Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux)

[wpsd]

- added temperature/fan monitoring for Linux version for gpu-pro drivers. Only monitoring is supported currently, not management.

The temp monitor will work in linux right ? if it is overheat it will stop the gpu

[1715520864]

1715520864

[1715520864]

1715520864

[1715520864]

1715520864

[1715520864]

1715520864

[1715520864]

1715520864

[anotherkiwi]

Claymore my miner thread hangs frequently, and it wont restart. I mean when remote-connect to the machine and on the miner window press down arrow "miner thread hanged, restarting" message pops up, but untill I do so, the machine just stays in freeze mode for 4 - 6 hours causing me significant loss. I need to keep checking on machines almost regular. I have set the miner to restart every 60 minutes with -r 60 but nothing happens. I'm using Ver 7.1 of your miner on RX 470's, 16.9.2 drivers on Win 10.

Hi Claymore, I changed from V7.1 to 7.2 and used the new low intensity dag creation and it works great. I can overclock my undervolted 390's a bit more and it starts up ok!

But I have the same problem now with the program hanging with the "miner thread hanged, restarting" dialog box that requires clicking to proceed.

Edit: It doesn't happen every time there is a problem and the miner restarts, sometimes it restarts OK. OS is Windows 10, Eth/SC

The dialog box says...

"EthDcrMiner64.exe has stopped working" A problem caused the program to stop working correctly. Windows will close the program and notify you if a solution is available.   > Close program <

The last few lines in the log file were...

17:08:19:259   bc4   recv: 51
17:08:19:259   bc4   srv pck: 50
17:08:19:290   bc4   srv bs: 0
17:08:19:290   bc4   sent: 208
17:08:19:446   c2c   got 36 bytes
17:08:19:446   c2c   buf: {"id":4,"result":true,"error":null}

17:08:19:446   c2c   parse packet: 35
17:08:19:446   c2c    SC: Share accepted (390 ms)!

17:08:19:446   c2c   new buf size: 0

[fittsy]

Riser or.. power cable.. I have GPU error very often and.. finally I realized that plug was slightly melted...

Which generation riser/cable?

I saw that with the narrow first generation ones. But haven't seen it with the wider ones with 3 capacitors.

[Claymore]

Is there any way to specify difficulty in dual mining? like lbry static difficulty setting?
Thank you

Difficulty means share target? Pool sets it.

Command tar failed with return code 0 and error message
Like in previous version... What does it mean?

I copied History.txt from Windows machine to Linux and Linux changed its time for some reason to future. Have no idea why it happend since other files have correct timestamp and original History.txt has correct time. May be another Linux "feature". You have to wait a day until it work, or change time on your system, or disable time checking in "tar" somehow.

Claymore, is there any workaround for gtx 1080 (372.90 / win 10) and not getting 5 Mhs ?

Yes, Windows 10 Anniversary + latest drivers wddm 2.1.

Claymore my miner thread hangs frequently, and it wont restart. I mean when remote-connect to the machine and on the miner window press down arrow "miner thread hanged, restarting" message pops up, but untill I do so, the machine just stays in freeze mode for 4 - 6 hours causing me significant loss. I need to keep checking on machines almost regular. I have set the miner to restart every 60 minutes with -r 60 but nothing happens. I'm using Ver 7.1 of your miner on RX 470's, 16.9.2 drivers on Win 10.

I'll try to reproduce it, PM me the complete log file.

- added temperature/fan monitoring for Linux version for gpu-pro drivers. Only monitoring is supported currently, not management.

The temp monitor will work in linux right ? if it is overheat it will stop the gpu

Yes it must work in Linux+gpupro drivers. All features must work except fan management.

[agis6]

Hi Claymore,

Thanks for the great miner!

I have a number of rigs with 380x cards, one with a mix of 380 and 380x and a "new" one with one ASUS STRIX-RX480-O8G-GAMING and one ASUS DUAL-RX480-O4G. They all run Ubuntu 14.04.4, except for the new one that runs Ubuntu 16.04.1

The rigs with the 380x and the 380/380x cards run smoothly since day 1, but the "new" one has the following problems:
1. The two cards cannot work together. ethdcrminer64 just crashes at random intervals. The situation is even worse under Windows 10: the system just reboots... I don't even get a decent BSOD. My solution is to run (under Ubuntu) two instances of the miner, one for very card.
2. The hashing speed goes up and down. The RX480 with 4GB hashes between 15MH/s and 21.5MH/s and the RX480 with 8GB hashes between 17MH/s and 25MH/s
3. I get LOTS of system errors. By lots, I mean 100GB of kernel and syslog messages in a couple of hours. I've disabled syslog and set -dbg -1, but still... I'm pasting a few dmesg messages below

[  848.570829] amdgpu 0000:01:00.0: GPU fault detected: 147 0x05d88801
[  848.570830] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x081048BB
[  848.570831] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x04088001
[  848.570832] VM fault (0x01, vmid 2) at page 135284923, read from 'TC6' (0x54433600) (136)
[  848.570894] amdgpu 0000:04:00.0: GPU fault detected: 147 0x05f88801
[  848.570895] amdgpu 0000:04:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x081048BF
[  848.570896] amdgpu 0000:04:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x02088001
[  848.570898] VM fault (0x01, vmid 1) at page 135284927, read from 'TC6' (0x54433600) (136)
[  848.570932] amdgpu 0000:01:00.0: GPU fault detected: 147 0x05e80801
[  848.570933] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x081048BD
[  848.570934] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x04008001
[  848.570935] VM fault (0x01, vmid 2) at page 135284925, read from 'TC0' (0x54433000) (
[  848.570946] amdgpu 0000:01:00.0: GPU fault detected: 147 0x05d04801
[  848.570947] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x081048BA
[  848.570948] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x04048001
[  848.570949] VM fault (0x01, vmid 2) at page 135284922, read from 'TC4' (0x54433400) (72)

Any help on any of the problems will be greatly appreciated.

TIA,
Agis

[adaseb]

In the Remote Monitor

Under RUNNING TIME what does the number inside the parenthesis mean?

Like

64:00 (1)

[citronick]

In the Remote Monitor

Under RUNNING TIME what does the number inside the parenthesis mean?

Like

64:00 (1)

I think its the number of times miner was restarted due to GPU hang...

[agis6]

In the Remote Monitor

Under RUNNING TIME what does the number inside the parenthesis mean?

Like

64:00 (1)

It shows how many time the remote manager has reconnected to the miner, during the time the remote manager has been running.
Reconnects can be because of miner restarts, computer restarts, etc.

[jewelparty]

Claymore, is there any workaround for gtx 1080 (372.90 / win 10) and not getting 5 Mhs ?

Would say it's ok if yours was 1070, but it's 1080.
I'm on GTX 970 Ti and Win 10 Anniversary x64 and get 4 Mhs (which i don't think it's ok). Drivers don't change it, latest or 352.84.

[yslyung]

@claymore, eth + SC or eth+lbc is 2%fee...
eth +dcr 1%fee?
that's according to op on 1st page.  mind to clarify?

thx.

[Za1n]

Is it just me or does EthMan display 1 extra GPU under the "Total Working GPUs:" summary when displayed on the webpage?

[Claymore]

Claymore, is there any workaround for gtx 1080 (372.90 / win 10) and not getting 5 Mhs ?

Would say it's ok if yours was 1070, but it's 1080.
I'm on GTX 970 Ti and Win 10 Anniversary x64 and get 4 Mhs (which i don't think it's ok). Drivers don't change it, latest or 352.84.

For 9xx on Win10 you should use old drivers like 352.xx and miner cuda 6.5.
From OP:

Code:

NVIDIA DRIVERS:
9xx cards in Windows 7 x64: just use latest/recent drivers from Nvidia website (for example, 368.81). Note that latest 372.54 is slower than 368.81.
9xx cards in Windows 10 x64: you have to use old drivers (for example, 352.xx) and miner built for cuda6.5. 
10xx cards in Windows 7 x64: just use latest 372.54 drivers from Nvidia website.
10xx cards in Windows 10 x64: just use latest 372.54 drivers from Nvidia website, note that you must have Win10 Anniversary update.

@claymore, eth + SC or eth+lbc is 2%fee...
eth +dcr 1%fee?
that's according to op on 1st page.  mind to clarify?
thx.

From OP or Readme:

This miner is free-to-use, however, current developer fee is 1% for Ethereum-only mining mode (-mode 1) and 2% for Ethereum+Decred mining mode (-mode 0)

Well, if you think it is not clear, I will change it to:
current developer fee is 1% for Ethereum-only mining mode (-mode 1) and 2% for dual mining mode (-mode 0)

[cptfisher]

at the moment  claymore you should improve the code for libry coin... im speaking as nvidia user... not profitable to use your miner....

genoil + ccminer running at same time on windows  will produce approximately 13% more than your miner  and for free


3 x 1070 gtx @ 70 % power  will produce 36 mhas eth and 610 lbry coin ... and no fee involved (except the pool fee dwarf pool... suprnova has no fee if i remember.) 365 w from the wall ( approximaterly 330 w REAL considering psu efficency. 100w /gpu considering 30w for system)

24h tested stable


alternatively we can give more priority to eth  and produce with same settings  67 Eth and  270 lbry

[sealofsun]

How to overclock memory to 2275MHz? I tried and maximum was 2250MHz (via Asus GPU Tweak II) rx480

[Odissey]

Good day. Tell me what kind of error and whether it affects what the realties.


ETH: Received error: {"id":4,"result":null,"error":[21,"jobnotfound"]}

[Claymore]

at the moment  claymore you should improve the code for libry coin... im speaking as nvidia user... not profitable to use your miner....
genoil + ccminer running at same time on windows  will produce approximately 13% more than your miner  and for free
3 x 1070 gtx @ 70 % power  will produce 36 mhas eth and 610 lbry coin ... and no fee involved (except the pool fee dwarf pool... suprnova has no fee if i remember.) 365 w from the wall ( approximaterly 330 w REAL considering psu efficency. 100w /gpu considering 30w for system)
24h tested stable
alternatively we can give more priority to eth  and produce with same settings  67 Eth and  270 lbry

Ok, I'll check it. Can you send the command lines for both miners? If I run both Genoil+ccminer1.8.3 I get only 18.8MH/s ETH + 71MH/s LBC (single stock 1070). If I start ccminer1.8.3 only, I see 278MH/s LBC.

Good day. Tell me what kind of error and whether it affects what the realties.
ETH: Received error: {"id":4,"result":null,"error":[21,"jobnotfound"]}

Stale share response. Probably your pool does not accept stale shares, which is not good for ETH.

[cptfisher]

-U --cuda-extragpu-mem 4096 --cuda-streams 1 --cuda-block-size 64 --cuda-grid-size 4096

for genoil

for ccminer i use 1.8.2 with -i 23 no special parameter (but also not specifing intensity is giving me same results)

settings :

70 % power limit

+ 212 core

+ 756 mem

60% fan speed

Temp limit 92 degrees ( videocards at 63 degrees anyway so no problem) power approximately 350  - 370 w from the wall (3x 1070 gtx)

alterantively i could specify no param to genoil and get  70 mhs  to ethereum and  240 from ccminer  with same card specs.

raising power limit and core or mem does not yeld any significant improvement but raising the engery consumed. alternatively there is also possibility to run x11revo  at 11mhs and eth at 9 mhs OR  23 mhs Eth and 6 mhs x11 revo for each card . same settings. Pretty good considering that just x11rev gives 36 mh max on 3 cards ( so 12 mhs per card). power is lower . 320 w

also i wanted to ask why if i put 50 % power the cores do not turbo bost and hangs at  low speed _ they should stay at 1400 - 1500 according to oc but they chill at 1000 slowing greatly eht (happens both with u and genoil... with cc miner instead card does not sit but keep working at full speed)

else 1070 gtx would be best card for eth. 70w @ 34 mhs by solving this issue

[Claymore]

-U --cuda-extragpu-mem 4096 --cuda-streams 1 --cuda-block-size 64 --cuda-grid-size 4096
for genoil
for ccminer i use 1.8.2 with -i 23 no special parameter (but also not specifing intensity is giving me same results)

Ok, I'll check it soon, though in my tests I always use stock settings, no OC, no changing power limits, no undervolting, no modded bios etc, it is not important when compare two miners.

[Nippy-ua]

Claymore's Dual Ethereum+Decred AMD+NVIDIA GPU Miner.

- systems with up to 32 GPUs are supported now (with some minor limitations).

Thanks BRO
it remains to solve the problem with the drivers...

[maxmad_x]

in GTX 980
Dual mining ETH+DCR
ETH=19 mh/s
DCR=197 mh/s

if i only DCR mining = 1750 mh/s

INCREASE DECRED INTENSITY--

You can increase the intensity for your secondary coin with the "dcri" flag.  I have mine set at "dcri 75", more than twice the default value of "dcri 30".  This will more than double your hash rate for Decred (DCR), and have little effect on your Ethereum (ETH) hash rate.  The cards from nVidia are very good for this, my 980ti gets a hash rate over 500MH/s on SiaCoin (SC).       --scryptr

Tried different settings, double mining not profitable.
The main ethereum power strongly falls. On SIA twice

Hey Bro,
Depending upon your cards you will have to find the best setting of dcri flag. Start low (20 or something like that ) and use "+" key to slowly increase until the ether mining starts to fall more than 5%. If you want keep an eye on power consumption. In certain situations dual mining may not be profitable more as your power consumption increases / AC power needed to keep mining cool increases / more downtime because of throttling after some time of mining etc. So just find your Efficiency (max) on the curve and then use that setting thereafter.

Single:
ETH=19
DCR=1750

Dual:
ETH=17
DCR=210

It is better to get something one

I guess you can use whattomine.com to find the right combo for you. Thats the last advice