ethminer 0.15.0rc2 -X Mixed Mode doesn't use all CUDA Cards & Reduces the speed

Created on 25 Jun 2018  路  12Comments  路  Source: ethereum-mining/ethminer

ethminer 0.15.0rc2 -X Mixed Mode Issue
See my Last Post (below) for a good work around
On one Miner all CUDA cards are not Used and CUDA card speed is shockingly reduced on those GPU used.
The other virtually identical rig it all works fine

The only difference I can detect is the GPU on the motherboard has a different driver.
This driver difference means the OPEN CL Platforms are:
    0,1 on the rig which works ok
    0,1,2 as per this rig which is not working properly
        Platform 2 is not mentioned in the help 
                     I'm not sure if this is part of the issue

Rigs have
6 x GeForce GTX 1080 8GB (MSI ) with latest Nvidia Driver 24.21.13.9811
4 x RX580 8GB (Azus) with latest AMD/Radeon Driver 24.20.11021.1000
Microsoft Windows 10 x64 [Version 10.0.17134.112] (Updates all current)
16GB RAM
PAGING FILE SIZE 30-60GB
CUDA_DEVICE_ORDER =PCI_BUS_ID
ethminer 0.15.0rc2
Same boards, memory, drive, Risers, PCI slots used, Power supplies (2 x 1.6kW) & Power Connections.

CUDA Only Mining (-U)
ethminer.exe --report-hashrate --farm-recheck 2000 --HWMON 1 -U --cuda-parallel-hash 8 -P stratum+ssl://[MyAddress].[MinerName]@us2.ethermine.org:5555
Uses only all 6 x GeForce GTX 1080 each stable at 39 Mh/s
Speed 236.69 Mh/s

Mixed CUDA & opencl Mining (-X)
ethminer.exe --report-hashrate --farm-recheck 2000 --HWMON 1 -X --cuda-parallel-hash 8 --opencl-platform 2 --opencl-devices 0 1 2 3 -P stratum+ssl://[MyAddress].[MinerName]@us2.ethermine.org:5555
4x (only) of the 6 x GTX 1080 (& the performance of each GTX 1080 is reduced severely to around around 27 Mh/s)
4x RX580
Speed 180.44 Mh/s

OUTPUT of
ethminer.exe" -X --list-devices

Listing OpenCL devices.
FORMAT: [platformID] [deviceID] deviceName
[0] [0] Intel(R) HD Graphics 630
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 6742532096
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 3371266048
CL_DEVICE_MAX_WORK_GROUP_SIZE: 256
[1] [0] GeForce GTX 1080
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 8589934592
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 2147483648
CL_DEVICE_MAX_WORK_GROUP_SIZE: 1024
[1] [1] GeForce GTX 1080
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 8589934592
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 2147483648
CL_DEVICE_MAX_WORK_GROUP_SIZE: 1024
[1] [2] GeForce GTX 1080
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 8589934592
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 2147483648
CL_DEVICE_MAX_WORK_GROUP_SIZE: 1024
[1] [3] GeForce GTX 1080
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 8589934592
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 2147483648
CL_DEVICE_MAX_WORK_GROUP_SIZE: 1024
[1] [4] GeForce GTX 1080
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 8589934592
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 2147483648
CL_DEVICE_MAX_WORK_GROUP_SIZE: 1024
[1] [5] GeForce GTX 1080
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 8589934592
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 2147483648
CL_DEVICE_MAX_WORK_GROUP_SIZE: 1024
[2] [0] Ellesmere
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 8589934592
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 4244635648
CL_DEVICE_MAX_WORK_GROUP_SIZE: 256
[2] [1] Ellesmere
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 8589934592
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 4244635648
CL_DEVICE_MAX_WORK_GROUP_SIZE: 256
[2] [2] Ellesmere
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 8589934592
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 4244635648
CL_DEVICE_MAX_WORK_GROUP_SIZE: 256
[2] [3] Ellesmere
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 8589934592
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 4244635648
CL_DEVICE_MAX_WORK_GROUP_SIZE: 256

Listing CUDA devices.
FORMAT: [deviceID] deviceName
[0] GeForce GTX 1080
Compute version: 6.1
cudaDeviceProp::totalGlobalMem: 8589934592
Pci: 0000:04:00
[1] GeForce GTX 1080
Compute version: 6.1
cudaDeviceProp::totalGlobalMem: 8589934592
Pci: 0000:05:00
[2] GeForce GTX 1080
Compute version: 6.1
cudaDeviceProp::totalGlobalMem: 8589934592
Pci: 0000:06:00
[3] GeForce GTX 1080
Compute version: 6.1
cudaDeviceProp::totalGlobalMem: 8589934592
Pci: 0000:0a:00
[4] GeForce GTX 1080
Compute version: 6.1
cudaDeviceProp::totalGlobalMem: 8589934592
Pci: 0000:0b:00
[5] GeForce GTX 1080
Compute version: 6.1
cudaDeviceProp::totalGlobalMem: 8589934592
Pci: 0000:0f:00

OUTPUT OF
ethminer.exe --report-hashrate --farm-recheck 2000 --HWMON 1 -X --cuda-parallel-hash 8 --opencl-platform 2 --opencl-devices 0 1 2 3 -P stratum+ssl://[MyAddress].[MinerName]@us2.ethermine.org:5555

m 16:44:26 main ethminer 0.15.0rc2
m 16:44:26 main Build: windows/release
i 16:46:04 main Found suitable OpenCL device [Ellesmere] with 8589934592 bytes of GPU memory
i 16:46:04 main Found suitable OpenCL device [Ellesmere] with 8589934592 bytes of GPU memory
i 16:46:04 main Found suitable OpenCL device [Ellesmere] with 8589934592 bytes of GPU memory
i 16:46:04 main Found suitable OpenCL device [Ellesmere] with 8589934592 bytes of GPU memory
cu 16:46:05 main Using grid size: 8192, block size: 128
i 16:46:37 main Selected pool us2.ethermine.org:5555
i 16:46:37 stratum Trying 34.213.1.169:5555 ...
i 16:46:38 stratum Connected to us2.ethermine.org [34.213.1.169:5555]
i 16:46:38 stratum Spinning up miners...
i 16:46:38 cuda-0 No work. Pause for 3 s.
i 16:46:38 cuda-1 No work. Pause for 3 s.
i 16:46:38 cuda-2 No work. Pause for 3 s.
i 16:46:38 cuda-3 No work. Pause for 3 s.
cl 16:46:38 cl-4 No work. Pause for 3 s.
cl 16:46:38 cl-5 No work. Pause for 3 s.
cl 16:46:38 cl-6 No work. Pause for 3 s.
cl 16:46:38 cl-7 No work. Pause for 3 s.
i 16:46:38 stratum Subscribed to stratum server
i 16:46:39 stratum Authorized worker [MyAddress].[MinerName]
i 16:46:39 stratum Job: #02b5e962... us2.ethermine.org [34.213.1.169:5555]
i 16:46:39 stratum Pool difficulty: 4.00K megahash
i 16:46:39 stratum New epoch 194
i 16:46:41 cuda-1 Initialising miner 1
i 16:46:41 cuda-0 Initialising miner 0
i 16:46:41 cuda-3 Initialising miner 3
i 16:46:41 cuda-2 Initialising miner 2
cu 16:46:41 cuda-0 Using device: GeForce GTX 1080 (Compute 6.1)
cl 16:46:41 cl-6 Platform: AMD Accelerated Parallel Processing
cl 16:46:41 cl-5 Platform: AMD Accelerated Parallel Processing
cl 16:46:41 cl-4 Platform: AMD Accelerated Parallel Processing
cu 16:46:41 cuda-3 Using device: GeForce GTX 1080 (Compute 6.1)
cl 16:46:41 cl-7 Platform: AMD Accelerated Parallel Processing
cu 16:46:41 cuda-2 Using device: GeForce GTX 1080 (Compute 6.1)
cu 16:46:41 cuda-1 Using device: GeForce GTX 1080 (Compute 6.1)
cl 16:46:41 cl-6 Device: Ellesmere / OpenCL 2.0 AMD-APP (2580.6)
cl 16:46:41 cl-5 Device: Ellesmere / OpenCL 2.0 AMD-APP (2580.6)
cl 16:46:41 cl-4 Device: Ellesmere / OpenCL 2.0 AMD-APP (2580.6)
cl 16:46:41 cl-7 Device: Ellesmere / OpenCL 2.0 AMD-APP (2580.6)
m 16:46:46 main Speed 0.00 Mh/s gpu/0 0.00 26C 0% 9W gpu/1 0.00 26C 0% 9W gpu/2 0.00 27C 0% 9W gpu/3 0.00 26C 0% 9W gpu/4 0.00 20C 0% 5W gpu/5 0.00 20C 0% 5W gpu/6 0.00 22C 0% 5W gpu/7 0.00 22C 0% 5W [A0+0:R0+0:F0] Time: 00:00
cu 16:46:46 cuda-3 Set Device to current
cu 16:46:46 cuda-2 Set Device to current
cu 16:46:46 cuda-1 Set Device to current
cu 16:46:46 cuda-1 Resetting device
cu 16:46:46 cuda-3 Resetting device
cu 16:46:46 cuda-2 Resetting device
cu 16:46:46 cuda-0 Set Device to current
cu 16:46:46 cuda-0 Resetting device
cu 16:46:47 cuda-3 Allocating light with size: 42204992
cu 16:46:47 cuda-3 Generating mining buffers
cu 16:46:47 cuda-2 Allocating light with size: 42204992
cu 16:46:47 cuda-1 Allocating light with size: 42204992
cu 16:46:47 cuda-0 Allocating light with size: 42204992
cu 16:46:47 cuda-3 Generating DAG for GPU #3 with dagSize: 2701130624 gridSize: 8192
cu 16:46:47 cuda-1 Generating mining buffers
cl 16:46:47 cl-6 OpenCL kernel: Stable kernel
cu 16:46:48 cuda-2 Generating mining buffers
cu 16:46:48 cuda-1 Generating DAG for GPU #1 with dagSize: 2701130624 gridSize: 8192
cu 16:46:48 cuda-0 Generating mining buffers
cu 16:46:48 cuda-2 Generating DAG for GPU #2 with dagSize: 2701130624 gridSize: 8192
cu 16:46:48 cuda-0 Generating DAG for GPU #0 with dagSize: 2701130624 gridSize: 8192
cl 16:46:48 cl-7 OpenCL kernel: Stable kernel
cl 16:46:50 cl-6 Build info:
cl 16:46:50 cl-6 Creating light cache buffer, size: 42204992
cl 16:46:50 cl-6 Creating DAG buffer, size: 2701130624
cl 16:46:50 cl-6 Loading kernels
cl 16:46:50 cl-6 Writing light cache buffer
cl 16:46:50 cl-6 Creating buffer for header.
cl 16:46:50 cl-6 Creating mining buffer
m 16:46:51 main Speed 0.00 Mh/s gpu/0 0.00 33C 0% 125W gpu/1 0.00 30C 0% 126W gpu/2 0.00 32C 0% 128W gpu/3 0.00 31C 0% 124W gpu/4 0.00 22C 0% 44W gpu/5 0.00 24C 0% 42W gpu/6 0.00 26C 0% 75W gpu/7 0.00 25C 0% 44W [A0+0:R0+0:F0] Time: 00:00
cl 16:46:51 cl-7 Build info:
cl 16:46:51 cl-7 Creating light cache buffer, size: 42204992
cl 16:46:51 cl-7 Creating DAG buffer, size: 2701130624
cl 16:46:51 cl-7 Loading kernels
cl 16:46:51 cl-7 Writing light cache buffer
cl 16:46:52 cl-7 Creating buffer for header.
cl 16:46:52 cl-7 Creating mining buffer
cl 16:46:52 cl-4 OpenCL kernel: Stable kernel
cl 16:46:53 cl-5 OpenCL kernel: Stable kernel
i 16:46:55 stratum Job: #ade3bc38... us2.ethermine.org [34.213.1.169:5555]
cl 16:46:55 cl-4 Build info:
cl 16:46:55 cl-4 Creating light cache buffer, size: 42204992
cl 16:46:55 cl-4 Creating DAG buffer, size: 2701130624
cl 16:46:55 cl-4 Loading kernels
cl 16:46:55 cl-4 Writing light cache buffer
cl 16:46:55 cl-4 Creating buffer for header.
cl 16:46:55 cl-4 Creating mining buffer
cu 16:46:55 cuda-3 Generated DAG for GPU #3 in: 8221 ms.
cu 16:46:56 cuda-1 Generated DAG for GPU #1 in: 8227 ms.
cu 16:46:56 cuda-2 Generated DAG for GPU #2 in: 8217 ms.
cu 16:46:56 cuda-0 Generated DAG for GPU #0 in: 8219 ms.
m 16:46:56 main Speed 0.00 Mh/s gpu/0 0.00 34C 0% 128W gpu/1 0.00 32C 0% 128W gpu/2 0.00 34C 0% 129W gpu/3 0.00 32C 0% 126W gpu/4 0.00 24C 0% 78W gpu/5 0.00 24C 0% 42W gpu/6 0.00 27C 0% 75W gpu/7 0.00 26C 0% 78W [A0+0:R0+0:F0] Time: 00:00
cl 16:46:56 cl-5 Build info:
cl 16:46:56 cl-5 Creating light cache buffer, size: 42204992
cl 16:46:56 cl-5 Creating DAG buffer, size: 2701130624
cl 16:46:56 cl-5 Loading kernels
cl 16:46:56 cl-5 Writing light cache buffer
cl 16:46:57 cl-5 Creating buffer for header.
cl 16:46:57 cl-5 Creating mining buffer
i 16:46:57 cl-6 2.51562 GB of DAG data generated in 7327 ms.
i 16:46:59 cl-7 2.51562 GB of DAG data generated in 7334 ms.
i 16:47:02 cl-4 2.51562 GB of DAG data generated in 7338 ms.
i 16:47:04 cl-5 2.51562 GB of DAG data generated in 7328 ms.
.
.
i 16:48:21 stratum **Accepted 203 ms. us2.ethermine.org [34.213.1.169:5555]
m 16:48:23 main Speed 180.44 Mh/s gpu/0 26.98 50C 0% 124W gpu/1 26.98 49C 0% 124W gpu/2 26.98 51C 0% 126W gpu/3 26.98 50C 0% 123W gpu/4 18.13 40C 0% 100W gpu/5 18.13 42C 0% 97W gpu/6 18.13 43C 0% 99W gpu/7 18.13 43C 0% 102W [A3+0:R0+0:F0] Time: 00:02
m 16:48:28 main Speed 179.23 Mh/s gpu/0 26.78 51C 0% 126W gpu/1 26.78 50C 0% 126W gpu/2 26.78 52C 0% 127W gpu/3 26.78 51C 0% 122W gpu/4 18.03 41C 0% 103W gpu/5 18.03 42C 0% 99W gpu/6 18.03 44C 0% 98W gpu/7 18.03 43C 0% 104W [A3+0:R0+0:F0] Time: 00:02
m 16:48:33 main Speed 179.88 Mh/s gpu/0 26.86 52C 0% 126W gpu/1 26.86 51C 0% 125W gpu/2 26.86 53C 0% 126W gpu/3 26.86 52C 0% 127W gpu/4 18.08 42C 0% 104W gpu/5 18.08 43C 0% 99W gpu/6 18.08 44C 0% 99W gpu/7 18.18 44C 0% 102W [A3+0:R0+0:F0] Time: 00:02

OUTPUT OF
ethminer.exe --report-hashrate --farm-recheck 2000 --HWMON 1 -U --cuda-parallel-hash 8 -P stratum+ssl://[MyAddress].[MinerName]@us2.ethermine.org:5555

m 18:15:38 main ethminer 0.15.0rc2
m 18:15:38 main Build: windows/release
cu 18:15:39 main Using grid size: 8192, block size: 128
i 18:17:52 main Selected pool us2.ethermine.org:5555
i 18:17:52 stratum Trying 52.37.202.243:5555 ...
i 18:17:53 stratum Connected to us2.ethermine.org [52.37.202.243:5555]
i 18:17:53 stratum Spinning up miners...
i 18:17:53 cuda-0 No work. Pause for 3 s.
i 18:17:53 cuda-1 No work. Pause for 3 s.
i 18:17:53 cuda-2 No work. Pause for 3 s.
i 18:17:53 cuda-3 No work. Pause for 3 s.
i 18:17:53 cuda-4 No work. Pause for 3 s.
i 18:17:53 cuda-5 No work. Pause for 3 s.
i 18:17:53 stratum Subscribed to stratum server
i 18:17:54 stratum Authorized worker [MyAddress].[MinerName]
i 18:17:54 stratum Job: #3897f818... us2.ethermine.org [52.37.202.243:5555]
i 18:17:54 stratum Pool difficulty: 4.00K megahash
i 18:17:54 stratum New epoch 194
i 18:17:56 cuda-1 Initialising miner 1
i 18:17:56 cuda-0 Initialising miner 0
i 18:17:56 cuda-4 Initialising miner 4
i 18:17:56 cuda-3 Initialising miner 3
i 18:17:56 cuda-2 Initialising miner 2
cu 18:17:56 cuda-0 Using device: GeForce GTX 1080 (Compute 6.1)
i 18:17:56 cuda-5 Initialising miner 5
cu 18:17:56 cuda-4 Using device: GeForce GTX 1080 (Compute 6.1)
cu 18:17:56 cuda-1 Using device: GeForce GTX 1080 (Compute 6.1)
cu 18:17:56 cuda-2 Using device: GeForce GTX 1080 (Compute 6.1)
cu 18:17:56 cuda-3 Using device: GeForce GTX 1080 (Compute 6.1)
cu 18:17:56 cuda-5 Using device: GeForce GTX 1080 (Compute 6.1)
m 18:17:57 main Speed 0.00 Mh/s gpu/0 0.00 35C 0% 10W gpu/1 0.00 34C 0% 10W gpu/2 0.00 38C 0% 9W gpu/3 0.00 36C 0% 10W gpu/4 0.00 28C 0% 9W gpu/5 0.00 24C 0% 9W [A0+0:R0+0:F0] Time: 00:02
cu 18:18:01 cuda-0 Set Device to current
cu 18:18:01 cuda-4 Set Device to current
cu 18:18:01 cuda-1 Set Device to current
cu 18:18:01 cuda-2 Set Device to current
cu 18:18:01 cuda-3 Set Device to current
cu 18:18:01 cuda-5 Set Device to current
cu 18:18:01 cuda-0 Resetting device
cu 18:18:01 cuda-4 Resetting device
cu 18:18:01 cuda-1 Resetting device
cu 18:18:01 cuda-2 Resetting device
cu 18:18:01 cuda-3 Resetting device
cu 18:18:01 cuda-5 Resetting device
m 18:18:02 main Speed 0.00 Mh/s gpu/0 0.00 35C 0% 46W gpu/1 0.00 34C 0% 10W gpu/2 0.00 38C 0% 46W gpu/3 0.00 36C 0% 10W gpu/4 0.00 28C 0% 46W gpu/5 0.00 24C 0% 9W [A0+0:R0+0:F0] Time: 00:02
cu 18:18:02 cuda-0 Allocating light with size: 42204992
cu 18:18:03 cuda-1 Allocating light with size: 42204992
cu 18:18:03 cuda-0 Generating mining buffers
cu 18:18:03 cuda-4 Allocating light with size: 42204992
cu 18:18:03 cuda-0 Generating DAG for GPU #0 with dagSize: 2701130624 gridSize: 8192
cu 18:18:03 cuda-1 Generating mining buffers
cu 18:18:03 cuda-2 Allocating light with size: 42204992
cu 18:18:03 cuda-4 Generating mining buffers
cu 18:18:03 cuda-3 Allocating light with size: 42204992
cu 18:18:03 cuda-5 Allocating light with size: 42204992
cu 18:18:03 cuda-2 Generating mining buffers
cu 18:18:04 cuda-1 Generating DAG for GPU #1 with dagSize: 2701130624 gridSize: 8192
cu 18:18:04 cuda-4 Generating DAG for GPU #4 with dagSize: 2701130624 gridSize: 8192
cu 18:18:04 cuda-2 Generating DAG for GPU #2 with dagSize: 2701130624 gridSize: 8192
cu 18:18:04 cuda-5 Generating mining buffers
cu 18:18:04 cuda-3 Generating mining buffers
cu 18:18:04 cuda-5 Generating DAG for GPU #5 with dagSize: 2701130624 gridSize: 8192
cu 18:18:04 cuda-3 Generating DAG for GPU #3 with dagSize: 2701130624 gridSize: 8192
m 18:18:08 main Speed 0.00 Mh/s gpu/0 0.00 40C 0% 126W gpu/1 0.00 39C 0% 122W gpu/2 0.00 42C 0% 124W gpu/3 0.00 40C 0% 123W gpu/4 0.00 32C 0% 121W gpu/5 0.00 29C 0% 128W [A0+0:R0+0:F0] Time: 00:02
cu 18:18:10 cuda-0 Generated DAG for GPU #0 in: 6904 ms.
cu 18:18:11 cuda-1 Generated DAG for GPU #1 in: 6963 ms.
cu 18:18:11 cuda-4 Generated DAG for GPU #4 in: 6964 ms.
cu 18:18:11 cuda-2 Generated DAG for GPU #2 in: 6908 ms.
cu 18:18:11 cuda-5 Generated DAG for GPU #5 in: 6938 ms.
cu 18:18:11 cuda-3 Generated DAG for GPU #3 in: 6953 ms.
.
.
m 18:18:34 main Speed 236.69 Mh/s gpu/0 39.41 45C 0% 127W gpu/1 39.41 44C 0% 127W gpu/2 39.41 47C 0% 125W gpu/3 39.62 45C 0% 128W gpu/4 39.41 38C 0% 123W gpu/5 39.41 34C 0% 125W [A3+0:R0+0:F0] Time: 00:02
m 18:18:39 main Speed 236.14 Mh/s gpu/0 39.45 46C 0% 126W gpu/1 39.26 45C 0% 127W gpu/2 39.45 48C 0% 124W gpu/3 39.26 46C 0% 123W gpu/4 39.45 39C 0% 128W gpu/5 39.26 35C 0% 126W [A3+0:R0+0:F0] Time: 00:02

bug committed duplicate opencl
Source
picture sirplus51

All 12 comments

Mixed mode is, unfortunately, broken.
As NVIDIA cards are usable in both modes (OpenCL or CUDA) the logic ethminer picks them is not efficient. Plus ... using NVIDIA cards in OpenCL mode reduces hashrate

As I understand your GPU configuration should be

| Id | Device Type
| --| ------|
|0| Intel Graphics
|1| Nvidia Gtx 1080
|2| Nvidia Gtx 1080
|3| Nvidia Gtx 1080
|4| Nvidia Gtx 1080
|5| Nvidia Gtx 1080
|6| Nvidia Gtx 1080
|7| AMD Rx 580
|8| AMD Rx 580
|9| AMD Rx 580
|10| AMD Rx 580

I suggest you to run two separate instances of ethminer: one full CUDA and the second one full OpenCL in this exact order

  1. ethminer -U [other params] _no need to specify --cuda-devices as ethminer will pick them all_
  2. ethminer -G --opencl-devices 7 8 9 10 [other params] _here you have to specify which opencl devices you want as all depicted above are valid for OpenCl_

This way you'll get the max out of your cards but will loose something also

  • You won't have the display of a cumulative hashing speed value
  • You don't cluster all cards as a single worker thus processing twice as much jobs

Nevertheless your revenues should not be affected.

Edit

The second instance (the one for OpenCl) could also be issued like

ethminer -G --opencl-platform 2 [other params]

AndreaLanfranchi picture AndreaLanfranchi on 25 Jun 2018
👍1

Duplicate of #1128

AndreaLanfranchi picture AndreaLanfranchi on 25 Jun 2018

Thanks for that and running 2 miners will certainly be my fallback position only it will make it interesting controlling everything with the bat files.

Yes, only my other rig works perfectly so is this really on the 馃挼 ?
Hard to believe that rig working correctly is a fluke.

I see it as an indication mixed mining can work (if some unknown variables are correct).

That a work-around can be massaged out of the Ether..
In fact I'm more than hopeful
Ether can in turn be massaged from a work-around.. 馃槃

I'm going to start with retrograding the upgraded driver.

Id | Device Type
-- | --
0 | Intel Graphics <-- Driver was upgraded
|Make it
0 | Microsoft Basic Display Adapter

Failing that I'll just swap the 2 x ssd containing the OS and see what happens.

sirplus51 picture sirplus51 on 26 Jun 2018

Yes, only my other rig works perfectly so is this really on the

Is it with mixed cards too ?

AndreaLanfranchi picture AndreaLanfranchi on 26 Jun 2018

Yes
Identical Rig (as per third line of original post -prob TLDR)

sirplus51 picture sirplus51 on 26 Jun 2018

The only difference being a different driver on the inbuilt GPU on the Identical motherboard.

The GPU cards are all consecutive serial numbers.

sirplus51 picture sirplus51 on 26 Jun 2018

I rolled back the MoBoard Driver

Listing OpenCL devices.
was

[platformID] |[deviceID] |deviceName
--|--|--
[0] |[0] |Intel(R) HD Graphics 630
[1] |[0] |GeForce GTX 1080
[1] |[1] |GeForce GTX 1080
[1] |[2] |GeForce GTX 1080
[1] |[3] |GeForce GTX 1080
[1] |[4] |GeForce GTX 1080
[1] |[5] |GeForce GTX 1080
[2] |[0] |Ellesmere
[2] |[1] |Ellesmere
[2] |[2] |Ellesmere
[2] |[3] |Ellesmere

Note there were Platforms 0,1,2 ie 3 Platforms
On the "working well mixed mining Rig" there are only 2 Platforms.
The MoBoard GPU with Microsoft Driver does not take any OpenCL Platform .
Listing OpenCL devices.

[platformID] |[deviceID] |deviceName
--|--|--
[0] |[0] |GeForce GTX 1080
[0] |[1] |GeForce GTX 1080
[0] |[2] |GeForce GTX 1080
[0] |[3] |GeForce GTX 1080
[0] |[4] |GeForce GTX 1080
[0] |[5] |GeForce GTX 1080
[1] |[0] |Ellesmere
[1] |[1] |Ellesmere
[1] |[2] |Ellesmere
[1] |[3] |Ellesmere

Rolling back the driver on the MoBoard from

Id | Device Type
-- | --
0 | Intel Graphics聽<-- Driver was upgraded 聽
Now it's |
0 | Microsoft Basic Display Adapter

On the PROBLEM WITH mixed mining Rig there are now only 2 Platforms
Listing OpenCL devices.

[platformID] |[deviceID] |deviceName
--|--|--
[0] |[0] |GeForce GTX 1080
[0] |[1] |GeForce GTX 1080
[0] |[2] |GeForce GTX 1080
[0] |[3] |GeForce GTX 1080
[0] |[4] |GeForce GTX 1080
[0] |[5] |GeForce GTX 1080
[1] |[0] |Ellesmere
[1] |[1] |Ellesmere
[1] |[2] |Ellesmere
[1] |[3] |Ellesmere

AND
Mixed Mining Now Works Better but NOT Correctly!!!
ethminer.exe --report-hashrate --farm-recheck 2000 --HWMON 1 -X --cuda-parallel-hash 8 --opencl-platform 1 --opencl-devices 0 1 2 3 -P stratum+ssl://[MyAddress].[MinerName]@us2.ethermine.org:5555

The flowing Helped the CUDA Cards to run at full Potential, however a max of only 4 Are used still:

  1. max of 2 Platforms (0 & 1)
  2. specify BOTH
    2.A. opencl platform
    2.B. opencl card to do mixed mining.

So I still have only 8 of the 10 cards being used although I now get Hash of 248 vs 179 from before.
NB running all the Cuda cards in CUDA only Mining gave 240 so I'll put the AMD cards in another Rig and should get around the mark

sirplus51 picture sirplus51 on 26 Jun 2018

GOT IT !!!

Mixed Mining Now Works Correctly!!!
ethminer.exe --report-hashrate --farm-recheck 2000 --HWMON 1 -X --cuda-parallel-hash 8 --opencl-platform 1 -P stratum+ssl://[MyAddress].[MinerName]@us2.ethermine.org:5555

The flowing Helped the CUDA Cards to run at full Potential, AND All opencl cards to be used

  1. max of 2 Platforms (0 & 1) (Set Onboard Driver to Microsoft Basic Display Adapter)
    1.A. This fixed the performance of the Cuda cards
  2. specify JUST opencl platform (1?) unless you have more cards on that platform than the CUDA Platform
    2.A. Do NOT Use --opencl-devices at all unless necessary (if you have less opencl cards than cuda then MUST let it use the whole Platform).

So I now have 10 of the 10 MIXED cards being used at MAXIMUM Potential

The Bug is with
--opencl-devices causes only those card Numbers in cuda platform to be used in Cuda AS WELL as the corresponding opencl cards in opencl platform X

Do NOT Use --opencl-devices

sirplus51 picture sirplus51 on 26 Jun 2018
👍1

Nice ... will keep open for other's reference.

Btw ... --farm-recheck 2000 is totally unneeded in stratum mode.

AndreaLanfranchi picture AndreaLanfranchi on 29 Jun 2018

isn't 2000 dangerously high? i imagine that will significantly increase stale shares 馃槻

divinity76 picture divinity76 on 29 Jun 2018

--farm-recheck has only meaning in getwork mode.
In stratum mode users keep setting it but it does nothing.

AndreaLanfranchi picture AndreaLanfranchi on 29 Jun 2018

Resolved by #1704

AndreaLanfranchi picture AndreaLanfranchi on 27 Nov 2018
Was this page helpful?
0 / 5 - 0 ratings

Related issues

gennadiv picture gennadiv  路  5Comments
rawsh picture rawsh  路  5Comments
krrkrr picture krrkrr  路  5Comments
RJBetancor picture RJBetancor  路  5Comments
sur1v picture sur1v  路  4Comments