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

[VTdude12]

Thanks to everyone who responded to my post!

To this...

reported hashrate depends on the valid shares you submit.

You can have lower or higher than the calculated one.

You should atleast wait 24 hours and see how much they are aligned.

If there's a -10% difference it's likely that you are having memory errors.

Are you saying maybe overclocked too much / under volt?  Or true hardware issue?  Didn't think this would be the case if claymore was able to run stably for 24/7 

[QuadraQ]

So I recently added two GTX 1060 cards to my mining rig to go with my trusty AMD 7970. I have the latest AMD and CUDA drivers installed, specifically AMD 17.6.1 and Nvidia CUDA 8.0.61. I'm mining in dual mode with ETH and DCR, and the 1060's are brand new and not overclocked (at least not yet).

So everything was going fine until about 5:20 am last night. Looking through the log files here are the relevant bits if info:

Code:

05:20:50:355	161c	NVML: cannot get current temperature, error 15
05:20:50:359	161c	NVML: cannot get fan speed, error 15
...
05:20:53:253	734	GPU 2, GpuMiner cu_k00 failed 30, unknown error
05:20:53:312	1748	GPU 1, GpuMiner cu_k00 failed 30, unknown error
05:20:53:465	1748	GPU 1, GpuMiner kx failed 1
05:20:53:325	1624	parse packet: 234
05:20:53:484	734	GPU 2, GpuMiner kx failed 1
05:20:53:477	1748	Set global fail flag, failed GPU1
05:20:53:483	1624	ETH: job is the same
05:20:53:499	734	Set global fail flag, failed GPU2
05:20:53:540	1748	GPU 1 failed
05:20:53:563	d04	GPU 1, GpuMiner cu_k00 failed 30, unknown error
05:20:53:592	d04	GPU 1, GpuMiner kx failed 1
05:20:53:604	734	GPU 2 failed
05:20:53:616	1188	GPU 2, GpuMiner cu_k00 failed 30, unknown error
05:20:53:532	1624	new buf size: 0
05:20:53:630	161c	NVML: cannot get current temperature, error 999 (an internal driver error occurred)
...
05:20:56:809	161c	NVML: cannot get current temperature, error 999 (an internal driver error occurred)
05:20:56:813	161c	NVML: cannot get fan speed, error 999 (an internal driver error occurred)
05:20:56:872	161c	NVML: cannot get current temperature, error 999 (an internal driver error occurred)
05:20:56:876	161c	NVML: cannot get fan speed, error 999 (an internal driver error occurred)
...
05:21:19:101	161c	WATCHDOG: GPU error, you need to restart miner :( 
...
05:21:20:339	161c	Restarting OK, exit...

When the next log file begins after the restart, one of the GTX 1060 cards isn't even recognized:

Code:

05:21:23:634	101c	OpenCL platform: AMD Accelerated Parallel Processing
05:21:23:638	101c	OpenCL platform: NVIDIA CUDA
05:21:23:644	101c	OpenCL initializing...
05:21:23:649	101c	AMD Cards available: 1 
05:21:23:654	101c	GPU #0: Tahiti, 3072 MB available, 28 compute units
...
05:21:24:269	101c	CUDA initializing...
05:21:24:273	101c	NVIDIA Cards available: 1 
05:21:24:277	101c	CUDA Driver Version/Runtime Version: 8.0/8.0
05:21:24:286	101c	GPU #1: GeForce GTX 1060 6GB, 6144 MB available, 10 compute units, capability: 6.1
05:21:24:291	101c	Total cards: 2 
05:21:33:321	101c	NVML version: 8.376.51

I noticed that when I ran GPU-Z to check on the driver versions, it recognized all three cards BUT one of the 1060's didn't show a checkmark for CUDA anymore! After I restarted and checked GPU-Z again, both cards showed checkmarks for CUDA again. Weird.

These are the driver versions according to GPU-Z

Code:

22.19.171.1 (Crimson 17.6.1) Beta
21.21.13.7651 (ForceWare 376.51) Beta

I'm using Windows 8.1 Pro 64-bit with all the latest updates, and Claymore's Dual miner v9.5.

I don't think the cards overheated since just before the errors above the miner reported temperatures that should be well within specifications:

Code:

05:19:46:774	161c	GPU0 t=80C fan=67%; GPU1 t=82C fan=75%, GPU2 t=79C fan=70%
05:19:46:778	161c	em hbt: 250, dm hbt: 0, fm hbt: 62, 
05:19:46:783	161c	watchdog - thread 0 (gpu0), hb time 266
05:19:46:788	161c	watchdog - thread 1 (gpu0), hb time 125
05:19:46:793	161c	watchdog - thread 2 (gpu1), hb time 63
05:19:46:798	161c	watchdog - thread 3 (gpu1), hb time 15
05:19:46:803	161c	watchdog - thread 4 (gpu2), hb time 46
05:19:46:808	161c	watchdog - thread 5 (gpu2), hb time 0
...
05:19:50:112	1624	ETH - Total Speed: 48.685 Mh/s, Total Shares: 154, Rejected: 0, Time: 04:02
05:19:50:116	1624	ETH: GPU0 12.885 Mh/s, GPU1 17.799 Mh/s, GPU2 18.001 Mh/s
05:19:50:120	1624	 DCR - Total Speed: 1737.263 Mh/s, Total Shares: 645, Rejected: 0
05:19:50:124	1624	 DCR: GPU0 901.935 Mh/s, GPU1 415.303 Mh/s, GPU2 420.025 Mh/s

I'm stumped. Obviously stability is king for any mining rig, and I'm concerned this may keep happening but I don't know what to do to address it. Any ideas?

[Claymore]

According to what I observed, it may be legitimate misreporting of Eth hashrate whilst dual-mining.

Probably you mean that different GPU rounds take different time in your tests, am I correct? There are two reasons for it at least:
1. You started two separate miners, both try to load GPU completely so different GPU rounds will have different time. My miner uses different approach. You can press "s" key to see hashrate for latest GPU round, press it many times and see if there are fluctuations, for Polaris card you should notice them but they are related to item #2 below.
2. At some circumstances GPU can be slower for a second or so, then work a few seconds at 100%, and so on. Especially Polaris cards. It was a problem in v9.0 but I've almost fixed it in recent versions, these fluctuations are low and rare now even on Polaris. v8 is free of this issue, but has lower speed for second coin.
Anyway, may be I should also show additional counter that measures hashrate for a minute when "s" is pressed, I will try to add it in next update.

[mmrj9]

Hello guys, can anyone post here a start.bat file ready for Eth + Pascal dual mining pls   

[coinpops]

Hello, I'm having an issue with Claymore not detecting my GPUs.

Claymore has been working great with my two RX 480s in the PCIe 16x slots in my motherboard. Since I want to see if I can use more GPUs, I bought two 1x->16x PCIe risers that are now plugged in to my mobo's PCIe 1x slots with my two 480s. They seem to be working fine in Windows -- they're shown in the device manager and in Radeon Settings. But when I start Claymore, it's saying that no GPUs were found.

Any ideas?

[QuadraQ]

Hello, I'm having an issue with Claymore not detecting my GPUs.

Claymore has been working great with my two RX 480s in the PCIe 16x slots in my motherboard. Since I want to see if I can use more GPUs, I bought two 1x->16x PCIe risers that are now plugged in to my mobo's PCIe 1x slots with my two 480s. They seem to be working fine in Windows -- they're shown in the device manager and in Radeon Settings. But when I start Claymore, it's saying that no GPUs were found.

Any ideas?

Did you connect power to the riser cards?

[coinpops]

Did you connect power to the riser cards?

Yes.

[coinpops]

Not sure why, but reinstalling the drivers seems to have fixed it

[cprieto]

Hello,

Have you guys tried Dual mining with GTX 1070 cards? hows your hash? its decent compared to rx 480/580 series

This is my hash dual mining on rx 480

ETH - Total Speed: 29.143 Mh/s  SC - Total Speed: 582.858 Mh/s

I tried to find some information, i found a guy on youtube hashing 200-300 when dual mining.. its normal?

Regards

[bitcoinmining123456]

Can we mining ethrium after November month?? If is it true, then what about others coin to mining?? Can we mining 3/4 years. ??

[martyroz]

Have you guys tried Dual mining with GTX 1070 cards? hows your hash? its decent compared to rx 480/580 series

This is my hash dual mining on rx 480

ETH - Total Speed: 29.143 Mh/s  SC - Total Speed: 582.858 Mh/s

I tried to find some information, i found a guy on youtube hashing 200-300 when dual mining.. its normal?

This is what I'm getting on my GTX1070 - 29.5MH/s ETH and 280-300MH/s SC
Not sure if that's as good as it gets.

Swapping to ZEN soon.

[RyanM923]

I'm having an issue with the hash rate of my 970sc cards.  They're both only getting 3.15Mh a piece, and I am about exhausted of ideas as to why.  I have tried installing the 347.52 drivers, and when I install the cuda 8.0 (while doing a custom install to de-select drivers that over-ride the 347.52 ones), the miner will not run for me (says that no gpu's are found).  With the latest updated nvidia drivers I get the slow speed.  I am running win10...has anyone has similar issues with a 970?

Also, I tried using the v7.3 miner and get the same hash rate as with the v9.5 one.

[myty]

Hello.
I am using R9 380 Nitro Sapphire.
When i am mining solo Ethereum all is working fine, but when i dual mine with DECRED I get GPU artefacts.
Why is that?
Thank you!

[Piradoxlanieve]

I read that in dual mining the graphics cards do not usually work at the same speed, it is normal that some work at 29mhs and another at 25mhs, normally the gpu that runs at 25 when they spend a few hours usually drops to 0 and a warning comes out The controller stopped working and my computer freezes or reboots, I had the voltage at -96mv now I'm testing at -66 mv I have bios mod, but I do not know if it's normal at all that 3 run at 29mbs and the rest do not, I feel my English bad.

Code:

ETH: 06/12/17-09:07:55 - New job from eu1.ethermine.org:4444
ETH - Total Speed: 167.161 Mh/s, Total Shares: 20, Rejected: 0, Time: 00:08
ETH: GPU0 29.321 Mh/s, GPU1 29.370 Mh/s, GPU2 26.522 Mh/s, GPU3 25.219 Mh/s, GPU4 27.659 Mh/s, GPU5 29.070 Mh/s
 SC - Total Speed: 3008.894 Mh/s, Total Shares: 9, Rejected: 0
 SC: GPU0 527.774 Mh/s, GPU1 528.661 Mh/s, GPU2 477.393 Mh/s, GPU3 453.944 Mh/s, GPU4 497.860 Mh/s, GPU5 523.262 Mh/s
GPU0 t=59C fan=74%, GPU1 t=58C fan=73%, GPU2 t=59C fan=71%, GPU3 t=60C fan=71%, GPU4 t=55C fan=70%, GPU5 t=67C fan=77%
ETH: 06/12/17-09:08:40 - New job from eu1.ethermine.org:4444
ETH - Total Speed: 167.523 Mh/s, Total Shares: 20, Rejected: 0, Time: 00:09
ETH: GPU0 29.347 Mh/s, GPU1 29.319 Mh/s, GPU2 26.909 Mh/s, GPU3 25.158 Mh/s, GPU4 27.705 Mh/s, GPU5 29.084 Mh/s
 SC - Total Speed: 3015.405 Mh/s, Total Shares: 9, Rejected: 0
 SC: GPU0 528.240 Mh/s, GPU1 527.748 Mh/s, GPU2 484.365 Mh/s, GPU3 452.846 Mh/s, GPU4 498.690 Mh/s, GPU5 523.516 Mh/s
GPU0 t=60C fan=74%, GPU1 t=58C fan=73%, GPU2 t=59C fan=71%, GPU3 t=60C fan=71%, GPU4 t=55C fan=71%, GPU5 t=67C fan=78%
 SC: 06/12/17-09:08:54 - New job from sia-eu1.nanopool.org:7777
ETH: 06/12/17-09:09:04 - SHARE FOUND - (GPU 2)
ETH: Share accepted (62 ms)!
ETH: 06/12/17-09:09:12 - SHARE FOUND - (GPU 4)
ETH: Share accepted (63 ms)!

[ilv]

I'm having an issue with the hash rate of my 970sc cards.  They're both only getting 3.15Mh a piece, and I am about exhausted of ideas as to why.  I have tried installing the 347.52 drivers, and when I install the cuda 8.0 (while doing a custom install to de-select drivers that over-ride the 347.52 ones), the miner will not run for me (says that no gpu's are found).  With the latest updated nvidia drivers I get the slow speed.  I am running win10...has anyone has similar issues with a 970?

Also, I tried using the v7.3 miner and get the same hash rate as with the v9.5 one.

I'm running a GTX 970 and in dual mode I'm getting about 3.2 Mh/s on ETH and 54 Mh/s on SC.

I've tried changing the nVidia driver to the newest as well as the recommended 368.81, hash rates stay the same.

I've disabled all the power options I've found on the nvidia panel. My GTX is running at around 1417mhz. GPU temps never get much above 70C, and tweaking the GPU CLOCK OFFSET/MEM CLOCK OFFSET change the hash rates slightly but eventually lead to the box halting. I tried copying the CUDA dll form one of the child directories on the Claymore 9.5 zip, the execution reported that CUDA 8/7.5 were available. This also didn't change the hash rate at all.

I've tried editing my start.bat to include the following settings:

Code:

setx GPU_FORCE_64BIT_PTR 0
setx GPU_MAX_HEAP_SIZE 100
setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_SINGLE_ALLOC_PERCENT 1

Nothing seems to have made a difference.  I honestly feel like some video games tax the GPU more than this... the GPU fans spin up more in some games.

[XManiac]

I'm running a GTX 970 and in dual mode I'm getting about 3.2 Mh/s on ETH and 54 Mh/s on SC.

I had 18mh/s in ETH only mode on my 970.

[ilv]

I'm running a GTX 970 and in dual mode I'm getting about 3.2 Mh/s on ETH and 54 Mh/s on SC.

I had 18mh/s in ETH only mode on my 970.

Great for you! What am I, and the person I quoted doing wrong then? I don't think the question is, "are these numbers right" I think the question is, "How do we get the right numbers from this card and config"

Thanks!

[XManiac]

Great for you! What am I, and the person I quoted doing wrong then? I don't think the question is, "are these numbers right" I think the question is, "How do we get the right numbers from this card and config"

Thanks!

I just wrote it for comparison. How i get 18 on 970? Download miner, edit *.bat and launch. I have no idea why some ppl have less than 5mh on the 970... Sry.

[drak24]

Hello,

Have you guys tried Dual mining with GTX 1070 cards? hows your hash? its decent compared to rx 480/580 series

This is my hash dual mining on rx 480

ETH - Total Speed: 29.143 Mh/s  SC - Total Speed: 582.858 Mh/s

I tried to find some information, i found a guy on youtube hashing 200-300 when dual mining.. its normal?

Regards

What are your settings to get 29? I have 2 rx 480s in my system dual mining the same coins. I get 24 on both cards for ETH and 1200+ for SC using dcri 22. If it's higher, I get a higher SC hash rate but slightly lower ETH hash rate. If I lower it under 22, ETH stays the same at 24/card, but SC gets lower. I'm running both cards at 0 mV (I keep it there to prevent crashes), 0 power limit, 90 temp limit, 1154 core clock, 2228 memory clock (I notice no difference increasing this to 2250), custom fan speed.

Any suggestions would be appreciated.

[davescream]

Claymore,
the problem with ETHMan 3.1 + Dualminer 9.5
If some rig go down, and then after some time goes up, it dosent restore in ETHMan, on screenshot you see that miner restarted and work, i launched web console and it works already 2 minutes, but ETHMan still show "lost ****". Even after 10 minutes ETHMan dosent see rig. So as it was with ETHMan 3.0, I need restart it to see all rigs stats again.

If you included cryptography and https, then maybe here some trouble with ssl? need to update certificate if connection was lost?

https://pp.userapi.com/c637223/v637223838/6dcbe/XV9iyhdO-c0.jpg