No. Fix from AMD is required, we have to wait it. Or I will have to do their job again and find some workaround.


It's different, a bit better.

Ok I'll check it.

v11.1

- fixed issue with stale shares, now effective hashrate must be higher.
- a few minor improvements.

PS. Linux version will be available in a few days.

Yes I will do it but I did not promise any timeframes, right now I'm busy with XMR miner update (bug fixes mostly), I will release it very soon.
Then I will update dual miner because some things must be updated.
After that I will update this miner and this update will include less devfee as promised.

Check GPU load in gpu-z (must be 100% all the time).
I will do some tests in 1-2 days to confirm that the miner produces correct number of shares, though I already did it for v11.

UPD: It seems I found the reason, when pool sends new job, v11.0 drops old job too late (loses 1-2 sec). I have fixed this bug but it will take 1-2 days to test it properly and release an update. However, you can test a fix right now if you want and tell me if it helps:
https://mega.nz/#!TlYQAJSZ!291Swk4GbJLC5JYmax_fc-tLWjfwKsFTnOUSVGqsDlA

For most cases, miner applies the best "-h" and "-dmem" values automatically. But you can find the best value manually:
1. Use "-dmem 0" and try several "-h" values to find the best hashrate. For example, you found that "-dmem 0 -h 512" is the best. It will require about 1GB GPU RAM.
2. Now set "-dmem 1" and divide "-h" value by two: "-dmem 1 -h 256". It will require about 1GB GPU RAM too.
3. Also try "-dmem 1 -h 512" if GPU has enough RAM, it will require about 2GB GPU RAM.
Note that using almost all available GPU memory usually causes slow working because drivers allocate memory incorrectly, reduce "-h" in this case.
In my tests "-dmem 1" must be used for all GPUs for best speed, except Vega which is not optimized properly yet.

Hmm, in my tests v11 is a bit faster for both Tahiti and Pitcairn, or at least the same. May be you use not optimal settings.


Sorry, I don't not support very old cards anymore.

v11.0:

- improved speed by 1-3%, depends on GPU model.
- fixed issue with invalid shares.
- added CryptoNight-Lite support, use "-lite 1" option to mine AEON.
- improved stability.
- adjusted default "-h" values for various GPUs for better performance, check your settings.
- fixed critical issues in remote management feature (attacker could crash miner even in read-only mode).
- fixed issue with possible GPU crash during miner shutdown.
- updated pools for devfee.
- several minor improvements and fixes.

Currently I don't have enough time for it, sorry, I'm too busy with other miners.

I also added CryptoNight-Lite support (AEON), it took some additional time. I will release new version within 48 hours.

Note about latest drivers version: I confirmed that AMD broke fan management again.
I have contacted them and hope they will fix it soon. Right now don't use 17.12.1 drivers if you want to manage GPU fan/temperature.

Miner uses the list of devfee pools if you specified "-allpools 1" option, it selects pool randomly, so it will find a good pool after a few attempts anyway, you don't lose anything because miner mines for you during these attempts.
In next update (2-3 days) I will update this list of pools and exclude nicehash.

In 2-4 days I will release new version with a bit better hashrate, improved stability, also I have fixed the issue with invalid shares on some cards.

Don't worry, this ticket was created on September 26 and still no response, so I don't care anymore...

The most recent is "Ticket N55965".

It seems this pool has wonderful level of support - all my tickets hang for two months with no response. Other pools answer my questions much, much faster.

v3.9

- fixed issue with devfee failover.
- fixed critical issues in remote management feature (attacker could crash miner even in read-only mode).
- added "-mpsw" parameter.

Ok I will check it and will probably add in next update.

Ok, it will be improved in next update.