SRBMiner-MULTI GPU & CPU Miner 0.9.4 - page 37.

doktor83

hero member

Activity: 2548

Merit: 626

Quote from: entryboy on February 12, 2021, 12:51:57 PM

Hi doctor83.

i have problem to mine ERGO (srb version 0.6.6) with 2 VEGA-64-LC (8 GB). It is not possible to mine more than 18 hours. After this time the cards need an reboot, because they can't reinit with a new instance -> mining stops.

Can you look. i know such problems when using timer (seconds clock) with only 16 bit resolution -> 16,8 hours -> overflow.

My driver version is 20.10.35.02 with HBCC ON (max) -> 2x230 Watt -> 225 MH (woolypool).

Thanks.

There is no problem with any timer or anything like that, i have uptimes of 40+ days

0.6.6 version has 2 threads / gpu for default, maybe that's what your vegas dont like, try --gpu-intensity 24, it's running 1 thread only.

doktor83

hero member

Activity: 2548

Merit: 626

Quote from: kavalerus on February 12, 2021, 09:39:12 AM

Hi doktor83,

Found an issue: Incorrect count of avg hashrate.
On my rig (6*rx480) example:
2 cards are died for some reason:
actual hashrate:
[2021-02-12 19:27:44] GPU0: 60.58 MH/s
[2021-02-12 19:27:44] GPU1: 63.86 MH/s
[2021-02-12 19:27:44] GPU2: 62.44 MH/s
[2021-02-12 19:27:44] GPU5: 69.83 MH/s
[2021-02-12 19:27:44] Total: 256.71 MH/s
But:
[2021-02-12 19:27:49] Avg. 1 min. : 379.61 MH/s
[2021-02-12 19:27:49] Avg. 5 min. : 374.77 MH/s
[2021-02-12 19:27:49] Avg. 15 min. : 375.71 MH/s
10 min later:
[2021-02-12 19:37:25] Avg. 1 min. : 376.24 MH/s
[2021-02-12 19:37:25] Avg. 5 min. : 374.24 MH/s
[2021-02-12 19:37:25] Avg. 15 min. : 374.30 MH/s

Thanks

Thanks for reporting, going to check it out

entryboy

newbie

Activity: 44

Merit: 0

Hi doctor83.

i have problem to mine ERGO (srb version 0.6.6) with 2 VEGA-64-LC (8 GB). It is not possible to mine more than 18 hours. After this time the cards need an reboot, because they can't reinit with a new instance -> mining stops.

Can you look. i know such problems when using timer (seconds clock) with only 16 bit resolution -> 16,8 hours -> overflow.

My driver version is 20.10.35.02 with HBCC ON (max) -> 2x230 Watt -> 225 MH (woolypool).

Thanks.

kavalerus

newbie

Activity: 2

Merit: 0

Hi doktor83,

Found an issue: Incorrect count of avg hashrate.
On my rig (6*rx480) example:
2 cards are died for some reason:
actual hashrate:
[2021-02-12 19:27:44] GPU0: 60.58 MH/s
[2021-02-12 19:27:44] GPU1: 63.86 MH/s
[2021-02-12 19:27:44] GPU2: 62.44 MH/s
[2021-02-12 19:27:44] GPU5: 69.83 MH/s
[2021-02-12 19:27:44] Total: 256.71 MH/s
But:
[2021-02-12 19:27:49] Avg. 1 min. : 379.61 MH/s
[2021-02-12 19:27:49] Avg. 5 min. : 374.77 MH/s
[2021-02-12 19:27:49] Avg. 15 min. : 375.71 MH/s
10 min later:
[2021-02-12 19:37:25] Avg. 1 min. : 376.24 MH/s
[2021-02-12 19:37:25] Avg. 5 min. : 374.24 MH/s
[2021-02-12 19:37:25] Avg. 15 min. : 374.30 MH/s

Thanks

remark78

newbie

Activity: 14

Merit: 0

Quote from: doktor83 on February 11, 2021, 05:39:30 AM

Quote from: remark78 on February 11, 2021, 05:22:09 AM

Please help,

I use SRBMiner-Multi 0.6.5 for mining Haven (XHV) coin on haven.herominers.com.

After some time an error occurred: "couldn't connect to a devfee pool for a long time ... pause mining"
After that I have to restart the miner to continue working.

Your firewall/antivirus is blockin miner from accessing the devfee pool probably

When the miner connects to devfee pool and what is its address?
The firewall/antivirus should not be a problem because it leaks properly haven.herominers.com.

doktor83

hero member

Activity: 2548

Merit: 626

Quote from: TeCe on February 11, 2021, 05:37:40 AM

Quote from: doktor83 on February 05, 2021, 07:37:09 AM

V0.6.5
+ Huge performance increase on 'autolykos2' algorithm on GPU's (up to ~600%)
+ Huge performance increase on 'autolykos2' algorithm on CPU's with AVX2 (up to ~500%)
+ Added abort mechanism for 'autolykos2' dataset creation on CPU

Hey Doc! Hugs you man and tnx a lot for such good work!

Have two qestions:
1. i test two days - MULTI vs 1.9.4 on Haven on four rigs(8 cards RX470).
For MULTI i try some gpu tweak profiles and boost ...(boost 15, tweak 3 is better ... BUT --->>> Still 1.9.4 on custom GPU config works better. (faster, get more shares in short time)
Can you approve this?

2. Is there any command in 1.9.4 miner to start numbering GPU from 1 (not 0)?

Thank you.

1. I think at poolside MULTI should show a little bit better results if you find all the tweaks/boosts and make GPU's work stable
2. Nope

doktor83

hero member

Activity: 2548

Merit: 626

Quote from: remark78 on February 11, 2021, 05:22:09 AM

Please help,

I use SRBMiner-Multi 0.6.5 for mining Haven (XHV) coin on haven.herominers.com.

After some time an error occurred: "couldn't connect to a devfee pool for a long time ... pause mining"
After that I have to restart the miner to continue working.

Your firewall/antivirus is blockin miner from accessing the devfee pool probably

doktor83

hero member

Activity: 2548

Merit: 626

Quote from: dfish987 on February 10, 2021, 02:34:01 AM

i'm using a few different cpus so i figured it out something to do with us.fastpool.xyz:3338 it gives me rejected shares (incorrect extranounce2 size) or doesn't find shares but if i connect to just fastpool.xyz it works fine, but your version 6.2 works with us.fastpool.xyz super strange
could you test us.fastpool.xyz:3338 and see if you get rejected shares... and curious part is if i use 6.2 it works with us.fastpool.xyz:3338

All good on latest version at us.fastpool.xyz:3338

TeCe

newbie

Activity: 14

Merit: 0

Quote from: doktor83 on February 05, 2021, 07:37:09 AM

V0.6.5
+ Huge performance increase on 'autolykos2' algorithm on GPU's (up to ~600%)
+ Huge performance increase on 'autolykos2' algorithm on CPU's with AVX2 (up to ~500%)
+ Added abort mechanism for 'autolykos2' dataset creation on CPU

Hey Doc! Hugs you man and tnx a lot for such good work!

Have two qestions:
1. i test two days - MULTI vs 1.9.4 on Haven on four rigs(8 cards RX470).
For MULTI i try some gpu tweak profiles and boost ...(boost 15, tweak 3 is better ... BUT --->>> Still 1.9.4 on custom GPU config works better. (faster, get more shares in short time)
Can you approve this?

2. Is there any command in 1.9.4 miner to start numbering GPU from 1 (not 0)?

Thank you.

abg00

jr. member

Activity: 288

Merit: 1

066 ergo coin
windows 20.11.2

rate 50%/ miner restart not restore rate/ only rig reboot

remark78

newbie

Activity: 14

Merit: 0

Please help,

I use SRBMiner-Multi 0.6.5 for mining Haven (XHV) coin on haven.herominers.com.

After some time an error occurred: "couldn't connect to a devfee pool for a long time ... pause mining"
After that I have to restart the miner to continue working.

dfish987

newbie

Activity: 27

Merit: 0

i'm using a few different cpus so i figured it out something to do with us.fastpool.xyz:3338 it gives me rejected shares (incorrect extranounce2 size) or doesn't find shares but if i connect to just fastpool.xyz it works fine, but your version 6.2 works with us.fastpool.xyz super strange
could you test us.fastpool.xyz:3338 and see if you get rejected shares... and curious part is if i use 6.2 it works with us.fastpool.xyz:3338

doktor83

hero member

Activity: 2548

Merit: 626

Quote from: dfish987 on February 10, 2021, 12:27:52 AM

tried to new 6.6 on randomhash2 doesn't submit any shares for over 4 minutes i waited and miner didn't find a single share and the diff at pool is not the problem your old 6.2 version works fine but 6.3,6.4,6.5 wont connect says something about extranouce wrong or something thanks

What cpu you have?

dfish987

newbie

Activity: 27

Merit: 0

tried to new 6.6 on randomhash2 doesn't submit any shares for over 4 minutes i waited and miner didn't find a single share and the diff at pool is not the problem your old 6.2 version works fine but 6.3,6.4,6.5 wont connect says something about extranouce wrong or something thanks

doktor83

hero member

Activity: 2548

Merit: 626

V0.6.6

+ Performance increase on 'autolykos2' algorithm on GPU's
+ Reduced power consumption on 'autolykos2' algorithm on GPU's
+ Faster dataset creation on 'autolykos2' algorithm on GPU's
+ Fixed miner crashing when creating dataset on 'autolykos2' algorithm
+ Fixed miner crashing when running --gpu-auto-tune on 'autolykos2' algorithm
+ Possible fix for 'duplicate share' issue on 'autolykos2' algorithm
+ Fixed 'k12' algorithm
+ Minor bug fixes

entryboy

newbie

Activity: 44

Merit: 0