|
doktor83
|
 |
July 23, 2018, 12:11:27 PM |
|
Looks like that 18.7.1 is trash. Have you tried on 18.3.4 or BC drivers as i suggested? |
|
|
|
|
|
|
|
|
|
|
|
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction. Advertise here.
|
|
|
|
|
fellowbeing
Newbie
 Offline
Activity: 15
Merit: 0
|
|
July 23, 2018, 01:26:43 PM |
|
Vega FEs are allergic to 18.3.4 and I'll get less hash-rate on Alloy/Heavy and CN7, on Alloy there is alot of STALE shares too, so the only viable driver is 18.6.1.
Results for a rig having 5 Vega FEs SRB Miner 1.6.4:
driver 18.6.1
CN7 = 11060 H/s
Alloy = 5670 H/s
note: Mining Alloy, on DevFee we get miner crash or error
driver 18.3.4
CN7 = 10900 H/s
Alloy = 5580 H/s
note: Mining Alloy, there is a lot of STALE shares.
Results for a rig having 5 Vega FEs SRB Miner 1.6.1:
driver 18.6.1
CN7 = 10910 H/s
Alloy = 5630 H/s
driver Blockchain Aug 23
CN7 = 10910 H/s
Alloy = 5630 H/s
=========================================
So in conclusion:
18.3.4 won't work for Vega FEs. For Alloy we have to use 1.6.1 and we get similar hashrates as blockchain driver.
Edited: The said crash/error occurs not on DevFee but right after when resuming!
Stales are not a problem if they are accepted,i guess the difficulty on the pool you use is low, and the pool sends jobs quickly, so that is why you get a lot of stales, but if they are accepted then you have no problem. Also could you turn on logging (use --logfile log.txt in bat added) maybe i can get more info on the crash ? If i understand you correctly only on 18.6.1 is the miner crashing, on 18.3.4 it isnt ? Yes, I'll do this and see what we get. Regarding the drivers, I think 18.3.4 is just not working for me, even on CN7 I get lower hash-rate simiar to what I used to get on Blockchain driver about 150 H/s less or 50 H/s less on heavy/Alloy, but 18.6.1 has been very stable for me. So I avoid 18.3.4 totally. As for heavy/Alloy algo while using the SRB 1.6.4, I get GPU crash right after DevFee sometimes recovers after resetting the miner but other times I have to close the miner and restart it manually. I noticed this happened on 1.6.1 too but apparently less often. I don't have a clue what the cause could be, but I guess it might be something to do with change of algo from Alloy to DevFee and back. Unfortunately I have not had the chance to try any other heavy algo yet. I'll set the log and see what happens and come back. Thanks |
|
|
|
|
|
doktor83
|
|
July 23, 2018, 01:55:44 PM |
|
Vega FEs are allergic to 18.3.4 and I'll get less hash-rate on Alloy/Heavy and CN7, on Alloy there is alot of STALE shares too, so the only viable driver is 18.6.1.
Results for a rig having 5 Vega FEs SRB Miner 1.6.4:
driver 18.6.1
CN7 = 11060 H/s
Alloy = 5670 H/s
note: Mining Alloy, on DevFee we get miner crash or error
driver 18.3.4
CN7 = 10900 H/s
Alloy = 5580 H/s
note: Mining Alloy, there is a lot of STALE shares.
Results for a rig having 5 Vega FEs SRB Miner 1.6.1:
driver 18.6.1
CN7 = 10910 H/s
Alloy = 5630 H/s
driver Blockchain Aug 23
CN7 = 10910 H/s
Alloy = 5630 H/s
=========================================
So in conclusion:
18.3.4 won't work for Vega FEs. For Alloy we have to use 1.6.1 and we get similar hashrates as blockchain driver.
Edited: The said crash/error occurs not on DevFee but right after when resuming!
Stales are not a problem if they are accepted,i guess the difficulty on the pool you use is low, and the pool sends jobs quickly, so that is why you get a lot of stales, but if they are accepted then you have no problem. Also could you turn on logging (use --logfile log.txt in bat added) maybe i can get more info on the crash ? If i understand you correctly only on 18.6.1 is the miner crashing, on 18.3.4 it isnt ? Yes, I'll do this and see what we get. Regarding the drivers, I think 18.3.4 is just not working for me, even on CN7 I get lower hash-rate simiar to what I used to get on Blockchain driver about 150 H/s less or 50 H/s less on heavy/Alloy, but 18.6.1 has been very stable for me. So I avoid 18.3.4 totally. As for heavy/Alloy algo while using the SRB 1.6.4, I get GPU crash right after DevFee sometimes recovers after resetting the miner but other times I have to close the miner and restart it manually. I noticed this happened on 1.6.1 too but apparently less often. I don't have a clue what the cause could be, but I guess it might be something to do with change of algo from Alloy to DevFee and back. Unfortunately I have not had the chance to try any other heavy algo yet. I'll set the log and see what happens and come back. Thanks There is no change of algo when doing devfee mining, if you mine alloy then the devfee is also doing alloy, if mining heavy then its heavy etc. Also did you try switching pools manually (by defining more than one pool in pools.txt and pressing p) ? Is this crashing the miner ? |
|
|
|
|
qwep1
|
|
July 23, 2018, 09:09:34 PM |
|
hi all what setting for bittube for 580 4 gb
|
|
|
|
UAGet
Newbie
Offline
Activity: 66
Merit: 0
|
|
July 24, 2018, 05:37:58 AM |
|
Hi! Are there plans to support the purk algorithm? |
|
|
|
|
|
doktor83
|
|
July 24, 2018, 06:33:03 AM |
|
hi all what setting for bittube for 580 4 gb
Auto intensity should be good enough, also you can try the value auto gives + 1. 4GB cards work better with 1 thread on tube. so this should be good:
{
"cryptonight_type" : "bittubev2",
"intensity" : 0,
"double_threads" : false,
} |
|
|
|
|
doktor83
|
|
July 24, 2018, 06:33:40 AM |
|
Hi! Are there plans to support the purk algorithm? Nope, it's not cryptonight algo. |
|
|
|
kryx
Newbie
Offline
Activity: 60
Merit: 0
|
|
July 24, 2018, 07:55:11 AM |
|
Does this miner works with the old R9 280x ?
Anyone tried?
|
|
|
|
|
|
doktor83
|
|
July 24, 2018, 07:56:14 AM |
|
Does this miner works with the old R9 280x ?
Anyone tried?
yes it works |
|
|
|
chafer99
Member
Offline
Activity: 77
Merit: 10
|
|
July 24, 2018, 09:34:50 AM |
|
SRB is best for my 4xR380 (CN7 - 740H/s per card - 1050/1600Mhz) but very unstable: "application has been blocked from accessing graphics hardware".
The same problems as default settings (1000/1500Mhz). Driver 18.5.1 whql, Win10 (1803). Claymore 11.3 is OK.
What settings for RX580 8Gb? I got only 820H/s in CN7 (intensity 60/8/2, 1430/2200Mhz).
|
|
|
|
|
|
doktor83
|
|
July 24, 2018, 09:44:35 AM |
|
SRB is best for my 4xR380 (CN7 - 740H/s per card - 1050/1600Mhz) but very unstable: "application has been blocked from accessing graphics hardware".
The same problems as default settings (1000/1500Mhz). Driver 18.5.1 whql, Win10 (1803). Claymore 11.3 is OK.
What settings for RX580 8Gb? I got only 820H/s in CN7 (intensity 60/8/2, 1430/2200Mhz).
Application blocked text = driver crash Try with 1 thread, i had better results with 1 thread when testing on 380 4g. 56/8/2 is ok for 580 8g. Are you using a custom strap? Memory brand? |
|
|
|
|
JuanHungLo
|
|
July 24, 2018, 11:01:50 AM |
|
Hi! Are there plans to support the purk algorithm? Nope, it's not cryptonight algo. From Purk's page: Based on Boolberry & CryptoNote technology,... |
Bull markets are born on pessimism, grow on skepticism, mature on optimism, and die on euphoria. - John Templeton
|
|
|
|
doktor83
|
|
July 24, 2018, 11:04:04 AM |
|
Hi! Are there plans to support the purk algorithm? Nope, it's not cryptonight algo. From Purk's page: Based on Boolberry & CryptoNote technology,... From purk.miner.rocks : Please use supported mining apps and configs for wildkeccak |
|
|
|
SpceGhst
Jr. Member
Offline
Activity: 224
Merit: 1
|
|
July 24, 2018, 11:12:21 AM |
|
Hi! Are there plans to support the purk algorithm? Nope, it's not cryptonight algo. From Purk's page: Based on Boolberry & CryptoNote technology,... If you read the ANN thread, algo is wild keccak... https://bitcointalk.org/index.php?topic=3213127.0 |
|
|
|
|
chafer99
Member
Offline
Activity: 77
Merit: 10
|
|
July 24, 2018, 11:31:26 AM |
|
SRB is best for my 4xR380 (CN7 - 740H/s per card - 1050/1600Mhz) but very unstable: "application has been blocked from accessing graphics hardware".
The same problems as default settings (1000/1500Mhz). Driver 18.5.1 whql, Win10 (1803). Claymore 11.3 is OK.
What settings for RX580 8Gb? I got only 820H/s in CN7 (intensity 60/8/2, 1430/2200Mhz).
Application blocked text = driver crash Try with 1 thread, i had better results with 1 thread when testing on 380 4g. 56/8/2 is ok for 580 8g. Are you using a custom strap? Memory brand? Thanks, now intensity 56, thread 1 - 700H/s (R380 2g, samsung) looks stable. (previous 31/8/2 - 740H/s) RX580 in default (no custom straps/mem). But not bad in Heavy algo - 1000H/s |
|
|
|
|
|
|
|
doktor83
|
|
July 24, 2018, 11:36:30 AM |
|
SRB is best for my 4xR380 (CN7 - 740H/s per card - 1050/1600Mhz) but very unstable: "application has been blocked from accessing graphics hardware".
The same problems as default settings (1000/1500Mhz). Driver 18.5.1 whql, Win10 (1803). Claymore 11.3 is OK.
What settings for RX580 8Gb? I got only 820H/s in CN7 (intensity 60/8/2, 1430/2200Mhz).
Application blocked text = driver crash Try with 1 thread, i had better results with 1 thread when testing on 380 4g. 56/8/2 is ok for 580 8g. Are you using a custom strap? Memory brand? Thanks, now intensity 56, thread 1 - 700H/s (R380 2g, samsung) looks stable. (previous 31/8/2 - 740H/s) RX580 in default (no custom straps/mem). But not bad in Heavy algo - 1000H/s When i was testing r9 380 with double threads i could not go beyond intensity 28-29, it simply always crashed the driver. |
|
|
|
|
doktor83
|
|
July 24, 2018, 11:37:54 AM |
|
so i checked out the opencl kernel of purk and it has nothing to do with cryptonight algo. That's why there are wildkeccak miners for purk, not cryptonight miners. |
|
|
|
chafer99
Member
Offline
Activity: 77
Merit: 10
|
|
July 24, 2018, 12:18:30 PM |
|
SRB is best for my 4xR380 (CN7 - 740H/s per card - 1050/1600Mhz) but very unstable: "application has been blocked from accessing graphics hardware".
The same problems as default settings (1000/1500Mhz). Driver 18.5.1 whql, Win10 (1803). Claymore 11.3 is OK.
What settings for RX580 8Gb? I got only 820H/s in CN7 (intensity 60/8/2, 1430/2200Mhz).
Application blocked text = driver crash Try with 1 thread, i had better results with 1 thread when testing on 380 4g. 56/8/2 is ok for 580 8g. Are you using a custom strap? Memory brand? Thanks, now intensity 56, thread 1 - 700H/s (R380 2g, samsung) looks stable. (previous 31/8/2 - 740H/s) RX580 in default (no custom straps/mem). But not bad in Heavy algo - 1000H/s When i was testing r9 380 with double threads i could not go beyond intensity 28-29, it simply always crashed the driver. at 31 intensity works, but after 4-5 hours driver crashed, (the same as 28-30 intensity). https://imgur.com/oTgtb0p |
|
|
|
|
hesido
Newbie
Offline
Activity: 85
Merit: 0
|
|
July 24, 2018, 09:42:20 PM |
|
SRB is best for my 4xR380 (CN7 - 740H/s per card - 1050/1600Mhz) but very unstable: "application has been blocked from accessing graphics hardware".
The same problems as default settings (1000/1500Mhz). Driver 18.5.1 whql, Win10 (1803). Claymore 11.3 is OK.
What settings for RX580 8Gb? I got only 820H/s in CN7 (intensity 60/8/2, 1430/2200Mhz).
Application blocked text = driver crash Try with 1 thread, i had better results with 1 thread when testing on 380 4g. 56/8/2 is ok for 580 8g. Are you using a custom strap? Memory brand? Thanks, now intensity 56, thread 1 - 700H/s (R380 2g, samsung) looks stable. (previous 31/8/2 - 740H/s) RX580 in default (no custom straps/mem). But not bad in Heavy algo - 1000H/s When i was testing r9 380 with double threads i could not go beyond intensity 28-29, it simply always crashed the driver. at 31 intensity works, but after 4-5 hours driver crashed, (the same as 28-30 intensity). https://imgur.com/oTgtb0pTry to adjust your OC, most instabilities I've had could be solved by throwing more voltage at the cards or reducing the frequencies. |
|
|
|
|
|
