Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 1001. (Read 2347601 times)

@djm34, @pallas, @sp_ your VTC addresses please. It will be easy for me to send VTC rather than BTC. For BTC you have to wait a bit longer when I have accumulated enough VTC & sold them.

@ scryptr

what diff do you use with Give-Me-Coins pool?

also if I was to mine solo what would be the recommended diff setting ?

thx

RELIABLE HASH RATES--

I don't know what is happening at some pools, but I am receiving reliable hash rates and payouts at Give-Me-Coins (GMC).  With Coinotron and Verters, GMC is a top pool.  They have 1/4 to 1/3 of the network hash rate for VTC on a regular basis, so blocks are found at the pool often enough for a steady payout for even a single-card miner.       --scryptr

Threw some rigs at verters to check, each doing ~25.5 mhs as per ccminer.
Pool reported hashrates with;
diff 1 - 20-22 mhs
diff 1.5 - 30-33 mhs
diff 2 - 31-39 mhs
Anything above 2.0 will cause rejects.

Verters is mining 43% of all the blocks and they can be abused with -diff 2. Or more precisely, they could be abused if they weren't so inefficient at finding blocks, but as it stands that 31-39 mhs translates into 19.3-24.3 "real" mhs due to their terrible luck at finding blocks.

actually on p2pool there are even more option to adjust the difficulty (through password).
It would be better to get fewer low diff shares actually (that would obviously screw the hashrate report) by setting an higher difficulty so that the miner doesn't lose time in sending worthless shares as for p2pool only count for the payout are shares which are above a certain difficulty target.

also don't forget that on p2pool, it takes time to get the full payout however those higher diff shares last several blocks...

ok - so the hashrate status is merely an 'issue'? ... not a problem? ...

because the coins are being mined and deposited into the wallet ...

i will recompile the latest git and have it running tonight - as im off to bed after that ... need much more rest ...

tanx djm34 ...

#crysx

ok - will recompile now ...

has been mining for a while now - and there have been a few coins coming through ... so its not all wrong ...

tanx sp ...

#crysx

p2pool estimate aren't reliable in any way and actually it doesn't even matter as the payout is totally independent of the # of low share you get (if you run your own node at least), the only things which matter are shares above a certain difficulty (which are the real one)

you need to rebuild with the last commit. The "undo pool fuck" commit

Use --diff 2. and mine for 12 hours.

hi again ...

ive just come back on to see how things have been going and found that p2pool is still showing nowhere near the amount of hashrate that the miners are showing ...

what is happening here? ...

the hashrate on the pool is literally one tenth of the hashrate that the miners are showing ...

can anyone explain to me what this could be? ...

for example - one of the farm miners is showing about 24MH but on the pool - the TOTAL hashrate is about 20MH - and I have a few miners on the pool hashing away for this test ...

i cant understand why there is such a MASSIVE discrepancy in figures ...

#crysx

best thing to do is ask nicehash to adjust their setting...

Still very low..

Remember that lyra2re was profitable with a miner that was 6 times slower.

You can see from the graph that the current net hashrate is lower than it was with the previos algorithm

If the difficulty reach 800, the new algorithm will still be 3 times more profitable than the old algorithm. (on the 750ti)

Difficulty have been increase to 298, Every 24 hours increases ~%10-15

Nicehash is currently only displaying half the reported hashrate in the miner when I run with default difficulty(as in the official djm34 release). If you change the diff, no accepted shares...

This meens that nicehash is only paying 25% of what you get on p2pool.pl and --diff 2 (sp-mod release 60) or when solomining...

Another interesting thing. The global hashrate of lyra has dropped with the new algo and the botnets gone:


The new GPU miner is 600% faster than the old one... (on the 750ti)

After 30 hours of mining I've got around 16($0.96) Vertcoins a day per 750ti (recalculated to 24h average))   (p2pool.pl with --diff 2

Nicehash is currently paying around $0.41 per 750ti per day.

So someone is renting your hashrate and make money...

Note that the difficulty has raised from around 230 to 282 (+10.9%) in 24 hours and the next payouts will be smaller.


Results 1 rig (6x750ti)

tanx sp ...

i totally agree with you when it comes to pools ripping the miner off ...

i just thought i would bring that part to your attention ...

i have just received word from yaamp that he is no longer able to build our multipool now - as he is so involved with the new projects he has his interest in ...

so we are now looking for someone that can do that for us - and with us ... if there is anyone interested - please pm me ... im happy to work with whoever is interested - in many many ways ... and there are a few prerequisites for our pool - as i have a few ideas that i would like to see implemented within the pool itself ... yaamp could do this - but as said earlier - is too busy with other projects ...

we want to build a fair and equitable multipool ( which was one of the backend projects i have on ) which the community can trust ... so im open to suggestion with whoever would like to work with me ...

as for lyra2rev2 - ive pulled the miners off mining that algo for now ... at least until i get a lot better than what i am now - so i can trouble shoot and 'play' with the miners ... at the moment - i just get up occasionally and check in here for a few minutes - then get back to bed ...

ill test your suggestion with p2pool.pl and the --diff 2 parameter a little later today ...

ive started working on the donation servers again ... if i keep going at this pace - they will be active by the weeks end ...

tanx again sp ...

#crysx

I submitted some debug output. not on purpose. I was getting half the hashrate on p2pool.pl so I changed the difficulty multiplier to 2 as default (hardcoded into the miner) The pool fuck comment. (fuck the currupt pools). has now been removed
and get a full hashrate displayed on the pool. I have now reverted it, because Nicehash doesn't work.

If you build between releases, you might get some bugs in your miner. When I flag a release number, it meens that the code is safe to buid,

for full hashrate on p2pool.pl use the --diff 2 parameter

hi all ...

ive been quite ill for the last week and am still not very stable health wise - but enough to get some things done then back to bed ...

so ill continue to pop in and out of here - to try and keep up to date with what is happening ... and a LOT has happened in the last week ...

i have compiled the latest ccminer-spmod v60 - and set the miners running on the lyra2rev2 ... initailly on verters for a few hours - and now on nicehash ...

a few things popped out which seem to be issues ...

- hashrates are NEVER correct ... even after a few hours - --diff changes and paramters ... miners are showing more than 10 times the pool hashrates ...
- not all exchanges have caught on - and i guess they will soon enough ...
- the latest ccminer ( compiled earlier today ) have had some changes which are nice and some which are questionable ...

the yay! has been changed to a yes! ... yay for that - or not ...

but when the miner cannot connect to the pool - the message is far from standard ... a copy below ...

-------

*** ccminer 1.5.60-git(SP-MOD) for nVidia GPUs by sp-hash@github ***
        Built with the nVidia CUDA SDK 6.5

  Based on pooler cpuminer 2.3.2 and the tpruvot@github fork
   CUDA support by Christian Buchner, Christian H. and DJM34
  Includes optimizations implemented by sp , klaust, tpruvot and tsiv.

[2015-08-13 13:12:03] Intensity set to 19.0, 524288 cuda threads
[2015-08-13 13:12:04] Starting Stratum on stratum+tcp://mining-ace.granitecoin.com:6001/
[2015-08-13 13:12:04] NVML GPU monitoring enabled.
[2015-08-13 13:12:04] Binding thread 1 to cpu 1 (mask 2)
1
[2015-08-13 13:12:04] Binding thread 3 to cpu 3 (mask
3
[2015-08-13 13:12:04] Binding thread 0 to cpu 0 (mask 1)
0
[2015-08-13 13:12:04] Binding thread 2 to cpu 2 (mask 4)
2
[2015-08-13 13:12:04] 5 miner threads started, using 'lyra2v2' algorithm.
[2015-08-13 13:12:04] Binding thread 4 to cpu 0 (mask 10)
[2015-08-13 13:12:04] Stratum difficulty set to 4
[2015-08-13 13:12:05] stratum_gen_work: pool fuck
[2015-08-13 13:12:05] mining-ace.granitecoin.com:6001/ lyra2v2 block 348786
[2015-08-13 13:12:05] stratum_gen_work: pool fuck
[2015-08-13 13:12:05] stratum_gen_work: pool fuck
[2015-08-13 13:12:05] stratum_gen_work: pool fuck
[2015-08-13 13:12:05] stratum_gen_work: pool fuck
[2015-08-13 13:12:05] stratum_gen_work: pool fuck


-------

since when does 'pool fuck' denote a disconnection or issue with the pool? ...

for the private miner - thats one thing ... maybe even funny ...

but for a miner like us that show our machines to prospective investors or buyers - this is a little less tasteful ...

sp - is this part of ccminer now? ... or have i magically done something here in the compile that i didnt know about? ...

can that be removed please? ...

far be it of me to a prude - im no such thing - but to have that within the miner itself? ... ...

anyway - apart from that - x11 and quark have improved and are getting good hashrates ...

as i muster the strength to get things sorted and catch up with everything ive missed in the times that im up - the donation servers will be functional in the coming days ( as i could not work on them all week ) and we will be shooting some donation mining after that ...

tanx sp ...

#crysx

I wasn't talking about small amounts... Verters pays you like 10-20% less then you're supposed to receive (this is by the pool estimate and other sites). You can compare it to other pools as well. This sort of thing wasn't that uncommon a year ago when personal pools were a big deal.


Lots of trading happening at Cryptsy too, even though you can't send coins in or out.