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

[Analyst2016]

Hello,

This regards the following error:

"OpenCL error - 4 - cannot create DAG on GPU"
https://drive.google.com/open?id=0B6HNx74FwQi2VXJDZjBQbjQ0QWs

4 cards mine and one doesn't. If I change the cards or PCI slots, the problem will move from GPU to another. Where before it was GPU1, then it became GPU 4.
I am running windows 8.1 from a 64usb drive. When formatting the USB drive I formatted it the maximum amount of virtual memory, which was 59gb or so.

I can't imagine what is giving this issue. After 10 minutes or less this dual miner requested to shut down because it was out of memory also.

Any ideas for solutions please? I've noticed I'm not the only one who has had this issue but I could not find a resolution for them.

This is the .bat file I am running:
setx GPU_SINGLE_ALLOC_PERCENT 100
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_HEAP_SIZE 100
setx GPU_FORCE_64BIT_PTR 0
EthDcrMiner64.exe -epool us2.ethermine.org:4444 -ewal address -epsw x -eworker Worker01 -ethi 8 -mode 1

I tested if it is this is hardware related also. Disconnected a random card.
Then 4 cards mined without errors. Reconnected that card, and now disconnected another card.
This second lot of 4 cards also mined without any errors, hence eliminating the idea that this is a hardware issue.

Cheers

[1715383107]

1715383107

[1715383107]

1715383107

[1715383107]

1715383107

[Trimegistus]

The 390-8GB is slightly OCed, stock is 1015Mhz/1500Mhz, changing memory seems not giving any reduction in total watts.
With standard ethminer.exe, I can achieve only 26++MHs and qtminer.exe 27+MHs, but with CDM 29.xxMHs.
So if my wattage comes under 1300w, with the underclock and still maintain CDM 29.xxMHs, I can now put another GPU to fully utilise the PSU with higher total hash of 177+ MHs per rig.
So underclocking has its benefits and the rig has been running for more than 2 weeks with CDM. Temp low 70s overall.
Upgrading the Catalyst to 16.3.2 has also no obvious effect, hash, temp and stability seems unchanged.
16-20GB virtual memory as per Claymore's guide applied too.

Thanks. I'm already testing it on my rigs.

Cheers!

PS: Claymore mentioned he was going to had some kind of power control feature. I wonder if we will get it in the next hours 

[arielbit]

It takes a lot of time to test solo mode, I put all my cards on it, but still not enough and I have to wait for many hours, network diff is too high. Anyway, even if I cannot find a block, I will release v4.2 in a few hours because it works with eth-proxy so it must work with geth as well.

I believe some members whom have asked for the solo feature have monster hash capability, so we would like to test it out and perhaps this way we can give you faster feedback.

for smaller hash rate miners, it is still advisable to use pools so they can monitor what's going on..

for me, patience for up to two days or maybe up to three days is the limit.

[Prelude]

It takes a lot of time to test solo mode, I put all my cards on it, but still not enough and I have to wait for many hours, network diff is too high. Anyway, even if I cannot find a block, I will release v4.2 in a few hours because it works with eth-proxy so it must work with geth as well.

Awesome! Can't wait to go solo with your miner.

[arielbit]

Hi Claymore

Could you please check 390x coreclock issue, when i mine with genoil miner clock is same as msi afterburner setting
but with your miner it drops around 950 with slight fluctuations.

tanx
 

we may have the same issue, but mine shows only when mining in dual mode..my temps (VRM) are up thread.

core clocks don't touch 80c except for my msi r9 390 sometimes during the heat of the day it reaches 83c..looking at hwinfo it is not the one down-throttling, it is my sapphire 390 and 390x that is down-throttling.

eth only mining does not down-throttle..so these 3 cards is on eth only mode due to loosing a lot of ETH hashes compared to gains in ETH-DCR dual mode.

[Trimegistus]

Is anyone having problems with Dwarf's pool?

My rigs are reporting accepted shares put the pool's website is claiming no shares are being sent 

WTF?

[arielbit]

Is anyone having problems with Dwarf's pool?

My rigs are reporting accepted shares put the pool's website is claiming no shares are being sent 

WTF?

me too, "last share 45mins ago"

[Trimegistus]

Is anyone having problems with Dwarf's pool?

My rigs are reporting accepted shares put the pool's website is claiming no shares are being sent 

WTF?

me too, "last share 45mins ago"

Ok, then it's not a problem with my stuff. It's on their side.

Thanks!

[spiete]

Help, shares didnt work on dwarfpool.

PC doing okay.


But last share was 2 hours ago



What to do?


Claymore's Dual Ethereum+Decred AMD GPU Miner v3.2 Beta - POOL

[tflux99]

try esm 1 for qtminer mode or does dwarf not support qtminer dont mine there anymore do to 51% hashrate hehe share your power dont let pools grow to much hehe

[Xdragon]

Help, shares didnt work on dwarfpool.

PC doing okay.

https://imgur.com/3kBc5jx

https://imgur.com/a/hH52y

But last share was 2 hours ago

http://dwarfpool.com/eth/address?wallet=0xA0F5578094e713c58B901eC43FB58bbbD7A63372

What to do?


Claymore's Dual Ethereum+Decred AMD GPU Miner v3.2 Beta - POOL

Dwarfpool currently has some problems, it is not due to mining software...

[Trimegistus]

Dwarf is back online 

[Claymore]

v4.2:

- added support for Ethereum solo mining.
- added "-gser" option.
- added "-eres" option.
- added "-powlim" option.
- added "-etht" option.
- "-tt 1" (default) now does not manage fans but shows GPU temperature and fan status.
- added support for "client.reconnect" stratum command for Decred.

[Osakiman]

-eres   this setting is related to Ethereum mining stability. Every next Ethereum epoch requires a bit more GPU memory, miner can crash during reallocating GPU buffer for new DAG.
   To avoid it, miner reserves a bit larger GPU buffer at startup, so it can process several epochs without buffer reallocation.
   This setting defines how many epochs miner must foresee when it reserves GPU buffer, i.e. how many epochs will be processed without buffer reallocation. Default value is 2.

If we use big eres, will that affect the mining speed? Like bigger DAG?

[melloyellow]

v4.2:

- added support for Ethereum solo mining.
- added "-gser" option.
- added "-eres" option.
- added "-powlim" option.
- added "-etht" option.
- "-tt 1" (default) now does not manage fans but shows GPU temperature and fan status.
- added support for "client.reconnect" stratum command for Decred.

so we can point this version to eth-proxy? 

[Claymore]

-eres   this setting is related to Ethereum mining stability. Every next Ethereum epoch requires a bit more GPU memory, miner can crash during reallocating GPU buffer for new DAG.
   To avoid it, miner reserves a bit larger GPU buffer at startup, so it can process several epochs without buffer reallocation.
   This setting defines how many epochs miner must foresee when it reserves GPU buffer, i.e. how many epochs will be processed without buffer reallocation. Default value is 2.

If we use big eres, will that affect the mining speed? Like bigger DAG?

-eres does not change mining speed at all. It just changes the buffer size. So if you don't have any problems with GPUs you can set "-eres 5" to make miner use the same buffer for five epochs (25 days). By default -eres is "2" so after 10 days miner will have to reallocate GPU buffers and may crash.
Buffer size itself is not related to the mining speed. Mining speed is slower with every next epoch because the data size in the buffer is growing.

[Claymore]

v4.2:

- added support for Ethereum solo mining.
- added "-gser" option.
- added "-eres" option.
- added "-powlim" option.
- added "-etht" option.
- "-tt 1" (default) now does not manage fans but shows GPU temperature and fan status.
- added support for "client.reconnect" stratum command for Decred.

so we can point this version to eth-proxy?  

Yes it will work, but currently HTTP mode is not intended to work with proxies, so probably worker names will not work, I did not check it.
Why do you want to use eth-proxy? Proxy between miner and pool can cause delays and other problems.

[adamvp]

With my 380s or 370 or 280x hashrate about 10% less than ethminer  by Genoil on dwarfpool...

[melloyellow]

v4.2:

- added support for Ethereum solo mining.
- added "-gser" option.
- added "-eres" option.
- added "-powlim" option.
- added "-etht" option.
- "-tt 1" (default) now does not manage fans but shows GPU temperature and fan status.
- added support for "client.reconnect" stratum command for Decred.

so we can point this version to eth-proxy?  

Yes it will work, but currently HTTP mode is not intended to work with proxies, so probably worker names will not work, I did not check it.
Why do you want to use eth-proxy? Proxy between miner and pool can cause delays and other problems.

It's much easier to change pools once than to change it on several miners.  The HTTP requests on a local network won't have much delay and eth-proxy is connected via stratum to the pool, right?  For some reason I just have a personal preference for it, I guess.

I thought you were always going to add a single-instance proxy to the miner setup and this was a step towards your monitoring implementation, maybe I misunderstood.

[Claymore]

v4.2:

- added support for Ethereum solo mining.
- added "-gser" option.
- added "-eres" option.
- added "-powlim" option.
- added "-etht" option.
- "-tt 1" (default) now does not manage fans but shows GPU temperature and fan status.
- added support for "client.reconnect" stratum command for Decred.

so we can point this version to eth-proxy?  

Yes it will work, but currently HTTP mode is not intended to work with proxies, so probably worker names will not work, I did not check it.
Why do you want to use eth-proxy? Proxy between miner and pool can cause delays and other problems.

It's much easier to change pools once than to change it on several miners.  The HTTP requests on a local network won't have much delay and eth-proxy is connected via stratum to the pool, right?  For some reason I just have a personal preference for it, I guess.

I thought you were always going to add a single-instance proxy to the miner setup and this was a step towards your monitoring implementation, maybe I misunderstood.

I will release an utility for monitoring and changing settings for all miners at once. So right now using a proxy can be useful, but in a few days it will have no benefits. Also, eth-proxy can help with Ethereum pools, but not with Decred pools.