Closed MonaxDK closed 6 years ago
That wasn't my experience with my 980GTX, for what it's worth - https://github.com/nicehash/NiceHashMinerLegacy/issues/235
i dont know what happned
but, 1.8.1.2 my profit is $usd/day 2.49 ( half hour test) and 1.8.1.1 my profit is $usd/day 3.60 ( half hour test)
i checked again, same results.
i have one gtx 1080 ti 11 gb. windows 10, nvidia driver 385.28, 16 gb ram, i7 first generation
overclock none.
On Sat, Sep 2, 2017 at 9:30 PM, MonaxDK notifications@github.com wrote:
Apparently, ccminer_tpruvot is not suitable for older video cards. GTX 970, NHML 1.8.1.1 [image: 970-1811] https://user-images.githubusercontent.com/29005588/29999608-d78caaf2-9057-11e7-8082-e83c3f93d1b4.png GTX 970, NHML 1.8.1.2 [image: 970-1812] https://user-images.githubusercontent.com/29005588/29999613-04e00274-9058-11e7-8870-b1c364735d62.png
But for 1080Ti real miracles happen :) 3x1080Ti NHML 1.8.1.1 [image: 1080-1811] https://user-images.githubusercontent.com/29005588/29999615-17100cdc-9058-11e7-8a36-8d8bbaab9f95.png 3x1080Ti NHML 1.8.1.2 [image: 1080-1812] https://user-images.githubusercontent.com/29005588/29999616-1fb6a79c-9058-11e7-867a-f664ea771e87.png
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/nicehash/NiceHashMinerLegacy/issues/261, or mute the thread https://github.com/notifications/unsubscribe-auth/ABf1HXxUPn5S47_wmkJxrzMx_rCts4Ueks5sefMLgaJpZM4PLDJF .
You can't really go by USD - the price of BTC/crypto is so variable.
Has your hash rate changed?
I got two EVGA GTX 980 Ti with 1.8.1.1 I had about ~0.88 MH/s on each. with 1.8.1.2 I have about ~0.75 MH/s on each.
please, use correct miner for NVIDIA 9xx.
Sorry tpruvot is giving you less performance on 900 cards, I'll get this fixed in next version
As a workaround for now you could copy ccminer_neoscrypt over the tpruvot ccminer.exe, but that could hamper other algorithms
Not a big problem, neoscrypt is never the most profitable algorithm on 9 series anyway. In fact, given the difference in performance on 10 series cards, it is very interesting if other algorithms can benefit from the use of tpruvot?
Gigabyte GTX 1070 G1 Gaming
i replace miner in folder NHML-1.8.1.1\bin\ccminer_neoscrypt to KlauT-ccminer v.8.13
in version 1.8.1.2, I get a much smaller hash
in version 1.8.1.2 the possibility of substitution of a miner in a folder was disabled
maybe it's worth making a customizable path to the right miner?
well Klaust's ccminer gives me better hashes too, for about 10%.
I thinks there should be multiple miners for Neoscrypt, so users can use Benchmark and NHML will choose the best solution for everyone.
and add one more screenshot with ccminer-alexis78...
@adem4ik Yeah that looks like a real good solution especially with multi GPU setups in mind.
Perhaps it is worth having the option of selecting a miner for each algorithm, not just for Neoscrypt? Given that unexpectedly on another miner you can get 4x hash rate...
That would be great. Equihash now has a choice of a couple or maybe more? EWBF rocks, for me, 350+ sol/s on a 980.
And all these things are why we use NHML and not NHM. NHM gives you no choice but to put up with lower-profit mining. NHML allows you to choose the algo and device.
In the next version, there will be .JSON files generated by NHML that you can edit to change which ccminer fork is used on each device family
Thank you for reporting faster miners though!
I think it is better to make it possible for every miner to test all the algorithms available on the NH pool.
and then, according to the results of the benchmark, you will see which miner on which hardware is the best
sorry for my bad English :)
Here is an example of the result of the NemosMiner-v2.1-Zpool benchmark for 2x GTX 1070 sorted by the current profitability
It would be fantastic if NHML had such a benchmark. It's very difficult to manually test all the algorithms on all miners ...
I do not do benchmarking of every video card on all the rigs in NHML, I do a benchmark for one video card on one of the rigs, and then I copy the results to the remaining .json files in all the rigs.
If such a benchmark is difficult to integrate into NHML, perhaps it's worth creating a separate topic, and invite users to share the results of their benchmarks in Zpool? Thus, it will be possible to collect statistics which miners work faster on certain video cards. Here are my results for 3x 1080Ti Not all miners / algorithms have been tested, because Zpool is buggy :)
Apparently, ccminer_tpruvot is not suitable for older video cards. GTX 970, NHML 1.8.1.1 GTX 970, NHML 1.8.1.2
But for 1080Ti real miracles happen :) 3x1080Ti NHML 1.8.1.1 3x1080Ti NHML 1.8.1.2