1.9.0.0 - Alpha thread

[DillonN]

Please post issues related to the 1.9.0.0 alpha here!

Known issues:

• Parallel benchmarking
  • May not work properly with specific miners such as dtsm, Prospector and ClaymoreCryptonote_old. If having issues you can force benchmarks to run sequentially or manually benchmark.
  • REMEMBER: For miner programs that haven't been updated, you can copy over your old benchmark files and they will already have speeds entered in your new install
  • On certain miners may immediately terminate if multiple start at the exact same time. Just retry and they should go through
  • The three dots that increment in the benchmark list on the current algo/miner may not update correctly

[Smeagol001]

somehow tzhe app isnt displaying the statistics anymore but looking on the cmd windows i can tell its working

[Pjalchemist]

Windows 10 x64 Pro, build 1709 Updates disabled Defender disabled

When downloading via Chrome on the laptop, Windows goes nuts about virus.

Installing the miner no issues, when benching it crashed the moment it starts to bench. Have tried reboots, right click run as admin.

[JCarangoH]

There is an error in exchange value in 1.9.0.0 version, look:

1.9.0.0:

1.8.1.10:

Both screenshots were taken at same moment...

[MaKCuMyC]

@DillonN why you ignore right name of DSTM equihash miner? it is DSTM, not DTSM.

[nerdatwork]

Why isnt the list of miner names screen resizable yet ? It has been forever and tedious to just scroll left and right when you got a huge screen.

[paulh69]

seems to have problems identifying or indexing devices?

[paulh69]

not sure how useful power setting would be unless set automatically, difficult to access settings from benchmarking screen and it seems doubtful that the power efficient algorithms would pay enough to make it worthwhile

[luky76]

hi, with new alpha version my system reset it when i try to benchmark somes of algorithms. only Equihash and Neoscrypt did successful. Whith old version work ok. i got w 10 , ssd , gtx 1070 and 650 w psu.

[paulh69]

something is causing a crash and it wipes the contents of all .json files? they’re all still there but empty.

[luky76]

benchmark algorithms one by one, just one at a time. no crash, no stop , no errors. all ok and work

[Smeagol001]

i suggest making it possible to just specify how much load on gpu/cpu can be allocated to nicehash in general like i want atleast 2 cores from my cpu to be free fro use of other programs since otherwise my os starts to lag and i cant even use the mouse same for gpu make it so that at least like 2-5% wont be used like with a chekbox or something for globally affecting all miners atleast if they support something like this since 1.9 i cant even do a spreadsheet while runnning nicehash :(

[shinnsd]

I've been getting rejected shares (a lot) from cryptonight. appearently it's a compatibility issue between xmr stak and the nicehash server (no issue if I use cast xmr). Looks like there is a new version of xmr stak that addressed this issue:

https://github.com/fireice-uk/xmr-stak/releases/tag/2.4.1

but he mentioned “workaround” so I assume it's not a complete fix.

Can you please look into it? Thanks

[boby361]

My 280x won't complete either equihash benchmark but my 7950 completes them fine. Something is up.

[OiHeldeke]

1. Compared to 1.8.2.0 pre2 both xmr cpu miners seem to be slower, especially xmrig. Older version was able to deliver ~59 H/s per thread, new does 48-49. This is on my slowest CPU. On faster CPU-s difference is smaller, but still some percent slower. Edit: xmr-stak performance looks fine, but xmrig is as I wrote.

2. If I go to Settings->Devices/Algorithms->Check All Profitability (or check single), then it tries to open page https://api.nicehash.com/?utm_source=NHM&p=calc&name=CUSTOM&speed22=0.00&nhmver=1.8.1.10&cost=1&power=1 that does not exist.

[paulh69]

also gets stuck on “Unable to get Nicehash profitability data. If you are connected to the internet, try again later.” does not retry automatically

[paulh69]

also gets stuck on “Unable to get Nicehash profitability data. If you are connected to the internet

[paulh69]

benchmarks don’t always run, comes up with “benchmarks successful ok” several times

[AtlAntA118]

Not really related to this specific release but it would be such an awesome feuture to right click on an algorythm and select "disable/enable on all cards"

[Smeagol001]

suggestion make all the collumns in the listviews sortable

[Smeagol001]

all sgminer benchmarks exept x11 fail and want me to use precise mode despie having used precise mode im on a amd rx580 and ryzen5 1600x

[Spin55555]

Do you plan to implement the latest version of Claymore Dual Ethereum+ GPU Miner with keccak algorithm support? At the moment it is the most profitable combination.

[techman05]

Maybe its been mentioned but the stop button doesn't stop all. I went to benchmark after it running for a while and the xmr cpu app kept restarting even with the window closed and had to kill a nicehash process to stop it.

Also probably mentioned but the cryptoV7 is in cpu and gpu but only the cpu is running it. I'd hope if the miner is implemented I could somehow trigger it for the gpu.

Thanks

[pbuva]

1 BTC = N/A USD often applications do not download data from the server and then evaluate mining as unprofitable.

EDIT: After 5 days, no answer, no sollution, still cant get profitability data for all rigs.

[jmhickman]

I've had great results with the software so far. Since we're in a testing phase, can we add the newest CastXMR to the suite? The latest version supports vastly higher cryptonight/v7 rates without having to use the stupid blockchain driver. http://www.gandalph3000.com/cast_xmr/cast-xmr-optimized-cryptonight-miner-for-rx-vega/

[bmassy]

Cannot Mine Cryptonightv7 with AMD cards. "AMD device not found, Backend AMD disabled, No miner backend enabled" But mines perfectly fine on my GTX 1050ti..

[bmassy]

So I deleted the bin folder for that miner so nicehash would re-download and setup the miner again and it worked.

[CryptoVaraan]

@DillonN I see many comments here, but no reply from you? While i wonder if it helps to post more... The displayed value in EUR is 10 times too high for Global rate and Balance in this version... Have a nice day :)

[YuriyNa]

Разработчик похоже забросил проект. Жаль, на Найсхеше теперь на AMD не помайнишь((((.

[YuriyNa]

Алгоритм NeoScrypt на картах AMD не работает. Не проходит бенчмарк, при запуске вручную окно майнера открывается и сразу закрывается. Проверено на разных картах на трех системах.

[Bathmat]

I am getting a lot (5-10%) of this error: AMD Invalid Result GPU ID 1 when using xmr-stak for V7.

This issue was brought up here: https://github.com/fireice-uk/xmr-stak/issues/1337

And is potentially resolved by: https://github.com/fireice-uk/xmr-stak/pull/1437

So perhaps releasing a newer version of xmr-stak might solve the issue.

[dandawn13]

Whenever I put value to electricity cost, Nicehash will only mine 1 GPU and shows the net profit of that. Other GPUs will not work until I put the electricity cost back to 0.

[DillonN]

Sorry for the delays everyone, April was busy.

@Smeagol001 has this been an intermittent problem? If I could get the log file for a period when this happened could help greatly.

@JCarangoH this issue has been fixed in code and will be in next update. Was related to differing decimal separators (comma/period)

@MaKCuMyC actually surprised I had missed that, brain fart there

@nerdatwork size of the main window is managed by code and difficult to make resizeable without a lot of effort. I probably won't be doing this soon because in the future I want to overhaul the main window somewhat (that is a bit further out)

@Smeagol001 intensity is determined by the miner programs and each one has their own way of doing it. A way to edit this from within NHML in a way that it abstracts all these different settings has been talked about but is not currently planned. Also the wiki currently leaves much to desire in explaining this. As it stands you can specify how many threads xmr-stak is using by editing the file bin\xmr-stak\cpu.txt and removing thread entries until you have the amount you want to use. Each thread entry looks like { "low_power_mode" : false, "no_prefetch" : true, "affine_to_cpu" : 10 }, but will have different values for affine_to_cpu which determines which physical thread they bind to (if you have HyperThreading make sure to only use every other value so you get a physical core and not an HT).

@shinnsd this is indeed due to an incompatibility with xmr-stak and NH servers, there is a more robust fix in their newest version which will be included in next release

@OiHeldeke 1. CNV7 still needs some time to catch up but they are making it better with every release. 2. The pages were changed server side and new links will be in the next version

@AtlAntA118 that is an interesting idea. I've been thinking of implementing groups for devices so that you only manage one list of algorithms/speeds (e.g. if you have identical cards) so if that is your use case that would help.

@Smeagol001 I am looking into sgminer instability, which seams to be very bad when remoting into the computer. This is a known issue in sgminer.

@Spin55555 yes

@jmhickman we are in the fix-and-don't-add part of this test phase but can certainly look into CastXMR next

@dandawn13 mining on specific GPUs will not commence if their profit is 0 or negative after deducting the estimated power costs

In general:

Problems with benchmarking/miners not stopping are attributed to one of the known issues which is fixed

Problems with connectivity are not exclusive to this release and due to a huge number of factors. We continue to strive to make these issues as little as possible on the NHML side

TLDR

• Benchmarking is mostly fixed

• CNV7 will be fixed with new miners

• Exchange rate problems are fixed

• ClaymoreDual w/ keccak will be supported

• sgminer is still problematic (not new to this rel)

[ryangray0]

RE: Power consumption / profit: This is going to be a great feature.

Here is what I see: System benchmarks and runs fine with no values in for energy cost, or energy consumption in algos. When I add an energy cost in settings and populate all algo enrgy consumption values for all cards, no miner initiates when I click 'start'. If I remove the energy consumption value from any algo, that algo will initiate when I click 'start'. If I leave all energy consumption values in algos, and change energy cost to 0, system runs normally.

I apologize if this was covered in another post, this is my first contribution to the effort. Great product!

[DillonN]

@ryangray0 there is currently no auto-filling in for power usage - you must fill it all in manually.

When you hit start mining, NHML searches for the most profitable algorithm after subtracting power costs based on the values you entered for power usage and price per kilowatt-hour. If no enabled algos are above 0 profit after this subtraction then none will start

If you think the profits should be above 0 then you can forward me the logs and I can take a look to see if something is up

[dandawn13]

@DillonN we're talking about the new feature "electricity cost" and not the profit margin. It doesn't work on multiple gpu. Only works when a single gpu is enabled.

[clizmax]

@DillonN why in the version of nhml 1.9.0.0 alpha at the start of the program pops up the window "can not get nicehash profit data". this began to appear when version 1.9.0.1 was released. especially if you did so ?? and I do not want to switch to the new version, because Miner xmrig was removed, it yields 6 times more than xmr-stak ... return everything as it was, please

[S74nk0]

Issue related to old nhm version/build. Please use latest version and open a new issue for any bugs or feature requests.