Closed aashidham closed 5 years ago
@jean-m-cyr @AndreaLanfranchi @chfast any insights about this?
^ ping on this?
I'm not ignoring this issue. I am just not able to reproduce it on my single 480 GPU. I'm not sure what 'Clover' is! I generally focus on Nvidia stuff.
@ddobreff Do you see any problem with this configuration?
Install opencl dependencies from 18.10 amdgpu-pro package. AMD added support for Vega and Lexa since this package.
What is the apt name for this package? Is it just amdgpu-pro?
On Jun 24, 2018, at 10:28 AM, Dobromir Dobrev notifications@github.com wrote:
Install opencl dependencies from 18.10 amdgpu-pro package. AMD added support for Vega and Lexa since this package.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.
I don't think there's a apt package for that version. I think you have to download and follow instructions here: https://support.amd.com/en-us/kb-articles/Pages/Radeon-Software-for-Linux-Release-Notes.aspx
OK, well that's the exact link I followed already; thats the AMD tarball that contains ~/projects/graphics_mining/amdgpu-pro-18.20-606296
. Are you saying I should downgrade from 18.20
to 18.10
?
If the latest version (18.20
) is fine, then maybe it would help to specify exactly which of the opencl packages above I should install that I haven't tried already? I've already tried libopencl1-amdgpu-pro_18.20-606296_amd64.deb
and opencl-amdgpu-pro_18.20-606296_amd64.deb
and they haven't worked, as I mentioned in the original post.
All packages have dependencies, just track dependencies of libopencl/opencl/amdgpu-core, you can use dpkg-deb -R to unpack amdgpu-core and remove amdgpu-pro-dkms package requirement.
Do you still have the issue?
Yes. I don't get which of the packages to install. I have trouble understanding your prior comment.
On Tue, Jun 26, 2018 at 4:54 PM Dobromir Dobrev notifications@github.com wrote:
Do you still have the issue?
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/ethereum-mining/ethminer/issues/1294#issuecomment-400498889, or mute the thread https://github.com/notifications/unsubscribe-auth/AAvvSY9-oVbRr_UAv7pTlyPJtNOc_c2Qks5uAsmcgaJpZM4U0jtj .
Just install the hybrid driver amdgpu-pro-18.20 and it should work.
Like I mentioned in the original post, those drivers are the first thing I installed.
On Wed, Jun 27, 2018 at 9:04 AM Dobromir Dobrev notifications@github.com wrote:
Just install the hybrid driver amdgpu-pro-18.20 and it should work.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/ethereum-mining/ethminer/issues/1294#issuecomment-400731698, or mute the thread https://github.com/notifications/unsubscribe-auth/AAvvSWKBk6JxZANVboSus6nw1-Uo79lMks5uA60XgaJpZM4U0jtj .
Could you please post the output of
ethminer -G --list-devices
please ?
This is what I get:
test@test:~$ ethminer -G --list-devices
Listing OpenCL devices.
FORMAT: [platformID] [deviceID] deviceName
[0] [0] Radeon 550 Series (AMD POLARIS12 / DRM 3.25.0 / 4.13.0-45-generic, LLVM 5.0.0)
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 15430149732
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 3857537433
CL_DEVICE_MAX_WORK_GROUP_SIZE: 256
[1] [0] gfx804
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 3967397888
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 3725058048
CL_DEVICE_MAX_WORK_GROUP_SIZE: 256
Would you please try to relaunch with this parameters ?
ethminer -G --opencl-platform 0 --cl-global-work -1
--pool stratum+tcp://0xdeadbeef@us1.ethpool.org:3333
Then report back
Unfortunately I still have 0 Mh/s:
test@test:~/projects/graphics_mining/ethminer_bin$ ./ethminer -G --opencl-platform 0 --cl-global-work -1 --pool stratum+tcp://0xdeadbeef@us1.ethpool.org:3333
m 03:46:11 ethminer ethminer 0.15.0rc2
m 03:46:11 ethminer Build: linux/release
i 03:46:11 ethminer Found suitable OpenCL device [Radeon 550 Series (AMD POLARIS12 / DRM 3.25.0 / 4.13.0-45-generic, LLVM 5.0.0)] with 15430149732 bytes of GPU memory
i 03:46:11 main Selected pool us1.ethpool.org:3333
i 03:46:11 stratum Trying 35.185.92.48:3333 ...
i 03:46:11 stratum Connected to us1.ethpool.org [35.185.92.48:3333]
i 03:46:11 stratum Spinning up miners...
cl 03:46:11 cl-0 No work. Pause for 3 s.
i 03:46:11 stratum Subscribed to stratum server
i 03:46:11 stratum Authorized worker 0xdeadbeef
i 03:46:11 stratum Job: #73fcc765… us1.ethpool.org [35.185.92.48:3333]
i 03:46:11 stratum Pool difficulty: 4.00K megahash
i 03:46:11 stratum New epoch 196
cl 03:46:14 cl-0 Platform: Clover
cl 03:46:14 cl-0 Device: Radeon 550 Series (AMD POLARIS12 / DRM 3.25.0 / 4.13.0-45-generic, LLVM 5.0.0) / OpenCL 1.1 Mesa 17.2.8
cl 03:46:14 cl-0 OpenCL 1.1 not supported, but platform Clover might work nevertheless. USE AT OWN RISK!
cl 03:46:15 cl-0 OpenCL kernel: Stable kernel
X 03:46:16 cl-0 Build info:<unknown>:0:0: in function ethash_calculate_dag_item void (i32, %union.compute_hash_share addrspace(1)*, %union.compute_hash_share addrspace(1)*, i32): unsupported initializer for address space
X 03:46:16 cl-0 OpenCL Error: clEnqueueWriteBuffer: CL_INVALID_MEM_OBJECT (-38)
m 03:46:16 ethminer Speed 0.00 Mh/s gpu/0 0.00 [A0+0:R0+0:F0] Time: 00:00
i 03:46:19 stratum Job: #c2e6750f… us1.ethpool.org [35.185.92.48:3333]
m 03:46:21 ethminer Speed 0.00 Mh/s gpu/0 0.00 [A0+0:R0+0:F0] Time: 00:00
m 03:46:26 ethminer Speed 0.00 Mh/s gpu/0 0.00 [A0+0:R0+0:F0] Time: 00:00
m 03:46:31 ethminer Speed 0.00 Mh/s gpu/0 0.00 [A0+0:R0+0:F0] Time: 00:00
i 03:46:36 stratum Job: #267002ed… us1.ethpool.org [35.185.92.48:3333]
m 03:46:36 ethminer Speed 0.00 Mh/s gpu/0 0.00 [A0+0:R0+0:F0] Time: 00:00
m 03:46:41 ethminer Speed 0.00 Mh/s gpu/0 0.00 [A0+0:R0+0:F0] Time: 00:00
@aashidham your OpenCL driver is not able to compile the OpenCL kernel. The Clover OpenCL implementations you are using is production ready and does not give good performance. You should install amdgpu-pro driver.
He stated that the driver is installed but the presence of Mesa/clover is also noted. Please remove mesa opencl installation and install amdgpu-pro-18.20 cleanly.
Apologies for my ignorance, but how do I remove Mesa / clover and do a clean install on Ubuntu 16.04? What specific packages should I delete / install? Still not clear about the specifics.
Same problem here, but on Windows.
m 23:57:52 main ethminer 0.15.0 m 23:57:52 main Build: windows/release i 23:57:52 main Found suitable OpenCL device [Intel(R) HD Graphics 520] with 3380779418 bytes of GPU memory i 23:57:52 main Selected pool europe.ethash-hub.miningpoolhub.com:20535 i 23:57:52 stratum Trying 35.156.160.57:20535 ... i 23:57:52 stratum Connected to europe.ethash-hub.miningpoolhub.com [35.156.160.57:20535] i 23:57:52 stratum Spinning up miners... cl 23:57:52 cl-0 No work. Pause for 3 s. i 23:57:52 stratum Subscribed to stratum server i 23:57:52 stratum Extranonce set to 2bb0 (nicehash) i 23:57:52 stratum Difficulty set to 1.00117 (nicehash) i 23:57:52 stratum Job: #6b01e0ee... europe.ethash-hub.miningpoolhub.com [35.156.160.57:20535] i 23:57:52 stratum Pool difficulty: 4.30K megahash i 23:57:52 stratum New epoch 208 i 23:57:52 stratum Authorized worker test.test i 23:57:52 stratum Job: #bd148c98... europe.ethash-hub.miningpoolhub.com [35.156.160.57:20535] cl 23:57:55 cl-0 Platform: Intel(R) OpenCL cl 23:57:55 cl-0 Device: Intel(R) HD Graphics 520 / OpenCL 2.0 m 23:57:57 main Speed 0.00 Mh/s gpu/0 0.00 [A0+0:R0+0:F0] Time: 00:00 cl 23:57:57 cl-0 OpenCL kernel: Stable kernel cl 23:57:58 cl-0 Build info: fcl build 1 succeeded. bcl build succeeded.
cl 23:57:58 cl-0 Creating light cache buffer, size: 44039104 cl 23:57:58 cl-0 Creating DAG buffer, size: 2818571648 X 23:57:58 cl-0 Creating DAG buffer failed: clCreateBuffer: CL_INVALID_BUFFER_SIZE (-61) X 23:57:59 cl-0 OpenCL Error: clEnqueueWriteBuffer: CL_INVALID_MEM_OBJECT (-38) i 23:57:59 stratum Job: #6ef718f8... europe.ethash-hub.miningpoolhub.com [35.156.160.57:20535] m 23:58:02 main Speed 0.00 Mh/s gpu/0 0.00 [A0+0:R0+0:F0] Time: 00:00 i 23:58:06 stratum Job: #45f71106... europe.ethash-hub.miningpoolhub.com [35.156.160.57:20535] m 23:58:07 main Speed 0.00 Mh/s gpu/0 0.00 [A0+0:R0+0:F0] Time: 00:00 i 23:58:09 stratum Job: #19241eff... europe.ethash-hub.miningpoolhub.com [35.156.160.57:20535] m 23:58:12 main Speed 0.00 Mh/s gpu/0 0.00 [A0+0:R0+0:F0] Time: 00:00 i 23:58:15 stratum Job: #e23fe665... europe.ethash-hub.miningpoolhub.com [35.156.160.57:20535] m 23:58:17 main Speed 0.00 Mh/s gpu/0 0.00 [A0+0:R0+0:F0] Time: 00:00
is it possible this gpu is not large enough in terms of ram size for dag file?
@lesjokolat: just checked it and yes, that seems to be the problem. I'll check if I can increase it. thanks for your help!
No feed back. Closing
Hello, I have
Ubuntu 16.04.4 LTS
(xenial
) with a Radeon RX 550 Series graphics card. I'm trying to use OpenCL 1.1 to mine with it. However,ethminer
just gives back 0.00 Mh / s when I'm trying to mine with it.Here is the output for
ethminer
:Here is the output of
glmark2
:Here is the output of
clinfo
:I had do
apt install mesa-opencl-icd
to get this to work. I have the drivers fromamdgpu-pro-18.20-606296
installed.I tried installing a newer OpenCL ICD loader (
sudo gdebi opencl-amdgpu-pro-icd_18.20-606296_amd64.deb
) installed but this stoppedclinfo
andethminer
from recognizing the graphics card.I also tried installing newer OpenCL libraries from the AMD drivers (
libopencl1-amdgpu-pro_18.20-606296_amd64.deb
andopencl-amdgpu-pro_18.20-606296_amd64.deb
) but this had no effect.Here are all the packages that are available from the AMD driver tarball:
How do I either correctly install newer OpenCL libraries on Ubuntu or make
OpenCL 1.1
work withethminer
?