Topic: SRBMiner Cryptonight AMD GPU Miner V1.9.3 - native algo switching - page 96. (Read 237247 times)
November 21, 2018, 08:56:44 AM
Miner stops working without any errors. It just stops, but does not hang. With what it can be connected.
November 21, 2018, 07:12:14 AM
Hello Doktor,
So I usually enjoy using your miner. But since the v8 fork, I cannot use "gpu_conf" : command anymore. When I add some values to the integrity, worksize etc, the miner will not start but just shuts down for no reason. But if I delete the "gpu_conf" : line with its values, miner works fine. Its giving me headache like for real. Can you help me on this one?
So I usually enjoy using your miner. But since the v8 fork, I cannot use "gpu_conf" : command anymore. When I add some values to the integrity, worksize etc, the miner will not start but just shuts down for no reason. But if I delete the "gpu_conf" : line with its values, miner works fine. Its giving me headache like for real. Can you help me on this one?
have error in config, post and we help you fix that
November 21, 2018, 05:53:51 AM
Hello Doktor,
So I usually enjoy using your miner. But since the v8 fork, I cannot use "gpu_conf" : command anymore. When I add some values to the integrity, worksize etc, the miner will not start but just shuts down for no reason. But if I delete the "gpu_conf" : line with its values, miner works fine. Its giving me headache like for real. Can you help me on this one?
So I usually enjoy using your miner. But since the v8 fork, I cannot use "gpu_conf" : command anymore. When I add some values to the integrity, worksize etc, the miner will not start but just shuts down for no reason. But if I delete the "gpu_conf" : line with its values, miner works fine. Its giving me headache like for real. Can you help me on this one?
can you post your gpu conf file that errors out?
November 21, 2018, 02:39:49 AM
Hello Doktor,
So I usually enjoy using your miner. But since the v8 fork, I cannot use "gpu_conf" : command anymore. When I add some values to the integrity, worksize etc, the miner will not start but just shuts down for no reason. But if I delete the "gpu_conf" : line with its values, miner works fine. Its giving me headache like for real. Can you help me on this one?
So I usually enjoy using your miner. But since the v8 fork, I cannot use "gpu_conf" : command anymore. When I add some values to the integrity, worksize etc, the miner will not start but just shuts down for no reason. But if I delete the "gpu_conf" : line with its values, miner works fine. Its giving me headache like for real. Can you help me on this one?
November 20, 2018, 10:15:31 PM
Hi Doctor,
As a sample running haven for 1.6.8 vs 1.7.0, I am using gpu card asus rog rx580 8gb, memory type micron
My Mem Strap : 777000000000000022AA1C0073626C41B0551016BA0D260B006AE60004061420EA8940AA0300000 01914292EB22E3B16 (From Pimp SRBPolaris)
cclock/mclock : 1160/2200
AMD Driver : 18.6.1
using 1.6.8 I can get stable @1052
using 1.7.0:
{ "id" : 0, "intensity" : 59 , "worksize" : 8, "fragment" : 4, "threads" : 2},
the same config using various "fragments" :
==> 0 = 906 ,1 = 1009 , 2 = 911 , 4 = 1010 , 8 = 1007 , 16 = 1009 , 32 = 1010 , 64 = 1005, 128 = 966
using fragment 4 with "worksize" = 08, various intensity : 59=1011 58=1000 57=997 56=891 55=882 54=886 52=972 48=958
using fragment 4 with "worksize" = 16, various intensity : 59=0908 58=0903 57=897 56= 55= 54= 52=867 48=840
using various "aes", the result is no different, but default/0 more stable
using config suggested here :
{ "id" : 0, "intensity" : 58 , "worksize" : 16, "fragment" : 4, "threads" : 2} = 903
{ "id" : 0, "intensity" : 59 , "worksize" : 8, "fragment" : 4, "threads" : 2} = 1000
any more suggestion?
Thanks.
As a sample running haven for 1.6.8 vs 1.7.0, I am using gpu card asus rog rx580 8gb, memory type micron
My Mem Strap : 777000000000000022AA1C0073626C41B0551016BA0D260B006AE60004061420EA8940AA0300000 01914292EB22E3B16 (From Pimp SRBPolaris)
cclock/mclock : 1160/2200
AMD Driver : 18.6.1
using 1.6.8 I can get stable @1052
using 1.7.0:
{ "id" : 0, "intensity" : 59 , "worksize" : 8, "fragment" : 4, "threads" : 2},
the same config using various "fragments" :
==> 0 = 906 ,1 = 1009 , 2 = 911 , 4 = 1010 , 8 = 1007 , 16 = 1009 , 32 = 1010 , 64 = 1005, 128 = 966
using fragment 4 with "worksize" = 08, various intensity : 59=1011 58=1000 57=997 56=891 55=882 54=886 52=972 48=958
using fragment 4 with "worksize" = 16, various intensity : 59=0908 58=0903 57=897 56= 55= 54= 52=867 48=840
using various "aes", the result is no different, but default/0 more stable
using config suggested here :
{ "id" : 0, "intensity" : 58 , "worksize" : 16, "fragment" : 4, "threads" : 2} = 903
{ "id" : 0, "intensity" : 59 , "worksize" : 8, "fragment" : 4, "threads" : 2} = 1000
any more suggestion?
Thanks.
November 20, 2018, 02:56:30 PM
Hi! I have a Ryzen 1700+Vega 56 on a MSI x370 Carbon Pro Gaming, recently reformated with Windows 10 version 1809, latest mother drivers + Radeon 18.11.1
I have a strange problem. If I start Windows and run SRB right away I get 1850H/s. If I leave it alone for a couple of minutes (?) and then I start SRB, I get 1200H/s. I have to restart Windows or suspend/resume Windows to get 1850H/s again. I know that the suggestion is to have other radeon drivers but last week I had 18.5.1 and was having the same problem.
Is there a way to fix this? Or maybe restart the driver without the need of restarting Windows or suspending/resuming?
I have a strange problem. If I start Windows and run SRB right away I get 1850H/s. If I leave it alone for a couple of minutes (?) and then I start SRB, I get 1200H/s. I have to restart Windows or suspend/resume Windows to get 1850H/s again. I know that the suggestion is to have other radeon drivers but last week I had 18.5.1 and was having the same problem.
Is there a way to fix this? Or maybe restart the driver without the need of restarting Windows or suspending/resuming?
November 20, 2018, 12:16:50 PM
Repeat of question:
Whats happening with the RX 550 2GB with the 1.7.0 version?
When i run benchmark mode the hashrate is 465 (lower than the previous 490), but when i run the miner it drops to 365 h/s. Algo is normalv8. The RX 570 perform better than before, the RX 460 are at the same level.
Doctor, do you have RX 550 cards to test this?
Any answer would be much appreciated because my 12xgpu RX 550 rig is hashing only 2/3 of what it did. As i said, the two RX 570 rigs work even better than before. Great job on that!!!
Whats happening with the RX 550 2GB with the 1.7.0 version?
When i run benchmark mode the hashrate is 465 (lower than the previous 490), but when i run the miner it drops to 365 h/s. Algo is normalv8. The RX 570 perform better than before, the RX 460 are at the same level.
Doctor, do you have RX 550 cards to test this?
Any answer would be much appreciated because my 12xgpu RX 550 rig is hashing only 2/3 of what it did. As i said, the two RX 570 rigs work even better than before. Great job on that!!!
Don't repeat, but read back. Try various values for 'fragments' parameter. Benchmark uses same config as normal mining.
I did read back and i will try the fragments parameter today. Is there a default value for fragments and whats the purpose of it?
Currently i have the same settings for normalv8 in benchmark and mining mode but the hashrate is different, that's why i popped the question.
Thanks:)
November 20, 2018, 12:10:43 PM
Repeat of question:
Whats happening with the RX 550 2GB with the 1.7.0 version?
When i run benchmark mode the hashrate is 465 (lower than the previous 490), but when i run the miner it drops to 365 h/s. Algo is normalv8. The RX 570 perform better than before, the RX 460 are at the same level.
Doctor, do you have RX 550 cards to test this?
Any answer would be much appreciated because my 12xgpu RX 550 rig is hashing only 2/3 of what it did. As i said, the two RX 570 rigs work even better than before. Great job on that!!!
Whats happening with the RX 550 2GB with the 1.7.0 version?
When i run benchmark mode the hashrate is 465 (lower than the previous 490), but when i run the miner it drops to 365 h/s. Algo is normalv8. The RX 570 perform better than before, the RX 460 are at the same level.
Doctor, do you have RX 550 cards to test this?
Any answer would be much appreciated because my 12xgpu RX 550 rig is hashing only 2/3 of what it did. As i said, the two RX 570 rigs work even better than before. Great job on that!!!
What is your settings in config , I set sensitivity to 27 worksize 24 thread 2, memmory 1970 core 1170-1200 880mV 460mh/s per card ,maybe can better ,but this is stable for me.
Hello. My settings are intensity 26, worksize 8, threads 2 and the same when i run the benchmark but the result is different on normal v8 algo.
November 20, 2018, 11:42:01 AM
Hello!
Spent a few days trying different settings, miners, and drivers.
This is my current setup:
DDR4 4GB
G3930
RX588 x8
-cclock 1200 -mclock 2250 -powlim -20 -cvddc 900 -mvddc 825
Adrenalin 18.3.4
900w total system (measured at the wall)
8300hs xmrig-amd286 (944x8), 8300hs with Adrelanin 18.5.1 and 864x8
8000hs jce33b4 (944x8)
8100hs srbminer170 (i54w8t2f16)
jce33b5 has "Crazy huge performance boost for Heavy/HVX/Tube + 20%" when what it really has is a "Crazy huge fee boost for dev + 233%"
When a dev releases a faster miner, everyone is forced to switch over or be left behind. After upgrading all will earn the same as before unless, the dev increases the fee!
But public development is good because we need to keep up with secretly developed miners.
SRBMiner has only 0.85% fee instead off 2.1%.
This project has a good dev and a great miner, and everyone should support it.
Altough inevitable, the world should not end in the hands of greedy people.
Over and out.
Spent a few days trying different settings, miners, and drivers.
This is my current setup:
DDR4 4GB
G3930
RX588 x8
-cclock 1200 -mclock 2250 -powlim -20 -cvddc 900 -mvddc 825
Adrenalin 18.3.4
900w total system (measured at the wall)
Quote
{
"cryptonight_type" : "haven",
"gpu_conf" :
[
{ "id" : 0, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 1, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 2, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 3, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 4, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 5, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 6, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 7, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
]
}
"cryptonight_type" : "haven",
"gpu_conf" :
[
{ "id" : 0, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 1, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 2, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 3, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 4, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 5, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 6, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
{ "id" : 7, "intensity" : 54, "worksize" : 8, "threads" : 2, "fragments" : 16},
]
}
8300hs xmrig-amd286 (944x8), 8300hs with Adrelanin 18.5.1 and 864x8
8000hs jce33b4 (944x8)
8100hs srbminer170 (i54w8t2f16)
jce33b5 has "Crazy huge performance boost for Heavy/HVX/Tube + 20%" when what it really has is a "Crazy huge fee boost for dev + 233%"
When a dev releases a faster miner, everyone is forced to switch over or be left behind. After upgrading all will earn the same as before unless, the dev increases the fee!
But public development is good because we need to keep up with secretly developed miners.
SRBMiner has only 0.85% fee instead off 2.1%.
This project has a good dev and a great miner, and everyone should support it.
Altough inevitable, the world should not end in the hands of greedy people.
Over and out.
November 20, 2018, 09:58:27 AM
1.7.0 is working great for me on my 4x RX 560 (XfX 2GB) rig mining CNv8. I'm getting 460-490H/s per card with intensity set to 25.0 and using MSI AB 4.5.0 to set core at 1200, mem at 1900 and 100% PL (running an Anorak powersave BIOS). I really like the restart miner if hashrate drops below a certain value function as it seems that one or both of the middle two cards on this rig will drop in hashrate by about 150H/s every so often.
All in all, I'm happy to pay the dev fee vs. running xmr-stak compiled to have no devfee, as the latter barely did better than 400H/s per card and it had stability problems when it did run. Unfortunately, I am running at a razor thin margin of profitability vs. power cost right now, and it is only because the weather turned cooler recently that I am keeping this rig running at all (everything else was turned off and/or sold earlier this year).
All in all, I'm happy to pay the dev fee vs. running xmr-stak compiled to have no devfee, as the latter barely did better than 400H/s per card and it had stability problems when it did run. Unfortunately, I am running at a razor thin margin of profitability vs. power cost right now, and it is only because the weather turned cooler recently that I am keeping this rig running at all (everything else was turned off and/or sold earlier this year).
Sounds like a power issue. My rig can exhibit problems like that if I re-apply ODT settings and/or restart the miner several times. It seems like some "power degradation" thingy happens when the cards/system are cycled too many times. The only remedy is to completely power off the rig including the psu for at least 5 minutes then power back on.
November 19, 2018, 11:21:03 PM
SRBMiner 1.7.0 already added to https://www.minermonitoring.com/ https://www.ethmonitoring.com/ Join our discord https://discord.gg/aGGaeqq
also, you can check the speed of the miner on different rigs using the benchmark page: https://minermonitoring.com/benchmark
also, you can check the speed of the miner on different rigs using the benchmark page: https://minermonitoring.com/benchmark
November 19, 2018, 04:50:30 PM
Repeat of question:
Whats happening with the RX 550 2GB with the 1.7.0 version?
When i run benchmark mode the hashrate is 465 (lower than the previous 490), but when i run the miner it drops to 365 h/s. Algo is normalv8. The RX 570 perform better than before, the RX 460 are at the same level.
Doctor, do you have RX 550 cards to test this?
Any answer would be much appreciated because my 12xgpu RX 550 rig is hashing only 2/3 of what it did. As i said, the two RX 570 rigs work even better than before. Great job on that!!!
Whats happening with the RX 550 2GB with the 1.7.0 version?
When i run benchmark mode the hashrate is 465 (lower than the previous 490), but when i run the miner it drops to 365 h/s. Algo is normalv8. The RX 570 perform better than before, the RX 460 are at the same level.
Doctor, do you have RX 550 cards to test this?
Any answer would be much appreciated because my 12xgpu RX 550 rig is hashing only 2/3 of what it did. As i said, the two RX 570 rigs work even better than before. Great job on that!!!
What is your settings in config , I set sensitivity to 27 worksize 24 thread 2, memmory 1970 core 1170-1200 880mV 460mh/s per card ,maybe can better ,but this is stable for me.
November 19, 2018, 04:10:43 PM
Repeat of question:
Whats happening with the RX 550 2GB with the 1.7.0 version?
When i run benchmark mode the hashrate is 465 (lower than the previous 490), but when i run the miner it drops to 365 h/s. Algo is normalv8. The RX 570 perform better than before, the RX 460 are at the same level.
Doctor, do you have RX 550 cards to test this?
Any answer would be much appreciated because my 12xgpu RX 550 rig is hashing only 2/3 of what it did. As i said, the two RX 570 rigs work even better than before. Great job on that!!!
Whats happening with the RX 550 2GB with the 1.7.0 version?
When i run benchmark mode the hashrate is 465 (lower than the previous 490), but when i run the miner it drops to 365 h/s. Algo is normalv8. The RX 570 perform better than before, the RX 460 are at the same level.
Doctor, do you have RX 550 cards to test this?
Any answer would be much appreciated because my 12xgpu RX 550 rig is hashing only 2/3 of what it did. As i said, the two RX 570 rigs work even better than before. Great job on that!!!
Don't repeat, but read back. Try various values for 'fragments' parameter. Benchmark uses same config as normal mining.
November 19, 2018, 04:03:31 PM
Repeat of question:
Whats happening with the RX 550 2GB with the 1.7.0 version?
When i run benchmark mode the hashrate is 465 (lower than the previous 490), but when i run the miner it drops to 365 h/s. Algo is normalv8. The RX 570 perform better than before, the RX 460 are at the same level.
Doctor, do you have RX 550 cards to test this?
Any answer would be much appreciated because my 12xgpu RX 550 rig is hashing only 2/3 of what it did. As i said, the two RX 570 rigs work even better than before. Great job on that!!!
Whats happening with the RX 550 2GB with the 1.7.0 version?
When i run benchmark mode the hashrate is 465 (lower than the previous 490), but when i run the miner it drops to 365 h/s. Algo is normalv8. The RX 570 perform better than before, the RX 460 are at the same level.
Doctor, do you have RX 550 cards to test this?
Any answer would be much appreciated because my 12xgpu RX 550 rig is hashing only 2/3 of what it did. As i said, the two RX 570 rigs work even better than before. Great job on that!!!
November 19, 2018, 12:39:18 PM
Thank you dudes, those timings did the job for me! :* :* :*
Are there any even better straps for Elpida or Samsung
Are there any even better straps for Elpida or Samsung
November 19, 2018, 11:49:08 AM
edited
580/8 srbminer 1.7.0 1230/2100 (880mv) - 1135-1140h/s haven
"gpu_conf" :
[
{ "id" : 0, "intensity" : 54, "worksize" : 8, "fragments" : 1, "threads" : 2},
{ "id" : 1, "intensity" : 54, "worksize" : 8, "fragments" : 1, "threads" : 2},
{ "id" : 2, "intensity" : 54, "worksize" : 8, "fragments" : 1, "threads" : 2},
{ "id" : 3, "intensity" : 54, "worksize" : 8, "fragments" : 1, "threads" : 2},
{ "id" : 4, "intensity" : 54, "worksize" : 8, "fragments" : 1, "threads" : 2},
{ "id" : 5, "intensity" : 54, "worksize" : 8, "fragments" : 1, "threads" : 2},
]
}
HYNIX H5GQ8H24MJR
999000000000000022339D00CEDD6B4470551414B80D960B0040C400750314204A8900A00200000 018123745B92B3A16
580/8 srbminer 1.7.0 1230/2100 (880mv) - 1135-1140h/s haven
"gpu_conf" :
[
{ "id" : 0, "intensity" : 54, "worksize" : 8, "fragments" : 1, "threads" : 2},
{ "id" : 1, "intensity" : 54, "worksize" : 8, "fragments" : 1, "threads" : 2},
{ "id" : 2, "intensity" : 54, "worksize" : 8, "fragments" : 1, "threads" : 2},
{ "id" : 3, "intensity" : 54, "worksize" : 8, "fragments" : 1, "threads" : 2},
{ "id" : 4, "intensity" : 54, "worksize" : 8, "fragments" : 1, "threads" : 2},
{ "id" : 5, "intensity" : 54, "worksize" : 8, "fragments" : 1, "threads" : 2},
]
}
HYNIX H5GQ8H24MJR
999000000000000022339D00CEDD6B4470551414B80D960B0040C400750314204A8900A00200000 018123745B92B3A16
This strap is good for this type of memmory ,but gpu conf isnt good, only 330 mhs per card
"gpu_conf" :
[
{ "id" : 0, "intensity" : 100, "worksize" : 8, "threads" : 2, "fragments" : 16, "bralock" : 50},
{ "id" : 1, "intensity" : 100, "worksize" : 8, "threads" : 2, "fragments" : 16, "bralock" : 50},
{ "id" : 2, "intensity" : 100, "worksize" : 8, "threads" : 2, "fragments" : 16, "bralock" : 50}
]
}
You can also try Intesity setup up to 116
king_pin Read this : use this strap and second "gpu config" 1230 gpu 2100 memmory 950mV power, happy v8 minning.
November 19, 2018, 10:45:18 AM
Cao brate, vidim da si nas, uzeo sam 4 riga RX580 8gb Hinyx memorija,probao razne biose i sa Anorak sajta i radio sa PBE1.6 i SRB editorom , i na CN V8 algo dobijam plafon 880-900 mhs ,
sa clockovima core 1200 memorija 2200-2250, i cak i ne radi stabilno ,miner je SRB miner i intensity je 0,to jest automatski ,probao sam i sa 60 i 70 intensity workload 8 i 16 thread 2 , i to neradi stabilno uospte, Recimo na 470-4 gb sa biosom za ETH koji je uradjen SRB polaris editorom dobijam 930 mhs po kartici ,reci mi jel bih mogao da mi pomognes sta jos da probam i pustim u rad vec jednom ova 4 riga , na ethu daju kartice 32 mhs i rade stabilno na V8 algo nikako. Hvala unapred
Problem je strap za te memorije. Na eth radi odlicno ali na monero bas lose. Jedino ako je neko napravio bolji do sada pa da se nadje.sa clockovima core 1200 memorija 2200-2250, i cak i ne radi stabilno ,miner je SRB miner i intensity je 0,to jest automatski ,probao sam i sa 60 i 70 intensity workload 8 i 16 thread 2 , i to neradi stabilno uospte, Recimo na 470-4 gb sa biosom za ETH koji je uradjen SRB polaris editorom dobijam 930 mhs po kartici ,reci mi jel bih mogao da mi pomognes sta jos da probam i pustim u rad vec jednom ova 4 riga , na ethu daju kartice 32 mhs i rade stabilno na V8 algo nikako. Hvala unapred
So my problem are the straps too? I tried all other setting (intensity, fragments, wroksize etc.)
Which straps for RX 580 8G Hynix or cards with HYNIX memory in general are the best for V8...
P.S. Greetings Slav brothers
November 18, 2018, 04:48:59 PM
Guys, please tell me how to fix this ?? >>
change the pool or port Code:
socket_error: A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond.
[2018-11-18 22:25:04] Connection to pool lost. Reconnecting in 10 seconds.
[2018-11-18 22:25:04] Connection to pool lost. Reconnecting in 10 seconds.
November 18, 2018, 04:37:02 PM
9 gpus x 8gb memmory = 72 gb x 1024 = 73728 gb of virtual memmory you need to set for these cards.
thank you. there is a feeling that need more. November 18, 2018, 04:32:50 PM
Guys, please tell me how to fix this ?? >>
Code:
socket_error: A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond.
[2018-11-18 22:25:04] Connection to pool lost. Reconnecting in 10 seconds.
[2018-11-18 22:25:04] Connection to pool lost. Reconnecting in 10 seconds.
Jump to: