Topic: Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux) - page 57. (Read 6590718 times)

This does sound like a virtual memory problem.

It's a simple enough procedure, but if you need help configuring your VM in Windows, I recently updated my guide here: https://www.cryptoyeti.com/cannot-write-buffer-for-dag-error/

W10 use for the OS 600mb of GPU memory cannot be solved, it is a ''feature'' of windows 10

Increase the virtual memory.

It is funny with 4Gb video card in windows 7 I can mine, but when I install windows 10 on the same computer I can not mine anymore because I get error
CUDA error - cannot allocate big buffer for DAG.

anyone knows why is that happend?
Can I disable something in windows 10?

Dear Claymore,
1. Do you plan to do something with the problem when AMD cards with 8GB cannot load the 379th dag file?
2. Do you plan to do something with the problem when amd cards with 4gb gets 0 mega-cache from 350 to 378 dag files?

Do you have any developments on these issues?

hi, what is the best miner for 1080 TI's for ETH right now, is it again Phoenix, and what are your ideal memory strap settings ?..

Also is ETHlargement pill uniformly recommended for all setups to improve performance ?.. I had switched to mining other coins for a while, so I would appreciate the latest cutting edge configurations specifically for 1080 ti's ...

Thanks in advance

dwarfpool is dead after 16 feb. but fee to dwarf. wait new version?

Just wanted to fire up some mining since it's actually an option again (long story) but i get same error

actually same for cc old miner, new miner, ella, rvn, etc,

I was running 6 months ago and believe a power shut down let an update into windows or something. Of course i forget all this till open pc, download latest version first

Anyone still mining with RX480?
Still worthed? What are best settings and estimated speed to obtain please?

Well, basicaly in that situatuon best practice is to disassamble the rig and start building from the beginning . Starting from 1 GPU on rig board. Geeting stability on 1 GPU then adding one by one.
ofcz if you are distant is looks not so convinient , but its most efficient way.  IMHO

Dear Friends, I need your help.

I have 2 rigs (3 cards + 6 cards) RX 570-580. Before Claymore v15.0 (up to November), it used to work quite stable... I had 75 + 155 MHs.
After the update, it basically stopped working at all... Drivers seem to fall leaving me with 1-2 cards running at 10mhs and consuming 140W (according to afterburner).  I'm using Aug. blockchain drivers.

So... I tried everything I could to make it work... I literally ran out of ideas.  
- I tried different configurations, reducing mc and cc, changing power limits, voltage, etc. kind of tradeoff efficiency for stability.
- I tried to auto-restart the miner (and computer) and run via .exe and bat file. (I even tried restarting pc every 20min or so).
- different drivers didn't work as well...
- I tried changing RAM to 4, 8, 16 GB...
- configuring autoburner so I have ±100 W; stable. (gives around 22mhs per card, which is not much... and still, it crushes.

I set up everything... eave the rig... it works 15 hours in the best case scenario... (usually 2-3 hours) and then ... drivers fall leaving 2 cards 10mhs. then I reinstall drivers, config auto burner and everything repeats. I think I was able to detect the error once before the crash, the red line said: "OpenCL error, -36, cannot write buffer for DAG". I'm not sure if it's related to what happens but this what I was able to see.

At this time it's more like a hobby than earning. I call it 'my little tech club' - I'm trying to make it work because I enjoy when it works stable... I'm not a tech or hardware guy and I learned a lot through the course of these years. But past experience was very disappointing. The rig is not in my house and I can visit it once a week in the best,  I probably spend more money on gas (not counting my time).  

Please give me some ideas about what can I do to make it work stable... maybe with fewer MHS, but stable. I would appreciate any advice on how can I make it work.

Thanks, dear community!

My suggestion is quite obvious. NOT TO DELETE UNCONFORTABLE FOR CLAYMORE POSTS.

The real meaning of the deleted post was not that Claymore is physically death. Its meaning was that Clamore has stopped developping and that he is humbly accepting that competitors like Phoenixminer are 2% faster on the new emerged AMD Cards like Radeon RX 5700.

So my concrete suggestion for improvement is Claymore to increase the performance of his miner at least 2% on RX5700.

What is it that you are missing? If you have a suggestion for an improvement, why don’t you post it here?

BTW, is anyone aware of a way to control the software using an IOT temperature sensor?

This forum has been censored (guess from who). There was an answer to this post right bellow it which mysteriously disappeared. Mind what you write here. The Big Brother is watching you.

Huge link to whattomine.com
Or you can try to use nicehash -> https://www.nicehash.com/profitability-calculator

You are welcome. 

Yes, that solved the problem !!!
I did everything by description and finally everything works again )
Thanks a lot

Try to make NVSMI folder in C:\Program Files\NVIDIA Corporation and copy nvml.dll file from C:\Windows\System32 to the NVSMI

I just found this solution in some russian blog: https://www.cryptoprofi.info/?p=5362

Hi everyone,

I'm getting this error message since a few weeks:

I've searched the web for a solution for this issue, they all say to update claymore, but since
I'm using v15.0 already this doesn't help. I've tried older versions aswell: same thing
One source mentioned to change the esm value, but that didn't help either.

Any ideas on how to fix this?

Systeminfo: Windows 10, RX580s

Is Claymore dead? No further development? 

I just watched a video from Red Panda Miner using Phoenix miner to achieve 61MH/s with a 5700 XT. I hope that Claymore can find some of those optimizations for his miner..