Author

Topic: Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux) - page 968. (Read 6590718 times)

sr. member
Activity: 689
Merit: 253
Hi guys,
I am getting around 164 mh/s on eth and around 4900 on drc
http://imgur.com/a/mDwd6
I am with 6 470 cards and the power consumption is around 1200 w, is this normal? I overclocked the cards to 1250/2050, -100 mv with the trixx tool.
Is there a way to reduce that consumption?
Thanks!

Try a lower core clock like 1150, 1125, or 1100 and check power consumption.. should only reduce mhs a bit.. I use 1130.. also try more than -100mv?
sr. member
Activity: 449
Merit: 251
All of my DCR shares are getting rejected using 9.0. I mine on supernova (dcr.suprnova.cc)...any idea why? It has been working fine with 8.1.

I believe this problem is related to this pool, sometimes it does not send response to auth message for some reason but still sends jobs, so miner mines but pool does not accept shares.
Miner will reconnect automatically after 10-15 invalid shares, usually the pool accepts auth message at second attempt.
Suprnova Decred pool is complete and utter garbage.  It should be blacklisted from the internet. I would get 95-100% rejects on most rigs, best case like 5-10%. Switched to coinmine, now get under 1% reject.

Some suprnova pools are ok, many are junk.
newbie
Activity: 7
Merit: 0
My seasonic 750w gold  kept up for the last week dual mining gpu-z showing 110w for each card. Now they are undervolted. You mean this is not good for longterm? Mind you i have 4 cards not 6.
Well ive been mining for a week with old settings dual mining gpu-z showing 110w for each card no problem. How will you explain this? They are even undervolted now.
gpuz power in not the same as wall power !
i've :
4*470 on ETH only, ~1075/1900 980mv , 750w at wall
4*470 on ETH+DCR, ~1150/2000 980mv, 850w at wall ( dcri 25 with v9.0, was 35 with v8.1 )



I know its not the same. Thats why i stated that the cards are undervolted.

So lets say they each pull 40 watts more than what GPU-Z shows. that is 92.5+40=132.5 watts for each and 530watts for all 4 cards. Even if the rest of the system pulls 170w  half idleing which i highly doubt ( x2 240 and 1 ram with 1 hdd) that adds up to 700w. I will get a kill a watt next week to verify. I should safely assume i am safe considering the cards ran with 20watts more each for a week when they were not undervolted.
newbie
Activity: 7
Merit: 0
My seasonic 750w gold  kept up for the last week dual mining gpu-z showing 110w for each card. Now they are undervolted. You mean this is not good for longterm? Mind you i have 4 cards not 6.
Well ive been mining for a week with old settings dual mining gpu-z showing 110w for each card no problem. How will you explain this? They are even undervolted now.
gpuz power in not the same as wall power !
i've :
4*470 on ETH only, ~1075/1900 980mv , 750w at wall
4*470 on ETH+DCR, ~1150/2000 980mv, 850w at wall ( dcri 25 with v9.0, was 35 with v8.1 )

That's some high power consumption.

My 6 RX480 rig pulling 178mh/s solo mining ETH only pulls 810w from the wall Cheesy

1125/2050...... 880mv though Wink


When i set core voltage below 935mV like say 900.. The gpu still uses around 925-935mV.  How do you achieve 880? And when i go below 1100mhz it slightly affects hashing speed.
donator
Activity: 1610
Merit: 1325
Miners developer
- A lot of minor improvements and bug fixes.
Hi there Smiley

It looks like there are still some minor issues with your implementation of NiceHash version of Ethereum stratum protocol. There is no eth_SubmitHashRate method in the NiceHash specification, but your program sends this request few times a minute. As the consequence, a lot of "Invalid stratum message" errors are being recorded in server's log file.


You can use "-erate 0" option for now. In next update miner won't send this message if "-esm 3" is specified.
hero member
Activity: 682
Merit: 500
My seasonic 750w gold  kept up for the last week dual mining gpu-z showing 110w for each card. Now they are undervolted. You mean this is not good for longterm? Mind you i have 4 cards not 6.
Well ive been mining for a week with old settings dual mining gpu-z showing 110w for each card no problem. How will you explain this? They are even undervolted now.
gpuz power in not the same as wall power !
i've :
4*470 on ETH only, ~1075/1900 980mv , 750w at wall
4*470 on ETH+DCR, ~1150/2000 980mv, 850w at wall ( dcri 25 with v9.0, was 35 with v8.1 )

That's some high power consumption.

My 6 RX480 rig pulling 178mh/s solo mining ETH only pulls 810w from the wall Cheesy

1125/2050...... 880mv though Wink
sr. member
Activity: 441
Merit: 252
I have two rigs with claymore, same gpus. One on v8.1 and the other just upgrade yesterday to v9

v9:
ETH - Total Speed: 174.042 Mh/s, Total Shares: 103, Rejected: 0, Time: 00:49
ETH: GPU0 29.389 Mh/s, GPU1 29.708 Mh/s, GPU2 27.896 Mh/s, GPU3 29.097 Mh/s, GPU4 29.042 Mh/s, GPU5 28.911 Mh/s
ETH - Total Speed: 159.680 Mh/s, Total Shares: 103, Rejected: 0, Time: 00:50
ETH: GPU0 21.503 Mh/s, GPU1 29.725 Mh/s, GPU2 21.440 Mh/s, GPU3 29.055 Mh/s, GPU4 29.038 Mh/s, GPU5 28.919 Mh/s
ETH - Total Speed: 174.459 Mh/s, Total Shares: 103, Rejected: 0, Time: 00:50
ETH: GPU0 29.717 Mh/s, GPU1 29.702 Mh/s, GPU2 27.966 Mh/s, GPU3 29.059 Mh/s, GPU4 29.055 Mh/s, GPU5 28.960 Mh/s
ETH - Total Speed: 166.282 Mh/s, Total Shares: 105, Rejected: 0, Time: 00:51
ETH: GPU0 29.676 Mh/s, GPU1 21.482 Mh/s, GPU2 28.008 Mh/s, GPU3 29.090 Mh/s, GPU4 29.085 Mh/s, GPU5 28.941 Mh/s
ETH - Total Speed: 163.535 Mh/s, Total Shares: 110, Rejected: 0, Time: 00:54
ETH: GPU0 29.680 Mh/s, GPU1 21.475 Mh/s, GPU2 27.965 Mh/s, GPU3 26.511 Mh/s, GPU4 29.080 Mh/s, GPU5 28.823 Mh/s


Look how fluctuate, on v8.1 they are pretty stable only a few decimals of fluctuations. I am dual mining on linux -dcri 30. Is it a normal behaivor for v9?

I believe that when card submits share the speed shows reduced temporarily immediately after.  That's why you get fluctuations.
legendary
Activity: 3108
Merit: 1359
- A lot of minor improvements and bug fixes.
Hi there Smiley

It looks like there are still some minor issues with your implementation of NiceHash version of Ethereum stratum protocol. There is no eth_SubmitHashRate method in the NiceHash specification, but your program sends this request few times a minute. As the consequence, a lot of "Invalid stratum message" errors are being recorded in server's log file.
full member
Activity: 254
Merit: 100
On 480 i got best value for SOLO dcri 7 , and for dual Mining Dcri 25
member
Activity: 79
Merit: 10
All of my DCR shares are getting rejected using 9.0. I mine on supernova (dcr.suprnova.cc)...any idea why? It has been working fine with 8.1.

Try setting a low DCRI like 5 or 10 to start. Then use the + key to raise the dcr gradually.

Or let it disconnect you and reconnect after 20s. It seems to run ok after that.

You'll have a lot of rejected shares initially but when it reconnects it will start submitting shares properly.


Thanks guys...got it working. The DCRI thing worked.
newbie
Activity: 41
Merit: 0
I have two rigs with claymore, same gpus. One on v8.1 and the other just upgrade yesterday to v9

v9:
ETH - Total Speed: 174.042 Mh/s, Total Shares: 103, Rejected: 0, Time: 00:49
ETH: GPU0 29.389 Mh/s, GPU1 29.708 Mh/s, GPU2 27.896 Mh/s, GPU3 29.097 Mh/s, GPU4 29.042 Mh/s, GPU5 28.911 Mh/s
ETH - Total Speed: 159.680 Mh/s, Total Shares: 103, Rejected: 0, Time: 00:50
ETH: GPU0 21.503 Mh/s, GPU1 29.725 Mh/s, GPU2 21.440 Mh/s, GPU3 29.055 Mh/s, GPU4 29.038 Mh/s, GPU5 28.919 Mh/s
ETH - Total Speed: 174.459 Mh/s, Total Shares: 103, Rejected: 0, Time: 00:50
ETH: GPU0 29.717 Mh/s, GPU1 29.702 Mh/s, GPU2 27.966 Mh/s, GPU3 29.059 Mh/s, GPU4 29.055 Mh/s, GPU5 28.960 Mh/s
ETH - Total Speed: 166.282 Mh/s, Total Shares: 105, Rejected: 0, Time: 00:51
ETH: GPU0 29.676 Mh/s, GPU1 21.482 Mh/s, GPU2 28.008 Mh/s, GPU3 29.090 Mh/s, GPU4 29.085 Mh/s, GPU5 28.941 Mh/s
ETH - Total Speed: 163.535 Mh/s, Total Shares: 110, Rejected: 0, Time: 00:54
ETH: GPU0 29.680 Mh/s, GPU1 21.475 Mh/s, GPU2 27.965 Mh/s, GPU3 26.511 Mh/s, GPU4 29.080 Mh/s, GPU5 28.823 Mh/s


Look how fluctuate, on v8.1 they are pretty stable only a few decimals of fluctuations. I am dual mining on linux -dcri 30. Is it a normal behaivor for v9?
sr. member
Activity: 661
Merit: 250
My seasonic 750w gold  kept up for the last week dual mining gpu-z showing 110w for each card. Now they are undervolted. You mean this is not good for longterm? Mind you i have 4 cards not 6.
Well ive been mining for a week with old settings dual mining gpu-z showing 110w for each card no problem. How will you explain this? They are even undervolted now.
gpuz power in not the same as wall power !
i've :
4*470 on ETH only, ~1075/1900 980mv , 750w at wall
4*470 on ETH+DCR, ~1150/2000 980mv, 850w at wall ( dcri 25 with v9.0, was 35 with v8.1 )
member
Activity: 129
Merit: 10
Hmm, i tested newest clyamore 9 on my RX470 and i see performance decreese from 171 to 167 MH Sad
I will do more tests.

OLD CLAYMORE 8.1



NEW CLAYMORE 9


What do you think guys about this ? Is it only numbers and real speed is the same or is it worse speed ?
Worse speed due to driver. I already told him about the diff 470/480 in linux/windows, while 470 appears to perform better on linux, 480 does better on windows, that was only seen in his miner.
Sgminer and Zawawa's GG did the same for 470s. That being said I think my 470s are now losing ~1MH and power usage is normal again.

MSI Armor 470 4G Samsung memory: GPU3 28.685 Mh/s at 1126/2050 cl v9.0, GPU3 29.462 Mh/s same clocks v8.1, 28.7 GG v1.0-pre6 with asm enabled.

All my rigs are ~0.75mh slower on new version, mining ETH only on Ubuntu 16.04 w/16.60 drivers.

Code:
Claymore 8.1 - RX 470 4G

Screen #0
^[0;36mETH - Total Speed: 29.419 Mh/s, Total Shares: 5, Rejected: 0, Time: 00:05
^[0;36mETH - Total Speed: 29.419 Mh/s, Total Shares: 5, Rejected: 0, Time: 00:05
Screen #1
^[0;36mETH - Total Speed: 29.427 Mh/s, Total Shares: 4, Rejected: 0, Time: 00:05
^[0;36mETH - Total Speed: 29.419 Mh/s, Total Shares: 4, Rejected: 0, Time: 00:05
Code:
Claymore 9.0 - RX 470 4G

Screen #0
^[0;36mETH - Total Speed: 28.690 Mh/s, Total Shares: 3, Rejected: 0, Time: 00:05
^[0;36mETH - Total Speed: 28.685 Mh/s, Total Shares: 3, Rejected: 0, Time: 00:05
Screen #1
^[0;36mETH - Total Speed: 28.681 Mh/s, Total Shares: 4, Rejected: 0, Time: 00:05
^[0;36mETH - Total Speed: 28.692 Mh/s, Total Shares: 4, Rejected: 0, Time: 00:05

full member
Activity: 219
Merit: 100
Claymore,

Will you be optimizing the Dual miner for the RX Vega once released to take advantage of the new architecture?

sr. member
Activity: 275
Merit: 258
Can't seem to get much over 24 MH/s on RX 480 Ref 8GB models when dual mining.. Changes ethi and dcri settings to no avail.  Anyone want to share their settings?  Thanks!
donator
Activity: 1610
Merit: 1325
Miners developer
All of my DCR shares are getting rejected using 9.0. I mine on supernova (dcr.suprnova.cc)...any idea why? It has been working fine with 8.1.

I believe this problem is related to this pool, sometimes it does not send response to auth message for some reason but still sends jobs, so miner mines but pool does not accept shares.
Miner will reconnect automatically after 10-15 invalid shares, usually the pool accepts auth message at second attempt.
sr. member
Activity: 441
Merit: 252
All of my DCR shares are getting rejected using 9.0. I mine on supernova (dcr.suprnova.cc)...any idea why? It has been working fine with 8.1.

Try setting a low DCRI like 5 or 10 to start. Then use the + key to raise the dcr gradually.

Or let it disconnect you and reconnect after 20s. It seems to run ok after that.

You'll have a lot of rejected shares initially but when it reconnects it will start submitting shares properly.
newbie
Activity: 7
Merit: 0
All of my DCR shares are getting rejected using 9.0. I mine on supernova (dcr.suprnova.cc)...any idea why? It has been working fine with 8.1.

Happened to me when i first tried 9.0.  It worked fine after second run. Also mining on dcr.supernova . Don't know what fixed it.
Try messing with dcri a bit.
member
Activity: 79
Merit: 10
All of my DCR shares are getting rejected using 9.0. I mine on supernova (dcr.suprnova.cc)...any idea why? It has been working fine with 8.1.
legendary
Activity: 1960
Merit: 1005
Hi guys,
I am getting around 164 mh/s on eth and around 4900 on drc
http://imgur.com/a/mDwd6
I am with 6 470 cards and the power consumption is around 1200 w, is this normal? I overclocked the cards to 1250/2050, -100 mv with the trixx tool.
Is there a way to reduce that consumption?
Thanks!
Jump to: