search

trexminer / T-Rex

T-Rex NVIDIA GPU miner with web control monitoring page
2.64k stars 438 forks source link

not enough memory to hold the full DAG for epoch 404 on a 5GB VRAM GPU #360

Closed mlamlam closed 2 years ago

mlamlam commented 3 years ago

My graphic card is GTX 1060 and has 5GB VRAM, OS is Windows Server 2019 and Driver version is 470.05.

it seems that now ETH DAG size is ~4.2GB and my graphic card has 5GB VRAM, but it still say not enough memory? I have only this graphic card on this computer, Page file set to 200GB already.

20210331 18:32:41 T-Rex NVIDIA GPU miner v0.19.14 - [CUDA v11.10 | Windows] 20210331 18:32:41 r.39dd950be1c7 20210331 18:32:41 20210331 18:32:41 NVIDIA Driver v470.05 20210331 18:32:41 CUDA devices available: 1 20210331 18:32:41 20210331 18:32:41 WARN: DevFee 1% (ethash) 20210331 18:32:41 20210331 18:32:41 URL : stratum+ssl://asia1.ethermine.org:5555 20210331 18:32:41 USER: XXXXXXXXXXXXXXXXXXXXXXXX 20210331 18:32:41 PASS: x 20210331 18:32:41 WRK : my 20210331 18:32:41 20210331 18:32:41 Starting on: asia1.ethermine.org:5555 20210331 18:32:41 Using protocol: stratum1. 20210331 18:32:41 Authorizing... 20210331 18:32:41 Authorized successfully. 20210331 18:32:41 ethash epoch: 404, block: 12146577, diff: 4.00 G 20210331 18:32:41 ethash epoch: 404, block: 12146577, diff: 4.00 G 20210331 18:32:41 ApiServer: HTTP server started on 127.0.0.1:4067 20210331 18:32:41 --------------------------------------------------- 20210331 18:32:41 For control navigate to: http://127.0.0.1:4067/trex 20210331 18:32:41 --------------------------------------------------- 20210331 18:32:41 ApiServer: Telnet server started on 127.0.0.1:4068 20210331 18:32:41 WARN: GPU #0(000200): NVIDIA GeForce GTX 1060 5GB, intensity set to 22 20210331 18:32:46 ethash epoch: 404, block: 12146578, diff: 4.00 G 20210331 18:32:47 WARN: GPU #0: not enough memory to hold the full DAG for epoch 404. 20210331 18:32:47 WARN: GPU #0: effective hashrate will be ~54.5% of its usual value 20210331 18:32:47 GPU #0: generating DAG 4.12 GB (99.06% of full size) for epoch 404 ... 20210331 18:32:48 ethash epoch: 404, block: 12146579, diff: 4.00 G 20210331 18:32:57 ethash epoch: 404, block: 12146580, diff: 4.00 G 20210331 18:33:00 ethash epoch: 404, block: 12146581, diff: 4.00 G 20210331 18:33:02 GPU #0: DAG generated [crc: ed73dc0f, time: 14795 ms], memory left: 0.00 B 20210331 18:33:33 ethash epoch: 404, block: 12146582, diff: 4.00 G 20210331 18:33:34 ethash epoch: 404, block: 12146583, diff: 4.00 G

-------------- 20210331 18:33:41 --------------- Mining at asia1.ethermine.org:5555, diff: 4.00 G GPU #0: NVIDIA GTX 1060 5GB - 10.32 MH/s, [T:68C, P:88W, F:63%, E:136kH/W] Shares/min: 0 Uptime: 1 min | Algo: ethash | T-Rex v0.19.14

-------------- 20210331 18:34:11 --------------- Mining at asia1.ethermine.org:5555, diff: 4.00 G GPU #0: NVIDIA GTX 1060 5GB - 10.31 MH/s, [T:70C, P:90W, F:62%, E:117kH/W] Shares/min: 0 Uptime: 1 min 30 secs | Algo: ethash | T-Rex v0.19.14

20210331 18:34:20 ethash epoch: 404, block: 12146584, diff: 4.00 G 20210331 18:34:22 ethash epoch: 404, block: 12146585, diff: 4.00 G 20210331 18:34:41 ethash epoch: 404, block: 12146586, diff: 4.00 G

-------------- 20210331 18:34:41 --------------- Mining at asia1.ethermine.org:5555, diff: 4.00 G GPU #0: NVIDIA GTX 1060 5GB - 10.28 MH/s, [T:71C, P:91W, F:62%, E:114kH/W] Shares/min: 0 Uptime: 2 mins | Algo: ethash | T-Rex v0.19.14

20210331 18:34:44 ethash epoch: 404, block: 12146587, diff: 4.00 G 20210331 18:34:45 WARN: shutdown t-rex, got Ctrl+C from console 20210331 18:34:46 Main loop finished. Cleaning up resources... 20210331 18:34:46 ApiServer: stopped listening on 127.0.0.1:4067 20210331 18:34:46 ApiServer: stopped listening on 127.0.0.1:4068 20210331 18:34:46 T-Rex finished. 1

trexminer commented 3 years ago

WDDM is not letting t-rex allocate enough memory, not our fault :) I'd recommend using Linux, there's a good chance you'll be able to mine at full speed.

mlamlam commented 3 years ago

Is it possible to use GPU pass thru on Hyper V and mining in Guest OS (ie Ubuntu) ?

pdieppa commented 3 years ago

PowerShell commands for you to investigate:

`Dismount-VMHostAssignableDevice -LocationPath gpu-device-location -force

Add-VMAssignableDevice -LocationPath gpu-device-location -VMName`

Reference:

https://docs.microsoft.com/en-us/windows-server/virtualization/hyper-v/deploy/deploying-graphics-devices-using-dda