Quote from: adaseb on Today at 07:40:42 PM

Quote from: wacholek on Today at 07:35:30 PM

Anyone can explain why I cannot mine using Claymore's Miner anymore?

Up to yesterday there was no problem with mining on my two rigs with 2 and 3 Vega 64. Both on Windows 10. Different builds ( update disabled, drivers update disabled). On one rig I had to restart the miner but after the restart that message showed up:

GPU0 - not enough GPU memory to place DAG, you cannot mine this coin with this GPU
GPU2 - not enough GPU memory to place DAG, you cannot mine this coin with this GPU
GPU1 - not enough GPU memory to place DAG, you cannot mine this coin with this GPU
GPU0 - OpenCL error -61 - cannot allocate big buffer for DAG. Check readme.txt for possible solutions.
GPU2 - OpenCL error -61 - cannot allocate big buffer for DAG. Check readme.txt for possible solutions.
GPU1 - OpenCL error -61 - cannot allocate big buffer for DAG. Check readme.txt for possible solutions.
GPU2, OpenCL error -38 - cannot write buffer for DAG
GPU1, OpenCL error -38 - cannot write buffer for DAG
GPU0, OpenCL error -38 - cannot write buffer for DAG'

In the same time on the second rig Claymore's Miner was mining without a problem. So I left that one and started to look for the problem on the first one.
After an hour of try and errors I moved to another miner (SRBminer) and there was no problem with mining.
On the second rig Claymore's Miner was still mining fine. So for the test purposes I restart this one to and it throw me the same error as on the first rig.

On both rigs I have 8GB or RAM and 64GB of swap file on SSD drive.
Claymore's Miner is started with administration rights and with:
GPU_FORCE_64BIT_PTR 0
GPU_MAX_HEAP_SIZE 100
GPU_USE_SYNC_OBJECTS 1
GPU_MAX_ALLOC_PERCENT 100
GPU_SINGLE_ALLOC_PERCENT 100

On both rigs SRBminer is running just fine.

You say you have 8GB of RAM which is system memory but how much GDDR do all your GPUs have? Because the DAG is growing and lately if you got any 4GB GPUs you can't mine with those anymore.

Your only solution is to put an 8GB GPU in the main PCIe 16x slot and leave 4GB in the other slots. You can also use a different OS like Ubuntu which extends the life of 4GB GPUs but its a pain initially to set up. Since undervolting and underclock is not as easy as Windows with afterburner.

Then of course the other solution is just to upgrade to all 8GB GPUs. So in 2 months when all the 4GB ASICS and GPUs stop mining the hashrate will drop and it will be very profitable with any >6GB GPUS or higher.

I have already stated that I have in total 5 AMD Vega 64 GPU's which has 8GB of HBM2 on board.