Thank you for clarify, may I ask if I want 10.3 to operate as 10.0 which command I have to add?
There is no way to make it work exactly as v10.0. Closest option is "-tt 0". |
|
|
|
Can you help me for this error : Temperature control thread hangs, need to restart miner!
My card has temps 62, the miner restart every 2 hours or less, every 20mn this morning
I have this error on claymore 10.2 and 10.3, i don't use previous version.
I haven't this problem with claymore for cryptonight or equihash
I also experienced same issue, Any idea how to disable the temperature control thread? Thanks You cannot. It is a driver issue. I agree, try different drivers if you are sure that hardware is ok. I see this issue on latest Adrenalin drivers on some cards, for example, I cannot make it work stable on a system with stock 1x RX480 + 1x Duo Pro on 17.12.2 drivers, looks like latest drivers are unstable in some configurations. Or you can specify "-tt 0" to disable temperature/fan control, but probably it will hang in OpenCL call in a few minutes anyway. I don't think it driver issue, it might compatibility issue with new version of claymore and blockchain driver. I use blockchain driver and having this peoblem with 10.2 and 10.3 after I went back to 10.0 never have this temperature thread hang again. (all same setting only change 10.3 to 10.0, everything stable for 24 hours, 10.3 give me hangs after 30-60 mins) Check History.txt: v10.2 ... - in rare cases ADL API calls can hang, now watchdog checks it as well. ... So now if ADL hangs, miner will be restarted. In v10.1 and earlier miner will work without temperature control (all related options will stop work too), watchdog will hang too (because it was in same thread). So you can select what behaviour is the best for you: restart miner when ADL hangs (use v10.3) or stop temperature control and related options (use v10.1), but the reason is not in miner anyway. |
|
|
|
Can you help me for this error : Temperature control thread hangs, need to restart miner!
My card has temps 62, the miner restart every 2 hours or less, every 20mn this morning
I have this error on claymore 10.2 and 10.3, i don't use previous version.
I haven't this problem with claymore for cryptonight or equihash
I also experienced same issue, Any idea how to disable the temperature control thread? Thanks You cannot. It is a driver issue. I agree, try different drivers if you are sure that hardware is ok. I see this issue on latest Adrenalin drivers on some cards, for example, I cannot make it work stable on a system with stock 1x RX480 + 1x Duo Pro on 17.12.2 drivers, looks like latest drivers are unstable in some configurations. Or you can specify "-tt 0" to disable temperature/fan control, but probably it will hang in OpenCL call in a few minutes anyway. |
|
|
|
|
I uploaded v11.2 for Linux.
|
|
|
|
|
I uploaded v10.3 for Linux.
|
|
|
|
|
I uploaded a small fix to v10.3, so re-download v10.3 if you have one AMD card and miner does not show its temperature.
|
|
|
|
|
If you have single AMD card, in some cases v10.3 will not manage fan/temperature for Adrenalin drivers, within 24 hours I will release a small update to fix it.
|
|
|
|
|
If fan/temperature control does not work for you, PM me what Windows version you use and send the complete miner log file for about two minutes of work.
|
|
|
|
|
v10.3:
- added support for Adrenalin drivers (GPU sorting, fan/temperature/clocks management).
- fixed issue with possible GPU crash during miner shutdown.
- now miner shows PCI bus number for every GPU.
- a few minor improvements.
- EthMan: some bug fixes and minor improvements.
PS. Linux version will be available in a few days.
|
|
|
|
Fixed the seg fault with NH?
I cannot see this bug (don't have it in Windows and don't have access to Linux system currently), but I found possible reason and fixed it. So need someone who sees this bug in Windows and can confirm that it is fixed in latest update. Or I will check it on Linux system directly in a few days. |
|
|
|
In 1-3 days I will release new version with Adrenalin drivers support (fan/temperature/clocks management) and some other improvements.
Good news... I guess. What's the benefit of using latest adrenalin drivers? Are there any improvements? I don't see any improvements, but some people want to use latest drivers anyway. |
|
|
|
|
In 1-3 days I will release new version with Adrenalin drivers support (fan/temperature/clocks management) and some other improvements.
|
|
|
|
I re-uploaded v11.2, fixed GPU sorting by PCI bus number, also now miner displays PCI bus number for every GPU. So if you use v11.2, please re-download it.
Please, add amd fan/clocks/voltages managment compatibility with 17.12.x drivers for your DUAL miner. Yes I will do it soon. On v11.0 (waiting for 11.2 Linux announcement) I'm only registering 2 of my 3 Vega 64 cards. I fixed most of the errors but am now getting this:
Anyone have some insight?
My miners don't support Vega in Linux. Hi I have the following problem: Claymoreminer 9.7 and 11.2 displays the GPU0 to GPU5 correctly. When I look in HWinfo, the GPU from Claymore is different. For example, Claymore GPU0 is the GPU4 in HWinfo. What can I do to make this happen? Also in Wattman it is another. Thanks
Miner sorts GPUs by PCI bus number, same as AfterBurner does. I don't know about sorting method in HWInfo. |
|
|
|
|
I re-uploaded v11.2, fixed GPU sorting by PCI bus number, also now miner displays PCI bus number for every GPU. So if you use v11.2, please re-download it.
|
|
|
|
Fiji (Fury x 1100/500) have lower hashrate -dmem 0 870H/s (-h 1825 have 935H/s) in 11.2 11.0/11.1 (970Hs/s) -dmem 1 def...
v11.1 uses "-dmem 1 -h 896" by default for Fiji chips, v11.2 uses "-dmem 0 -h 1024". First option works faster, but it requires more memory (2*896*2MB instead of 1*1024*2MB) and may fail or show very low hashrate in some cases. For example, in my tests it works fine for blockchain drivers but often fails in latest Adrenalin drivers, so I decided to reduce memory usage by default, for compatibility reasons. Of course you are free to try various settings and select the best for you. |
|
|
|
|
v11.2:
- added workaround for bug in Adrenalin drivers (failed fan/temperature management).
- changed default settings for Fiji chips.
Please let me know if this workaround works properly because I did not test it much yet. I will apply it for other miners too because it seems we can wait a fix from AMD for ages.
|
|
|
|
Is 11.1 can handle fan/temp/clocks info/control with new adrenalin 17.12.x drivers?
No. Fix from AMD is required, we have to wait it. Or I will have to do their job again and find some workaround. Allready a month has passed. Maybe it's time to stop waiting AMD and find some workaround? It's the great chance that this AMD bug will stay a long time... That's AMD... Ok I will create a workaround, it will take 1-2 days. claymore version 12.6 equiashe kindly you should fix it to decrease the devfee, and if you could try to improve we have been waiting months Wrong thread. If you check correct thread you will see my answer: https://bitcointalk.org/index.php?topic=1670733.msg27163302#msg27163302 |
|
|
|
Is 11.1 can handle fan/temp/clocks info/control with new adrenalin 17.12.x drivers?
No. Fix from AMD is required, we have to wait it. Or I will have to do their job again and find some workaround. Allready a month has passed. Maybe it's time to stop waiting AMD and find some workaround? It's the great chance that this AMD bug will stay a long time... That's AMD... Ok I will create a workaround, it will take 1-2 days. |
|
|
|
|
Show Posts
