The oldest driver that I tested is the 18.50 Linux driver (CL version 2766.4) that is over 1.5 years old. Which one you are using (Note: its worth updating them from time to time ... especially that old blockchain driver 17.something ... that special functions this one had should now be given with every current one).


This does not seem to be related to miner... thats more a wallet issue.. BUT: First: which coin? If you need wallet help thats relevant... Second: I would never post a private key publicly ...




I wonder if it would be worth mining that. Likely: no. The card may do maybe 6-8 sol/s or so ... At current network speed that would be maybe 30 xgm/day. Sounds a lot, but is actually 11ct or so - Unless you have free electricity... 
That said: the whole GRIMM network currently produces ~450$ / day in new coins. Even if I would ramp up to 10% (highly unlikely ... Equihash on AMDs sucks a lot) of network at 2% fee that would mean I would make less then a dollar a day. Just to explain why reactivating this old algorithm (that would need some rework to support Navi and to code in the changes Grimm made over Beam) has lesser priority then many other things I got on the to do list. This does not mean I will never do it - but already adding CTXC that produces more then 10 times more $ / day in its network was a big risk for me to never ROI.

Nope, I am not gone XD

I just have the problem that home office was not invented by parents with small kids  
Well ... and I had a hard drive crash two weeks ago, else I would be quicker with the next release

Still ... lolMiner 1.0 is in the making - and today you can get a preview of it with lolMiner 1.0 alpha 1

the download is here:

https://github.com/Lolliedieb/lolMiner-preview/releases/tag/1.0a1

Whats new:
- Added Cuckaroo-30 for mining Cortex. Use --coin CTXC to mine with it. The algorithm needs 7.6 G of GPU memory, fee is 2.5%
- Added temperature, fan speed and consumption readings (Buggy in Windows atm for cards newer then 580 ... will fix soon. Linux and Nvidias should work)
- Some bug fixes for missing driver supports (Grin algorithms)

Expected performance (Cortex):
- Radeon VII: 3.05 g/s (0.073 h/s)
- Vega 64 2.2 g/s (0.053 h/s)
- Vega 56 2.0 g/s (0.048 h/s)
- RX 5700 1.85 g/s (0.044 h/s)
- RX 580 1.25 g/s (0.030 h/s)

Nope - at current point it should work with C29M (Grin) and Beam, more to come


That can be a problem with 19.50 driver - I can not say why exactly, but for some reason Navis + 19.50 are really slow on Grin. Why? I have no clue, because 19.30 and 19.50 use the same kernel - it was even build with 19.50.
So well ... advice: if you can install 19.30 instead you should be fine


To the other two: Bring back old Beam Hash I and Temp readings are noted

Sorry for the recent silence - needed to do some other work, but will be back soon.
The next bigger release for lolMiner will likely be 1.0 ... so well ...
Since this is going to be something special I would like to ask you: what are the features that you urgently miss in lolMiner Let me know

--------------------------



No, not really ... what exactly are you searching for? Stay cooler and mine slower? That can best be done by lowering your GPU clocks


Good idea, will do so
Not yet - I never did implement it, because I started with grin when it forked to cuckarood. But it is a candidate for integration for lolMiner 1.0.

Sorry for all the driver problems on last version ... I hope this is fixed now with 0.9.8

Changelog:

• New GRIN-C29M solver, Less memory usage (fits 6G) & 8-12% better performance
• Added BEAM support for AMD Navi Generation
• Fixed a lot of driver issues of the last build with GRIN codes
• Renamed GRIN-AD29 to MWC-C29D

Link: https://github.com/Lolliedieb/lolMiner-releases/releases/tag/0.98



---------------------



Well no. The problem are two things
1) It is Equihash - that algorithms is starting with a compute intense phase and then switches to a memory intense one, where compute is much less used. As a result power consumption goes up and down and so GPU % usage
2) Aion is on a strange mix of pow and pos - every 2nd block is pos, such that in that time there is no work to mine on for GPUs...

Is this Windows 10? Because the 0.9.4 code changes increased the memory demand a bit ... could be that this causes the problem :/


In which version does this happen? Latest should work, but would like to test it directly. I had one version in between where it was indeed broken :/

The 5600xt not yet, but soon.

Hope to have a 6G version of C29M (with perfo of the current 8G code) soon and will activate the 4G C31 solver for this cards. C32 is on a bigger rework, so when thats done maybe a designated 6G version, too. Also all new codes will have Navi support

Well yes / no.
You can try to put in the command for killing the current instance and to restart it into the .bat, the question is if it will work. That depends on why exactly the card had crashed. In case of an algorithm problem it will work what you described, in case it did dead lock due to too high clock for your set voltage its very likely that the system will need a complete reboot - so it depends a bit on what had happened.

Well it is a new algorithm - specs not yet released, but soon will be. Will see how this performs

On Nicehash it would sense not to optimize the grin output but instead the btc you get for mining it. Hmm definitely possible when I can pull Nicehash API. I will check that, good idea.
But doesn't Nicehash miner already do this (it uses lolMiner for Grin algorithms on amds anyways ^^)




Let me check that ... did for ages not mine AION on Nvidias - but will have a look if its a general miner but there

Hey, just pushed 0.9.7

Changes

• Windows: Increased GRIN-C32 performance on Navi cards by 15-18%
• Linux: Increased GRIN-C32 performance on Vega, VII & Navi cards by 15-18%
• Added a Grin Auto profit switcher (use: --coin GRIN-AUTO + further options). Please read https://github.com/Lolliedieb/lolMiner-releases/wiki/Grin-Automatic-Profit-Switching carefully how to use it and which pools are supported.

If you got any questions about the automatic algorithm selector let me know (but please read that page first )

Preview of what is does:






Its 1% since 0.7.0 - and will stay there



The beam algorithm at the moment needs a big rework on Navi - a lot of things not working there. I will try to fix that with 0.9.8 (but honestly: wait for BeamHash III and mine something else and buy Beam from the mining return... currently thats the more profitable way to get Beams...)


0.9.6.1 needs a bit more memory - maybe that is the issue here. Does one of the card run the screen or is the rig headless? Also can you check how large your Windows pagefile (virtual memory) is? Please set it to at least 8G * number of cards (stupid Adrenaline Windows requirement that the virtual CPU memory needs to be as big as the used actual GPU memory... )

Were able to resolve the issue in Telegram
For some reason the 18.50 driver in Hive did not work for him, while 19.10 did (what is strange since I compiled with 18.50...)
Any ways: it worked then


For Windows users: updated version to 0.9.6.1 - fixing a bug with 8G cards on Win 10 using 4G solver on C31. Download location is same  as before, Linux is not affected afaik.

I would not say broken, but ETC is in a later dag epoch which may hurt your card more then it should (but its normal etc is slower then eth due to this). I assume at least the general implementation for etc and eth has same quality.

Just updated  lolMiner to version 0.9.6

Overall should give a nice speed bump for all GRIN algorithms on AMDs including the new C29M.

Expected Performance
Test Platform: Ubuntu 18.40, amdgpu-pro 19.30, stock cards

Algorithm	RX 570 8G	Vega 56	Vega 64	RX 5700	Radeon VII
Cuckaroom-29	2.4 g/s	4.1 g/s	4.6 g/s	4.4 g/s	6.4 g/s
Cuckatoo-31	0.67 g/s	1.15 g/s*	1.37 g/s*	1.12 g/s	2.05 g/s*
Cuckatoo-32	0.18 g/s	0.36 g/s	0.42 g/s	0.34 g/s	0.65 g/s

   
* On Windows sadly 5-7% slower due to a driver issue AMD Adrenaline has on Vegas

Just updated the version to 0.9.6

Overall should give a nice speed bump for all GRIN algorithms including the new C29M.

Expected Performance
Test Platform: Ubuntu 18.40, amdgpu-pro 19.30, stock cards

Algorithm	RX 570 8G	Vega 56	Vega 64	RX 5700	Radeon VII
Cuckaroom-29	2.4 g/s	4.1 g/s	4.6 g/s	4.4 g/s	6.4 g/s
Cuckatoo-31	0.67 g/s	1.15 g/s*	1.37 g/s*	1.12 g/s	2.05 g/s*
Cuckatoo-32	0.18 g/s	0.36 g/s	0.42 g/s	0.34 g/s	0.65 g/s

 

* On Windows sadly 5-7% slower due to a driver issue AMD Adrenaline has on Vegas

Can you share operation system.and driver version? Would like to debug, but need more info

Problem is that my solver at the moment needs 8G to work. So 390 does work, but Fury only got 4G ... I am thinking about a 4G solver but its more complicated then anticipated.

Well no I would say ... When you put in the actual values its 2nd best algo after ethash - but well that still leading the pack. Well on Vegas its different (just wtm using totally unreasonable clock settings ^^)

C31 very profitable at the moment - because the grin network needed to adjust diff down due to the C29 fork.
Miner where are you?


Ps: good progress on 4G C29M code

I will check - seems some of the cleanup does not work for unselected cards. That lolMine is displayed at all comes from.OpenCL context that is created...


By the way we had C29M blocks now, but the diff balancibg in grin takes a while to adjust to the new scheme. Give it 3-4 hrs then every thing should be normal