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

[Quatrier]

Hello all,
 
I'm starting to run 2 Titan X (Maxwell) with Claymore. Latest version, same for nVidia drivers and CUDA toolkit. I can never reach more than 7 Mh/s per card. It seems very low. I've switched the 2 cards in "P0" state instead of "P2" in nVidia settings. And I didn't find any "Optimise for Compute Performance" option by the way (supposed to speed up).
 
Here is my .bat file:

EthDcrMiner64.exe -epool eu1.ethermine.org:4444 -ewal 0xXXXXXXXXXX.XXXX -epsw x -ethi 1 -dpool stratum+tcp://dcr.suprnova.cc:3252 -mode 1

I had to add "-ethi 1" to avoid Windows freezing (it slows down the GPU load from 100% to 90% which is fine). Same bad benchmark without that "-ethi 1" switch anyway.

Config : Windows 7 64 bit / SSD / P9X79 Pro / Xeon E5 12 cores / No overclocking.
 
Any help appreciated.

Thank you.

[1714640551]

1714640551

[1714640551]

1714640551

[heavyarms1912]

Hello.
Need help.

i've got 8x1070 and 5x580.

When i start Claymore...

POOL/SOLO version
GPU #0: algorithm ASM
GPU #1: algorithm ASM
GPU #2: algorithm ASM
GPU #3: algorithm ASM
GPU #4: algorithm ASM

ANd the miner freezes and windows.

Tried to use 2x bat files one for AMD and one for NVIDIA (using -platform 1 and -platform 2 config)
All start but i get "bind failed with error: 10048 (port for remote management is busy, use different -mport value), next attempt in 20sec..." in one of the brands.

all i want is start all GPUs at the same time.
What could i do?

If i use only 3 rx 580 i've got no problem.

For this one you need to use -mport <somePortNo> in one bat and -mport <someOtherPortNo> in other bat. Say -mport 3333 and -mport 4444

[B00gieman]

Is that so Simple as that?.... Dam.... I Will try that... Thanks in advance

Edit:
@heavyarms1912

PERFECT. All working.

[Yohow]

Hi! I bought a power meter recently, I have a 6x Sapphire RX 480 Nitro Plus OC 8GB rig, while dual mining (ETH+DCR) I’m pulling 1070W from the wall. I undervolt the GPUs with MSI AB 4.4.0 final version to -110mV and -75mV for GPU0. Someone could post his power consumption for a similar rig? I think AB doesn’t apply changes to voltage. Thanks in advance!

Note: "-dcri" value for DCR is 24. GPU´s "mclock": between 2100-2140 and "cclock": 1150 for all (BIOS modded with PBE, copy 1750 to 2000)

I will share my experiences with you.  I am running 24 rigs with the same cards dual mining DCR and LBRY.  I use Claymore to set my control parameters and all my machines run 970W - 1030W and I am getting 170-175 Mh/s on each rig.  I run with cclock of 1160, mclock 2020 (doesn't seem to affect hashrate much but if set too high will result in cards going into open CL), cvddc of 870 and mvddc of 890.  If you want to get your power down, lower your mclock by 50 and then 100 and see if it affects your hash rate.  It most likely will not have a huge drop in performance, but you may see a reduction in power consumption.  Lowering the mvddc has a big impact on power reduction efforts as well, and you can play with that to understand power consumption over performance.  Also, running a powlim of -10 will get your power down as well with minimal reduction in performance.  I am using the uber straps mod'd bios and August 23rd block chain drivers.  I have to run ethi at 5 or below or the machines become unstable, and each machine has to be tweaked slightly to get it stable since no two machine act exactly the same.  Most of these machines run for weeks without issue and only get reset when I update the miner or change what coin I am mining.  I run DCR at 28 on the nitro+ 480s which seems to be the point that I get the max of both before it starts to slow down the ETH mining.  Hope that helps.

If I lower the "mclock" to 2100 or 2080 it affect the hashrate (i.e. 1150/2140 = 29.7 - 29.9 MH/s; but If I lower it to 1150/2100 = 29 - 29.2 MH/s) not too much but a little. I´m getting about 176-177 MH/s (very stable). I´m going to try "-dcri 28" but in the past I saw that any value above 25 becomes the rig unstable... I´m going to try "powlim" -10% too and observe, however I´d like to know how much shares I´d lose if I lower the hashrate to 173 MH/s for example, it would be worth it? Many thanks for help!

Hi! I bought a power meter recently, I have a 6x Sapphire RX 480 Nitro Plus OC 8GB rig, while dual mining (ETH+DCR) I’m pulling 1070W from the wall. I undervolt the GPUs with MSI AB 4.4.0 final version to -110mV and -75mV for GPU0. Someone could post his power consumption for a similar rig? I think AB doesn’t apply changes to voltage. Thanks in advance!

Note: "-dcri" value for DCR is 24. GPU´s "mclock": between 2100-2140 and "cclock": 1150 for all (BIOS modded with PBE, copy 1750 to 2000)

That's about right with what i'm pulling with 6x of my Asus Dual OC 580s dual-mining eth and dcr as well, at -dcri 25 hashing 178 Mh eth and 4500 Mh dcr.

Okay, then I'm calm, my consumption is normal

[crd007]

Is it possible to fix the difficulty on some eth pools using the miner software?
When mining cryptonight currencies you can do it like: YOUR_ADRESS.DIFFICULTY
Is this somehow also possible on eth pools?

[MadSciencist]

Hi.
I just added second GPU to my PC.
Sapphire RX480 8GB was the first one, second is XFX RX580 8GB. Both with Samsung memory.
After adding second GPU hashrate dropped significantly from Sapphire. I was able to get (stable) ~29-30.5Mh/s (with a little OCed vram).
Both cards has ~110W power usage (GPU-Z)
Now it's something like:
https://i.imgur.com/M0MT0Lq.png
PS - no, it's not a problem of throttling, it keeps stable 1150MHz.
PS2 - yes, crossfire is disabled
PS3 - yes, both cards are working in 'compute' mode.
PS4 - increasing 'power limit' does literally nothing.

Additional args:
-dcri 30,33 -ethi 8,16

My specs:
Asus Sabertooth 990FX 2.0
FX8350 (tried underclocking/volting it in case of too weak PSU)
PSU - Chieftec 750W
Newest AMD drivers (tried older too, but even newest worked fine before adding 2nd GPU).
Is it normal?

My friend has exactly the same setup (with exactly the same PSU), but he has 2x RX580 (Sapphire with hynix memory and XFX with samsung memory). Everything works fine.

[nathan_tek]

Which is the best driver to use with fall creators update?? I am using Rx 480/580 cards...

[SpartaNLD]

Which is the best driver to use with fall creators update?? I am using Rx 480/580 cards...

blockchain driver

[Ursul0]

Check your risers they be burned.

are you sure that "burned raiser" is not an urban legend?
bad molex contacts is something I encounter very often, but actually burned raiser is something that I haven't seen yet, but then maybe I'm just that good

[NewbieMiner17]

Check your risers they be burned.

are you sure that "burned raiser" is not an urban legend?
bad molex contacts is something I encounter very often, but actually burned raiser is something that I haven't seen yet, but then maybe I'm just that good

Yeah I had a burned molex (peripheral cables)... It actually burned out on the PSU itself and damaged one of my outputs. Have just ordered a new cable from corsair and used a different output... What could have caused this? Can I do anything to prevent it? Could a claim be made vs corsair for warranty on my 1200w PSU?

[P00P135]

Check your risers they be burned.

are you sure that "burned raiser" is not an urban legend?
bad molex contacts is something I encounter very often, but actually burned raiser is something that I haven't seen yet, but then maybe I'm just that good

Yeah I had a burned molex (peripheral cables)... It actually burned out on the PSU itself and damaged one of my outputs. Have just ordered a new cable from corsair and used a different output... What could have caused this? Can I do anything to prevent it? Could a claim be made vs corsair for warranty on my 1200w PSU?

How many things did you have plugged into your molex line?  Did you use a corsair cable or a different brand?

[NewbieMiner17]

Check your risers they be burned.

are you sure that "burned raiser" is not an urban legend?
bad molex contacts is something I encounter very often, but actually burned raiser is something that I haven't seen yet, but then maybe I'm just that good

Yeah I had a burned molex (peripheral cables)... It actually burned out on the PSU itself and damaged one of my outputs. Have just ordered a new cable from corsair and used a different output... What could have caused this? Can I do anything to prevent it? Could a claim be made vs corsair for warranty on my 1200w PSU?

How many things did you have plugged into your molex line?  Did you use a corsair cable or a different brand?

3 RX480's N+ 4GB on each with the final socket unused. Corsair that came with it

[heavyarms1912]

Check your risers they be burned.

are you sure that "burned raiser" is not an urban legend?
bad molex contacts is something I encounter very often, but actually burned raiser is something that I haven't seen yet, but then maybe I'm just that good

Yeah I had a burned molex (peripheral cables)... It actually burned out on the PSU itself and damaged one of my outputs. Have just ordered a new cable from corsair and used a different output... What could have caused this? Can I do anything to prevent it? Could a claim be made vs corsair for warranty on my 1200w PSU?

Are you using SATA to 4-pin/6-pin power adapters?  If so then most likely that's the reason.  Most of the bundled SATA to 6-pin/4-pin converters are of cr@p quality. 
They are not crimped but are molded.  At high temperatures the molds basically starts to melt and can short the wires. 

Avoid any SATA -> xyz power adapters.  Use molex splitters.  They might look flimsy but are constructed better than any other molded SATA connectors. 
Also avoid loading more than higher number of gpus (using splitters) on single molex cable.

p.s. The SATA cables that come bundled with PSU's are crimped and not molded so they're good. 

[P00P135]

Check your risers they be burned.

are you sure that "burned raiser" is not an urban legend?
bad molex contacts is something I encounter very often, but actually burned raiser is something that I haven't seen yet, but then maybe I'm just that good

Yeah I had a burned molex (peripheral cables)... It actually burned out on the PSU itself and damaged one of my outputs. Have just ordered a new cable from corsair and used a different output... What could have caused this? Can I do anything to prevent it? Could a claim be made vs corsair for warranty on my 1200w PSU?

How many things did you have plugged into your molex line?  Did you use a corsair cable or a different brand?

3 RX480's N+ 4GB on each with the final socket unused. Corsair that came with it

That was your problem, you can only put a maximum of 2 on a single molex cable or you risk burning.

[wobblewoo]

Hello all

I'm having a nightmare with Claymore. Can anyone help?
So ive 4x AMD RX480s (Crimson 16.9.2) and 2x Nvidea 1070s running on Windows 10. I've removed the overclocking if that matters

Im using Claymore 10.2:
EthDcrMiner64.exe -epool eth-eu1.nanopool.org:9999 -ewal <wallet address>.<worker>/<email> -epsw x -mode 1 -ftime 10

It works fine when benchmarking
EthDcrMiner64.exe -benchmark 1

But when i try and mine, i get:
GPU4, GpuMiner cu_k1 failed4, unspecified launch failure
GPU5, GpuMiner cu_k1 failed4, unspecified launch failure
GPU4, GpuMiner kx failed 1
GPU5, GpuMiner kx failed 1
GPU4 failed
GPU5 failed
... repeats above...
WATCHDOG: GPU error, you need to restart miner

Sometimes is 3 and 4, the numbers seem to change.

Can anyone help please

W

[Call_Me_Bambi]

Check your risers they be burned.

are you sure that "burned raiser" is not an urban legend?
bad molex contacts is something I encounter very often, but actually burned raiser is something that I haven't seen yet, but then maybe I'm just that good

"If you're not failing, you're not trying hard enough."

Here is a pic of one of my failed risers, I usually caught them before they went this far, but this is an extreme example.

[fredeq]

Check your risers they be burned.

are you sure that "burned raiser" is not an urban legend?
bad molex contacts is something I encounter very often, but actually burned raiser is something that I haven't seen yet, but then maybe I'm just that good

"If you're not failing, you're not trying hard enough."

Here is a pic of one of my failed risers, I usually caught them before they went this far, but this is an extreme example.

Thats a molex burn, exactly what the guy pointed out - that molexes burn, not risers.

[Blocklock]

Hello!
This program will work on my computer with features 5gb ram, 2gb graphics card? 

[Experimenter]

Claymore,

I have a minor bug report (or maybe an improvement suggestion) for the Web interface of EthMan.
The "More Details" button currently links to "http://<IP>:PORT/detail".
Can you please change it to just "/detail", omitting the <IP>:PORT ?
This will allow to maintain proper URL in case if special routing / DNS name is used to access the web page.

Thanks!

[deadsix]

Hello!
This program will work on my computer with features 5gb ram, 2gb graphics card? 

No.