lolMiner 1.64: Fastest Kaspa miner in market

[jgonzi]

I have been testing the Watts reduction from these differents versions:

I did it with the Navi cards in a small testing RIG... before changing my miners.



The results are really interessant... Without changing any parameters it has been a reduction of 1 Watts in each card for each new realeases (we take out the 1.14, that a small fix).

That means nearly 2 Watts from cards from version 1.12 to version 1.15. 1.5W less for card... going from 384.5W to 379.6W

That reduction was also visible at wall.. So it is interesant to update to last version not also for the updates also for the improves of watts... less see what brings newer versions.

I will do an update after changing the all farm.

[jgonzi]

i found answer  



thanks for help

So now is working?

[sk0mi]

Please,
Can you implement an option for minimal rig speed and restart if that speed is lower?
I am using lolminer through SMOS and several times now I found it to be working with 0 H/s.
Nicehash is mixing ETH and ETC and due to that miner just won’t keep working normally.
Sometimes it’s one graphic cards, sometimes it’s few, sometimes it’s all cards affected.
They have 0 H/s but miner keeps working and both SMOS and Nicehash showing rigs online but with 0 H/s.
I appreciate all you did with the miner, for me it’s the best option honestly. I wan’t to keep using it.
Please, put minimal speed option in miner, so it could restart if speed is lower than minimal.
Thank you in advance

[Lolliedieb]

Its time to say hello to lolMiner 1.16.

You can load it here: https://github.com/Lolliedieb/lolMiner-releases/releases/tag/1.16

Here is the - again not so short - change log

•     Added support of Ethash and Beam Hash III for RX 6000 generation of GPUs
•     All supported algorithms now show the share difficulty and have best share statistics.
•     New feature: use --rebuild-defect n to trigger a rebuild of DAG if a GPU produced n defect shares on the current one. Default is 3, use 0 to deactivate this feature.
•     New feature: Use --workmulti n to modify the amount of Ethash work a GPU does per batch. Higher values will cause less CPU load but more stale shares, lower values will give less stale shares but higher CPU load. Performance may vary for different values. Default is 128. 
•     New feature: if Ethash pool disconnects within 2 seconds from connection attempt (immediate reconnect), then the other stratum mode is tested to login.
•     New feature: AMD Vega and newer cards now display memory temperature in statistics & api (only visible if there is at least one such GPU in the rig).
•     Default ethstratum changed from ETHV1 to ETHPROXY for better pool compatibility.
•     Stratum pool addresses now understand "stratum+tcp://", "stratum+ssl://" and "ssl://" prefixes (turning on or of ssl / tls automatically) for better compatibility with existing configurations.   
•     Slightly reduced CPU load when mining Ethash
•     New coloring scheme with more friendly colors. For terminals that do not have rgb colors (e.g. shellinabox) use --basecolor to restrict to a simpler set. Use --nocolor to deactivate coloring completely.
•     Fixed bug: Cards may crash when switching from ZIL cache back to normal mining.
•     Fixed bug: Wavy hashrate - especially for rigs with many AMD Navi GPUs.
•     Fixed bug: (Linux:) Watchdog not called when a GPU is stuck & extremely high CPU load on crashed GPU. (1)
•     Fixed bug: Hashrate reporting not working on some pools (e.g. sparkpool)
•     Fixed bug: Miner can crash after trying to reconnect to same pool over 5 minutes.
•     Fixed bug: Miner crashes when mixing TLS and non-TLS pools for fail-over.

So pretty lot of new stuff - new cards added (I must say: 6800 is slower then most have hoped, but energy efficiency is damn impressive) - new colors (hope you like it) and a ton of fixes

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

I will do an update after changing the all farm.

Sadly no watts reduction this time - only some CPU stuff / speed stability fixes
But next I will likely put hands to code again to see what we can do

Can you implement an option for minimal rig speed and restart if that speed is lower?
I am using lolminer through SMOS and several times now I found it to be working with 0 H/s.

Problem is: if an amd card drops to 0, then usually the full card is stuck and can not recover unless you reboot the rig. In fact if you try to touch it again often the rig will completely lock up, so system reboot does not work via terminal - you then often need to power cycle it by hand. To cover this the new version finally should be able to trigger restart scripts gracefully. Also smos should be able to detect the 0 hr and reboot - or isn't it.
By the way if a card is most of the time stable but then does suddenly fall to 0 its often its voltage is one small bit to low for current settings. Increase it by 5 or 10mV and see if card stable then again.

[geck]

just swapped a couple of rigs over to lolminer. looking good so far!

by the way, what efficiency numbers are you seeing for the 6800 vs 5700?

[jgonzi]

just swapped a couple of rigs over to lolminer. looking good so far!

by the way, what efficiency numbers are you seeing for the 6800 vs 5700?

The numbers are moreorless getting 63Mhs-65Mhs with 20W less than the 5700... of course that will also depend on Silicon Lottery...

Why these Numbers... Memory is faster from 14Gbps to 16Gbps and with more CU you need less voltage to achieve that Mhs...

[alumar]

Here is my active testing thread on the 6800 non-xt

https://twitter.com/BitsBeTrippin/status/1330884361475026950

Bottom line initial testing 81MH/s on ETH @ 170w

Now time to tweak powerplay tables and see how low we can get the power down to.

[miner29]

81Mhs but it shows 2 gpus (0 and 1) so is it only 40mhs each....

Somebody got more info on what is going on there?

Ok..i see it is treating it (the miner sft) as two separate gpus with 8gb each....makes more sense

[Metroid]

Here is my active testing thread on the 6800 non-xt

https://twitter.com/BitsBeTrippin/status/1330884361475026950

Bottom line initial testing 81MH/s on ETH @ 170w

Now time to tweak powerplay tables and see how low we can get the power down to.

Not bad, $579 / 81 = 7.1 usd per 1 mhs, 3080 is $700 / 98 = 7.1, 3080 still a little better due to density but the 6800 uses a lot less power, so that is a draw.

[UnclWish]

Lolliedieb, What about this:

Another question:
RX 580-8 rig. On 1st card of rig when monitor connected and trying to look video on youtube or play any videofile (even in bad quality f.e. 320p) video playback goes with freezes. On all other miners (Claymore, TR, PM) video plays without problems during eth mining.

[mineypiney]

I have this same issue.

Another question:
RX 580-8 rig. On 1st card of rig. When monitor connected and trying to look video on youtube or play any videofile (even in bad quality f.e. 320p) video playback goes with freezes. On all other miners (Claymore, TR, PM) video plays without problems during eth mining.

My 6800XT by default treats the GPU as one, so my result is abysmal in comparison. I am still not sure if I fully understand this result.

81Mhs but it shows 2 gpus (0 and 1) so is it only 40mhs each....

Somebody got more info on what is going on there?

Ok..i see it is treating it (the miner sft) as two separate gpus with 8gb each....makes more sense

[alumar]

Here is a latest video on it. YouTube https://youtu.be/xd6qTNvWsKg

Note: My next steps today are the following

1) Run continuous for 2 hours against a pool and validate shares/hashrate

2) Test same config using 1x to 16x riser to see if SAM works same on riser

3) Test Power Usage at these rates, see about powerplay tweaking on what we can get out of it.

[mineypiney]

Is this speculation that Shared Access Memory is causing two logical devices appear? Have you toggled this on and off to confirm the scenario?

Here is a latest video on it. YouTube https://youtu.be/xd6qTNvWsKg

Note: My next steps today are the following

1) Run continuous for 2 hours against a pool and validate shares/hashrate

2) Test same config using 1x to 16x riser to see if SAM works same on riser

3) Test Power Usage at these rates, see about powerplay tweaking on what we can get out of it.

[alumar]

I did previous video with it turned off, I enabled it today and tested and got those results.

[mineypiney]

I am not sure why as a consequence the smart access memory would create a second logical GPU. Very odd. Neat in your case but seems like a strange consequence.

[alumar]

I am not sure why as a consequence the smart access memory would create a second logical GPU. Very odd. Neat in your case but seems like a strange consequence.

It's definitely a bug. I disabled 4G Decoding, BAR in the BIOS, rebooted, DDU'd the drivers, reinstalled drivers, tested without Smart Access Memory, then tested again with it back on and now back to 1 GPU still.

Looking to see if something else I was doing caused it to pop two instances of the miner. The limited time I was running, it was in fact getting 80mh at the pool and had both compute 0 and compute 1 under load in the GPU monitor; now I just get compute 0 with the single thread.

Bottom line, bug or not, there is headroom here for dev to enable that second thread officially.

[rdluffy]

It's definitely a bug. I disabled 4G Decoding, BAR in the BIOS, rebooted, DDU'd the drivers, reinstalled drivers, tested without Smart Access Memory, then tested again with it back on and now back to 1 GPU still.

Looking to see if something else I was doing caused it to pop two instances of the miner. The limited time I was running, it was in fact getting 80mh at the pool and had both compute 0 and compute 1 under load in the GPU monitor; now I just get compute 0 with the single thread.

Bottom line, bug or not, there is headroom here for dev to enable that second thread officially.

Never thought that is possible to miner recognize 2 GPUs like this, it makes no sense 

But now, with miner recognizing only one GPU, what's the hasrate? 62 to 65mhs as in your first video?

And maybe you found a bug and devs could use this approach to modify how miner works, at least with new AMDs cards

[hellscreamer]

Imagine this working with 3080 , 5GB is just right

[wishxy]

I've got no positive results on RX 470 4GB rig even I tried all different miners including the new lol miner 1.16.
I haven't change the old blockchain driver and the result is still poor, max 10 Mh/s for the whole rig.
Now I'm waiting for ETC Thanos fork at the end of November.

[jgonzi]

I've got no positive results on RX 470 4GB rig even I tried all different miners including the new lol miner 1.16.
I haven't change the old blockchain driver and the result is still poor, max 10 Mh/s for the whole rig.
Now I'm waiting for ETC Thanos fork at the end of November.

But this is normal... the DAG file is too big that it cannot be all loaded in mem... you have been using the Zombie mode that allow you to mine but with lot less Mhs... at least you were able with this miner. If you want the best improvements mine ETH until the fork of ETC. That will give you better performance until the change.  It is only a recommendation...