNA/RDNA2 GPU's
+ Added GPU power consumption details to stats & api
+ Decreased time needed for algorithm switch when using --multi-algorithm-job-mode 3 with cpu mining
+ Minor bug fixes

* - might apply to other AVX2 CPU's too - not tested

for navi cards -- ethash  080 bad
                                    078 best

Specifically which one? RX5700 ?
hero member
Activity: 2548
Merit: 626
August 25, 2021, 05:50:45 AM
V0.8.0
+ Added algorithm 'randomgrft' (GRFT - Graft coin) for CPU mining, fee 0.85%
+ Performance increase on 'ethash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'etchash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'ubqhash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'verthash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'heavyhash' algorithm for RX6800/6900 GPU's
+ Performance increase on 'autolykos2' algorithm on ZEN2+ AMD CPU's (up to ~90%) (*)
+ Fixed 'cryptonight_gpu' algorithm for RDNA/RDNA2 GPU's
+ Fixed 'cryptonight_xhv' algorithm for RDNA/RDNA2 GPU's
+ Added GPU power consumption details to stats & api
+ Decreased time needed for algorithm switch when using --multi-algorithm-job-mode 3 with cpu mining
+ Minor bug fixes

* - might apply to other AVX2 CPU's too - not tested
hero member
Activity: 2548
Merit: 626
August 22, 2021, 07:51:30 AM
Do you wish to add support for 4000G or 5000G iGPU in the Miner?

list-gpu-devices says and lists the device but the Actual miner application just ignored it when trying to use it.

Said no Suitable GPU devices found?

Not something you wish to add? I was just curious what kind of performance one could get but could not start to even test.

I don't have a ryzen with an integrated gpu so i can't do much here.

Do you need a device ID or something to add it in the code?

I presume you have a list of approved GPU's or something for the miner to allow?


Yeah, but you assume that its enough to allow a codename, and it will magically work Smiley

You need sku specific optimization/code and such to be done? Bummer.
It's supposed to be "VEGA", so was guessing it shouldn't have been too much work to do a test. But it's work, nonetheless.

Will not bother to ask you more about it.
I don't expect them to be any performance killers either.

i could allow those codenames and set them to use gfx900 kernels for example, but i can't test it if it works. might work, who knows.
And if it works, it would probably still be better if it was optimised specifically for that GPU, and not just use the gfx900 ones.
jr. member
Activity: 133
Merit: 5
August 22, 2021, 05:18:19 AM
Do you wish to add support for 4000G or 5000G iGPU in the Miner?

list-gpu-devices says and lists the device but the Actual miner application just ignored it when trying to use it.

Said no Suitable GPU devices found?

Not something you wish to add? I was just curious what kind of performance one could get but could not start to even test.

I don't have a ryzen with an integrated gpu so i can't do much here.

Do you need a device ID or something to add it in the code?

I presume you have a list of approved GPU's or something for the miner to allow?


Yeah, but you assume that its enough to allow a codename, and it will magically work Smiley

You need sku specific optimization/code and such to be done? Bummer.
It's supposed to be "VEGA", so was guessing it shouldn't have been too much work to do a test. But it's work, nonetheless.

Will not bother to ask you more about it.
I don't expect them to be any performance killers either.
hero member
Activity: 2548
Merit: 626
August 22, 2021, 05:05:16 AM
Do you wish to add support for 4000G or 5000G iGPU in the Miner?

list-gpu-devices says and lists the device but the Actual miner application just ignored it when trying to use it.

Said no Suitable GPU devices found?

Not something you wish to add? I was just curious what kind of performance one could get but could not start to even test.

I don't have a ryzen with an integrated gpu so i can't do much here.

Do you need a device ID or something to add it in the code?

I presume you have a list of approved GPU's or something for the miner to allow?


Yeah, but you assume that its enough to allow a codename, and it will magically work Smiley
jr. member
Activity: 133
Merit: 5
August 22, 2021, 05:00:57 AM
Do you wish to add support for 4000G or 5000G iGPU in the Miner?

list-gpu-devices says and lists the device but the Actual miner application just ignored it when trying to use it.

Said no Suitable GPU devices found?

Not something you wish to add? I was just curious what kind of performance one could get but could not start to even test.

I don't have a ryzen with an integrated gpu so i can't do much here.

Do you need a device ID or something to add it in the code?

I presume you have a list of approved GPU's or something for the miner to allow?
hero member
Activity: 677
Merit: 500
August 19, 2021, 07:08:17 AM
I already do a sleep 1.  Do I need 2 seconds?

Did you get it to run in a terminal with the suggestions?

Try different sleep settings to see if it works. It also depends what cpu/mem you have in other words the faster your system is the longer sleep you might need.


I tried it with longer sleep (up to 10 seconds) same issue...just exits out of the script as soon as it self terminates.

Can you give sample of the script you use? Just in case it twixes my mind what it may be?  Smiley


Unlike the other miners, when you CTRL+C to break, SRBMiner just dunps you back to the shell prompt rather than going to the next line.  The other miners will go to the next line after you break.

Code:
#!/bin/bash

sudo sysctl -w vm.nr_hugepages=1280

whoknows=0
while [ $whoknows -le 1 ]
do
    sleep 2
    ./SRBMiner-MULTI --give-up-limit 0 --disable-gpu --algorithm scryptn2 --pool stratum+tcp://scryptn2.na.mine.zpool.ca:4339 --wallet xxxx --password myrigname --cpu-threads 8

    sleep 2
    ./SRBMiner-MULTI --give-up-limit 0 --disable-gpu --algorithm Rx2 --pool stratum+tcp://rx2.mine.zergpool.com:8335 --nicehash true --send-stales true --api-enable --api-port 4068  --wallet xxx --password myrigname --cpu-threads 8

    sleep 2
    ./SRBMiner-MULTI --give-up-limit 0 --disable-gpu --algorithm curve --pool stratum+tcp://curvehash.na.mine.zergpool.com:3343 --wallet xxxx --password myrigname --cpu-threads 8

    sleep 2
    ./xmrig-cpu -r 0 -a cn/upx2 -o stratum+tcp://cryptonight_upx.na.mine.zergpool.com:4457 -u xxxx -p myrigname -t 8

    sleep 2
    ./xmrig-cpu -r 0 -a rx/0 -o stratum+tcp://randomx.na.mine.zergpool.com:4453 -u xxxx -p myrigname -t 8

    sleep 2
    ./cpuminer-opt-sse2 -r 0 -a yescrypt -o stratum+tcp://yescrypt.na.mine.zergpool.com:6233 -u xxxx -p myrigname --cpu-affinity AAAA --cpu-priority 5 -t 8

done
echo Loop ended.
hero member
Activity: 2548
Merit: 626
August 18, 2021, 03:48:09 AM
Do you wish to add support for 4000G or 5000G iGPU in the Miner?

list-gpu-devices says and lists the device but the Actual miner application just ignored it when trying to use it.

Said no Suitable GPU devices found?

Not something you wish to add? I was just curious what kind of performance one could get but could not start to even test.

I don't have a ryzen with an integrated gpu so i can't do much here.
member
Activity: 325
Merit: 42
August 16, 2021, 07:05:15 AM
I already do a sleep 1.  Do I need 2 seconds?

Did you get it to run in a terminal with the suggestions?

Try different sleep settings to see if it works. It also depends what cpu/mem you have in other words the faster your system is the longer sleep you might need.


I tried it with longer sleep (up to 10 seconds) same issue...just exits out of the script as soon as it self terminates.

Can you give sample of the script you use? Just in case it twixes my mind what it may be?  Smiley
Pages:
Jump to: