Topic: PhoenixMiner 6.2c: fastest Ethereum/Ethash miner with lowest devfee (Win/Linux) - page 462. (Read 784954 times)

you misunderstand me - 0.35% is the difference between Phoenix Miner and Claymore - that difference IS significant

oh yeah sorry, I misread the message.

YES, my point was paying fee is better.


it is your problem you cannot inquire someone about your private miner and have a stable hashrate with no fees.

Because you just replied to a post directed at Claymore acting like it was at you. Easy enough conclusion, if far fetched in reality.


No worries, as I said the only difference I can give is that I used the latest ISO for Win10 from MS, but then all the other machines are fully up to date AFAIK, and we have 2 other builds with the exact same hardware (built from ISO from around June 17). The machine itself has crashed/rebooted a couple of times over the last day so maybe it's a problem with the mobo. Will visit the mine later and see about copying bios settings from one of the fully working ones incase some arcane setting balls deep in the menus got missed!

wait... w0lf is claymore?

I pay 1% fee.
If I disable fee I loose about 4% hashrate (from 30.5 to 29.3. Dualmining).

I havent tried phoenixminer though.

Phoenix, is your 2.7B optimized for the latest AMD drivers ? It seems slower than the 2.6 with my 2 RX570 and blockchain drivers.

Do what you've got to do of course - obviously you feel strongly about this and believe that this is well worth your time It's just not exactly clear what you are trying to prove though - we already conceded that our old/honeypot kernels might be the same as your old kernels but this is not very likely. But since you are willing to provide the blobs of your kernels in these PM sessions, our kernel guy would certainly like to take a look, so thank you.

Frankly we've got a lot of respect for you, and if you have chosen to contact us via PM with your concerns, we would have respected that and checked if there is any merit in your claims but you've come swinging and accusing us in our thread and if you expect us to just sit there and take it, you are quite wrong.

It is also quite wrong of you to assume these things:

• That you can't be fooled by our anti-reversing measures (really do you think that we can't detect dll-injection attacks?).
• And that, if you do succeed in extracting our real/current kernels and leak them in the open, we can't do the same with yours. "We know" a guy that was part of the driver development team of AMD and he is itching to test its kernel-level emulator by trying to extract your current kernels. Heck, if he's successful, we can even put your kernels in our miner and let the users select them explicitly and see for themselves how they compare with ours.  

Anyway, we prefer to concentrate our efforts on improving our miner bit by bit (we are still lacking in many areas) but if you want to play this game, we can play it too. Just keep in mind that we were nobody in comparison with you just a few weeks ago, and you seem resolute to put us on the map so to speak, for which we thank you - it well known that controversial or bad publicity is always better than no publicity at all.  

disabling the fee ("-nofee" command) will lose "optimizations" according to claymore.. and it does, i've tried.

if you use the -nofee option, miner becomes slower for about 1MH/s ... i can confirm also.

I have significant hashpower and believe me, 0.35% makes a substantial difference at current prices

how come 1% is more than 1Mh/s ? what is wrong with you?

I disabled fee in Claymore. Lost 1Mh/s on each gpu.

So it' s worth to pay the fee.

What a drama coming!

For those who wants Claymore to lower dev fee - suck it up dudes, he makes constant developments of miner things and he's the only one who always deliver.

If you are so greedy and for you 0.35% makes a huge difference (sick joke, seriously), it would be better to throw yourselves out of the window.

I am 100% sure those who have industrial mines have their own miner software and not have any problem with comission.

So, most of baby cryers outta here have 1 rig with 300$ monthly income -- 3$ per month for actually MAKING this income possible -- it is too much for you guys?




Coming back to reverse engineering thing, I do not appreciate such kind of actions. Hope it has nothing with stealing Claymore's code.

Have a nice day everyone!

I decided to check your miner as soon as I noticed that you have same tuning option ("+" and "-" keys in runtime) that works exactly as my -dcri option, also your miner has no "optimized" kernels for same set of chips as my miner. After checking I found 100% match in GPU kernel on GPU.
Now you are talking about some "old unknown" kernels and you don't use them anymore and have super-protection that can fool me and so on... okay.
I will create step-by-step guide how to confirm that you send my kernels to GPU and you will say that your miner uses "old unknown" kernels because it detects dumping but still mines at the same speed. That's funny answer, but you don't have better one, I understand.
So my word against yours. The result is obvious for me, however, since I call you a liar, anyone who wants to confirm by yourself that this miner uses my kernels can PM me and I will send step-by-step guide how to do it, then you can say the results; no newbies please, be at least sr.member on this forum, just to be sure that you don't want to become another "phoenixminer" and spend my time again...

   It's definitely not a hardware problem - this is the anti-debug code activating without good reason apparently. We are trying to make these checks smarter and lighter but obviously we aren't there yet. Still, we are hesitant to relax the checks too much for obvious reasons - apparently there is no lack of highly esteemed developers trying to reverse our miner
   
   Not yet, unfortunately, it is proving to be harder that we expected, so we better not throw around optimistic schedules without delivering.

    Thank you for the logs. Unfortunately we can't determine much from there just that the GPU thread is freezing after the crash. You've got quite a bit of memory overclock though maybe try dialing it back a little. We also had problems with Vega because of the card high power consumption - even 600W quality PSU led to sporadic crashes when powering a single Vega56 even with big underclock. It seems like Vega has high peak currents that lead to crashes if the PSU is not ridiculously overpowered. We have some work to do with hardware controls support for Vega though, as it appears to use some new (and undocumented) ADL functions and structures.

   Stale shares. The high mining intensity increases the time window within which a stale share can occur. Therefore the default is 10 even though 12 gives a slightly higher hashrate. Still, the stale shares are sometimes dominated by the pool latency so you may see a better results with -mi 12 but keep an eye on stale shares to avoid losing more via stale shares than you have gained from the higher hashrate.

   Yes, but only the latest beta version which can be found here: https://bitcointalksearch.org/topic/m.31091660 The last official release (2.6) doesn't support these options yet.

  It isn't 

   Yes, you can. Start a two copies of PhoenixMIner and add the -amd command-line option on the first, and -nvidia on the second. The first will mine only with the AMD card and the second - with the Nvidia card.

   Well, actually you don't profit at all from the higher difficulty shares - it only matters that the share is above the difficulty specified by the pool. There is nothing sinister there though - this is the reason the pool exists: to equalize the luck of the individual miners according to their actual hashrates so it doesn't matter who actually finds the block.

   No, we certainly DO NOT confirm it. We believe that you've tried (and apparently you think you succeeded) in extracting our kernels using reverse engineering techniques. That wasn't very nice of you to say the least but we have expected something like this, just not that fast. In some sense it is flattering but also a little annoying.

   The first version of our kernels was obtained from someone on this very forum for quite a bit of coins. They were a little bit slower than the best miner on the market at this time - yours - but we decided to use them to jump start our project anyway. After a lot of tinkering with Polaris GPUs we found out how to increase the speed a bit and improved the kernels a lot, achieving higher speed and/or lower consumption than your miner (in most cases). The original versions were left as a honeypot for particularly persistent reverses which apparently did its job quite well

   Now, are these old kernels the same as old versions of yours? We don't know for sure but we doubt it, and we can't just take your word for it (especially after throwing accusations like this after basically admitting that YOU are trying to reverse OUR miner), and because they were slower than your miner, and they do not support dual mining at all (and dual mining is still the killer feature). But even if they were, frankly this isn't our problem - we've paid handsomly for them, and we haven't used them in any of the publicly released versions, so we won't lose any sleep over this. If you beleive so, search for a digruntlet employee, or perhaps your unnamed source(s) at AMD?

   Last but not least, we take offense in your description of our miner as a "wrapper" around the kernels. While the kernels are very important, the miner is wrapper no more than a car is a "wrapper" around the engine's crankshaft The source code of the whole miner is two orders of magnitude larger and more complex than the kernels, so we beg to differ here in our opinions. We may have less than percent of your market share but we are proud of our miner and (after we finally crack dual mining, and release Linux version   ) we believe that it is going to be the best in the market.

i was having this also, for me it was related to some OC settings.

@Claymore competition is always healthy for every party.. despite if PM dev is using Claymore's code, he should give him credits for it at least..

Hitting with this error for the past 2 days, mining ETC on ethermine.
Both gtx 1060 and 1070 no hang, i still can see them using after burner.
Anyone hitting with this kind of errors?

17591:07:20:22.496: main Eth speed: 79.625 MH/s, shares: 47/0/0, time: 0:46
17591:07:20:22.496: main GPUs: 1: 18.465 MH/s (9) 2: 31.832 MH/s (21) 3: 29.329 MH/s (17)
17591:07:20:25.706: eths Eth: Received: {"id":0,"jsonrpc":"2.0","result":["0xeb77fae561c0afcd983a8ff0a1c8372dbead6d4264853049fc44725f2dcbfa7f","0x5683748fa7f53897f1e1a80a8a321d3693761838135be37d1d132c0cd3d7d937","0x0112e0be826d694b2e62d01511f12a6061fbaec8bc02357593e70e52ba","0x5382a3"]}
17591:07:20:25.706: eths Eth: New job #eb77fae5 from eu1-etc.ethermine.org:14444; diff: 4000MH
17591:07:20:27.512: main Eth speed: 78.413 MH/s, shares: 47/0/0, time: 0:46
17591:07:20:27.512: main GPUs: 1: 17.995 MH/s (9) 2: 31.103 MH/s (21) 3: 29.316 MH/s (17)
17591:07:20:29.897: GPU2 CUDART error in CudaProgram.cu:422 : unspecified launch failure (4)
17591:07:20:29.897: GPU2 GPU2 search error: unspecified launch failure
17591:07:20:29.948: GPU1 CUDART error in CudaProgram.cu:422 : unspecified launch failure (4)
17591:07:20:29.949: GPU1 GPU1 search error: unspecified launch failure
17591:07:20:32.334: eths Eth: Send: {"id":6,"jsonrpc":"2.0","method":"eth_submitHashrate","params":["0x4bdeed8","0x94d6170da9ae5da346e6ef1432826c5f7c8d2db8c37ea7de36306d9565f93789"]}

17591:07:20:32.526: eths Eth: Received: {"id":6,"jsonrpc":"2.0","result":true}
17591:07:20:32.544: main Eth speed: 79.620 MH/s, shares: 47/0/0, time: 0:46
17591:07:20:32.544: main GPUs: 1: 18.463 MH/s (9) 2: 31.829 MH/s (21) 3: 29.329 MH/s (17)
17591:07:20:37.575: main Eth speed: 29.327 MH/s, shares: 47/0/0, time: 0:46
17591:07:20:37.575: main GPUs: 1: 0.000 MH/s (9) 2: 0.000 MH/s (21) 3: 29.327 MH/s (17)
17591:07:20:38.346: eths Eth: Received: {"id":0,"jsonrpc":"2.0","result":["0x020e766278c1681f610557f3dfcc9dbf2713e80daafe51c570d60cf53b50a20d","0x5683748fa7f53897f1e1a80a8a321d3693761838135be37d1d132c0cd3d7d937","0x0112e0be826d694b2e62d01511f12a6061fbaec8bc02357593e70e52ba","0x5382a4"]}
17591:07:20:38.347: eths Eth: New job #020e7662 from eu1-etc.ethermine.org:14444; diff: 4000MH
17591:07:20:38.350: GPU1 GPU1: Starting up...
17591:07:20:38.397: GPU2 GPU2: Starting up...
17591:07:20:38.431: GPU1 CUDART error in CudaProgram.cu:230 : all CUDA-capable devices are busy or unavailable (46)
17591:07:20:38.431: GPU1 GPU1 initMiner error: all CUDA-capable devices are busy or unavailable
17591:07:20:38.469: GPU2 CUDART error in CudaProgram.cu:230 : all CUDA-capable devices are busy or unavailable (46)
17591:07:20:38.469: GPU2 GPU2 initMiner error: all CUDA-capable devices are busy or unavailable
17591:07:20:42.592: main 
17591:07:20:42.592: main Eth: Mining ETC on eu1-etc.ethermine.org:14444
17591:07:20:42.592: main Eth speed: 29.332 MH/s, shares: 47/0/0, time: 0:46
17591:07:20:42.592: main GPUs: 1: 0.000 MH/s (9) 2: 0.000 MH/s (21) 3: 29.332 MH/s (17)
17591:07:20:42.592: main Eth: Accepted shares 47 (0 stales), rejected shares 0 (0 stales)
17591:07:20:42.592: main Eth: Incorrect shares 0 (0.00%), est. stales percentage 0.00%

Higher difficulty shares means more profit reflected on your earnings from the pool, I do not think there's a way to regulate it, I think it's just simply luck, sometimes you solve higher diff shares, some other times you solve lower diff shares and, sometimes you solve nothing

I might be wrong tho, this is my understanding so far, havent come up with a good read on how the pools handle difficulty.

Guys, I have seen that the miner finds shares 4950MH and goes up to 928.2GH, have seen also results that go to 1TH+.
What is the meaning of this and how to regulate it, if I have to?
Could anyone give me an advice, please?