[ANN] sph-sgminer: multi-coin multi-algorithm GPU miner | added MaruCoin

[digitalindustry]

There is target adjustment in sph-sgminer depending on the coin, this is what DM_SELECT(1, 256, 65536) does. Currently for darkcoin and myriadcoin-groestl truediffone is multiplied by 1, for quarkcoin and qubitcoin it's multiplied by 256 and for scrypt it's multiplied by 65536.

so here may be the problem - quarks should have diff1 similar as sha256 coins (so multiplier = 1)

Are you sure about this? Because for example hash of BTC block 10 (difficulty = 1) is:

000000002c05cc2e78923c34df87fd108b22221ac6076c18f3ade378a4d915e9

and hash of QRK block 600 (difficulty = 1.01576304) is:

00000096b99f154706b957c0e36cc4bf3789849e8a0684278dffac607b404641

so QRK definitely has higher target than BTC for this difficulty, otherwise this block would not be accepted. I'm not an expert in this, though.

Also take a look at:

https://github.com/bitcoin/bitcoin/blob/master/src/rpcblockchain.cpp#L36

and

https://github.com/MaxGuevara/quark/blob/master/src/rpcblockchain.cpp#L31

In BTC dDiff is multiplied/divided by 256 until nShift is 29 while in QRK it's multiplied/divided by 256 until nShift is 30 - so IMHO for two difficulties with the same value target value will be shifted by 8 bits.

Note that sph-sgminer displays network difficulty correctly for Quark and QubitCoin. It wouldn't be correct with wrong multiplier.

on our forum i'm about to ask about the possibility of forking this version and adapting it to Quark, so hopefully we can raise a bounty for this , i'd rather you yourself get this phm, as you seems to have done a great job adapting the original SGminer for the C code algos,  and it seems like the work to refine this to a Quark GPU miner would not be that extensive?

BtW re the suprnova mix of valid and invalid i can confirm i was mining with both CPU and GPU on the same worker, so the GPU would have been generating invalid and CPU valid.

curious thing about it was it did generate valid shares, it was showing found blocks for a time.

[1715036517]

1715036517

[1715036517]

1715036517

[1715036517]

1715036517

[1715036517]

1715036517

[feeleep]

There is target adjustment in sph-sgminer depending on the coin, this is what DM_SELECT(1, 256, 65536) does. Currently for darkcoin and myriadcoin-groestl truediffone is multiplied by 1, for quarkcoin and qubitcoin it's multiplied by 256 and for scrypt it's multiplied by 65536.

so here may be the problem - quarks should have diff1 similar as sha256 coins (so multiplier = 1)

Are you sure about this? Because for example hash of BTC block 10 (difficulty = 1) is:

000000002c05cc2e78923c34df87fd108b22221ac6076c18f3ade378a4d915e9

and hash of QRK block 600 (difficulty = 1.01576304) is:

00000096b99f154706b957c0e36cc4bf3789849e8a0684278dffac607b404641

so QRK definitely has higher target than BTC for this difficulty, otherwise this block would not be accepted. I'm not an expert in this, though.

Also take a look at:

https://github.com/bitcoin/bitcoin/blob/master/src/rpcblockchain.cpp#L36

and

https://github.com/MaxGuevara/quark/blob/master/src/rpcblockchain.cpp#L31

In BTC dDiff is multiplied/divided by 256 until nShift is 29 while in QRK it's multiplied/divided by 256 until nShift is 30 - so IMHO for two difficulties with the same value target value will be shifted by 8 bits.

Note that sph-sgminer displays network difficulty correctly for Quark and QubitCoin. It wouldn't be correct with wrong multiplier.

I understand this but there is one problem - in original CPU miner from Neisklar (https://github.com/Neisklar/quarkcoin-cpuminer) I am pretty sure diff1 is set the same as sha256 and it is used by thousands of users...

However I set up a separate stratum port (once again) for GPU users on Securecoin pool. As I dont have any GPU rig now could anyone test it for a while and post results?

mine1.coinmine.pl:6021

feeleep

[digitalindustry]

There is target adjustment in sph-sgminer depending on the coin, this is what DM_SELECT(1, 256, 65536) does. Currently for darkcoin and myriadcoin-groestl truediffone is multiplied by 1, for quarkcoin and qubitcoin it's multiplied by 256 and for scrypt it's multiplied by 65536.

so here may be the problem - quarks should have diff1 similar as sha256 coins (so multiplier = 1)

Are you sure about this? Because for example hash of BTC block 10 (difficulty = 1) is:

000000002c05cc2e78923c34df87fd108b22221ac6076c18f3ade378a4d915e9

and hash of QRK block 600 (difficulty = 1.01576304) is:

00000096b99f154706b957c0e36cc4bf3789849e8a0684278dffac607b404641

so QRK definitely has higher target than BTC for this difficulty, otherwise this block would not be accepted. I'm not an expert in this, though.

Also take a look at:

https://github.com/bitcoin/bitcoin/blob/master/src/rpcblockchain.cpp#L36

and

https://github.com/MaxGuevara/quark/blob/master/src/rpcblockchain.cpp#L31

In BTC dDiff is multiplied/divided by 256 until nShift is 29 while in QRK it's multiplied/divided by 256 until nShift is 30 - so IMHO for two difficulties with the same value target value will be shifted by 8 bits.

Note that sph-sgminer displays network difficulty correctly for Quark and QubitCoin. It wouldn't be correct with wrong multiplier.

I understand this but there is one problem - in original CPU miner from Neisklar (https://github.com/Neisklar/quarkcoin-cpuminer) I am pretty sure diff1 is set the same as sha256 and it is used by thousands of users...

However I set up a separate stratum port (once again) for GPU users on Securecoin pool. As I dont have any GPU rig now could anyone test it for a while and post results?

mine1.coinmine.pl:6021

feeleep

I can do this for you, is securecoin a Quark fork, I've barely even heard of it?

where is the miner ?  or do you want me to compile a source, i have access to a windows option and Linux.

i.e where is the GPU miner i should use to test GPU for securcoin ?

[aakashkumar]

For MYR coin  
 5870 - 2.86 Mh/s
 5970 - 5.5  Mh/s
Is this bad ?  
both or overclocked and tem is below 65  
 

[feeleep]

I can do this for you, is securecoin a Quark fork, I've barely even heard of it?

where is the miner ?  or do you want me to compile a source, i have access to a windows option and Linux.

i.e where is the GPU miner i should use to test GPU for securcoin ?

yes - this is quark clone and just download miner from this thread

[dingoKoin]

can somebody share his/her .sgminer.conf for an r9 290 card, please? i assume i can use a .sgminer.conf in the same manner i use a .cgminer.conf, right?

thank you.

[murraypaul]

I understand this but there is one problem - in original CPU miner from Neisklar (https://github.com/Neisklar/quarkcoin-cpuminer) I am pretty sure diff1 is set the same as sha256 and it is used by thousands of users...

It used to be that the Neisklar miner was hardcoded to never submit shares below a certain difficulty, which masked the problem. This was fixed 7 months ago.
Not sure if that is related to this issue?

https://github.com/Neisklar/quarkcoin-cpuminer/commit/b1af442712ee82fe9764df3812d134a99e11e3f2

Fix issue that miner could only work with diff 0.0039... or higher (not that that matters much now:)

[feeleep]

I understand this but there is one problem - in original CPU miner from Neisklar (https://github.com/Neisklar/quarkcoin-cpuminer) I am pretty sure diff1 is set the same as sha256 and it is used by thousands of users...

However I set up a separate stratum port (once again) for GPU users on Securecoin pool. As I dont have any GPU rig now could anyone test it for a while and post results?

mine1.coinmine.pl:6021

feeleep

I saw some guys started to test it - any results?

[digitalindustry]

I can do this for you, is securecoin a Quark fork, I've barely even heard of it?

where is the miner ?  or do you want me to compile a source, i have access to a windows option and Linux.

i.e where is the GPU miner i should use to test GPU for securcoin ?

yes - this is quark clone and just download miner from this thread

dam feeleep , legend you fixed it, can i ask what you did ?

[digitalindustry]

I understand this but there is one problem - in original CPU miner from Neisklar (https://github.com/Neisklar/quarkcoin-cpuminer) I am pretty sure diff1 is set the same as sha256 and it is used by thousands of users...

However I set up a separate stratum port (once again) for GPU users on Securecoin pool. As I dont have any GPU rig now could anyone test it for a while and post results?

mine1.coinmine.pl:6021

feeleep

I saw some guys started to test it - any results?

results :

i used the original QRK miner compile for win .

so i can confirm i can mine with this and get {share above target}  on Quark or say other clones (other pools)

i then created a worker on your pool and i have accepts the dash shows the correct Mh# and i will check the blocks.

[digitalindustry]

yes everything appears to be functioning i'm mining src (which i don't even know what is ) looks like DGC dev  made this yes ?

hash as represented is about 3x but no probs i assume.

**edit hash stabilized everything functions, we have lift off. feeleep does it again : )

[paulnsk]

I saw some guys started to test it - any results?

Hi! Just registered especially for this theme because of its significance.
So, only one of many kernels are working, the quarkcoin one.
But it's working well!
On my test 7950 it shows 1.500 M avg hashrate with 5.000 M WU.
On the site side I see increasing value of my hashrate aiming to WU. After 25 min its amount reached 3.5 M.
All diffs are acceptable with no rejects and HW errors.
Amount of diffs are 0.5 and 1.0.
Good work, Feeleep!
But this coin is not profitable even with this results for me.

[feeleep]

I saw some guys started to test it - any results?

Hi! Just registered especially for this theme because of its significance.
So, only one of many kernels are working, the quarkcoin one.
But it's working well!
On my test 7950 it shows 1.500 M avg hashrate with 5.000 M WU.
On the site side I see increasing value of my hashrate aiming to WU. After 25 min its amount reached 3.5 M.
All diffs are acceptable with no rejects and HW errors.
Amount of diffs are 0.5 and 1.0.
Good work, Feeleep!
But this coin is not profitable even with this results for me.

does miner hashrate equals what pool was saying?

[ocminer]

I saw some guys started to test it - any results?

Hi! Just registered especially for this theme because of its significance.
So, only one of many kernels are working, the quarkcoin one.
But it's working well!
On my test 7950 it shows 1.500 M avg hashrate with 5.000 M WU.
On the site side I see increasing value of my hashrate aiming to WU. After 25 min its amount reached 3.5 M.
All diffs are acceptable with no rejects and HW errors.
Amount of diffs are 0.5 and 1.0.
Good work, Feeleep!
But this coin is not profitable even with this results for me.

does miner hashrate equals what pool was saying?

Hmm.. So to understand correctly:

The miner which gets rejected on for example suprnova works well on coinmine now and reports the correct hashrate ? no more fake shares ? and it is fixed by a tweak at stratum on coinmine only

[paulnsk]

Hmm.. So to understand correctly:

The miner which gets rejected on for example suprnova works well on coinmine now and reports the correct hashrate ? no more fake shares ? and it is fixed by a tweak at stratum on coinmine only

Yes, unfortunately the one and only pool is correct 

[paulnsk]

does miner hashrate equals what pool was saying?

As I said the avg hashrate for session sgminer is showing is 1.5 MH/s.
Because of using 4 threads the theoretical hashrate must be 6 MH/s but Work Utility that is efficiency of a card is now showing 5.2 MH/s.
The site hashrate showing 3.5 MH/s and I think will not rise anymore. That was results one of my 7950 cards with some overclocking.
The pure qubit and groestl algo much more effective on video, have you any idea to put up pools with some corresponding coins?

[darius08]

I understand this but there is one problem - in original CPU miner from Neisklar (https://github.com/Neisklar/quarkcoin-cpuminer) I am pretty sure diff1 is set the same as sha256 and it is used by thousands of users...

However I set up a separate stratum port (once again) for GPU users on Securecoin pool. As I dont have any GPU rig now could anyone test it for a while and post results?

mine1.coinmine.pl:6021

feeleep

I saw some guys started to test it - any results?

Looks good!

[digitalindustry]

I saw some guys started to test it - any results?

Hi! Just registered especially for this theme because of its significance.
So, only one of many kernels are working, the quarkcoin one.
But it's working well!
On my test 7950 it shows 1.500 M avg hashrate with 5.000 M WU.
On the site side I see increasing value of my hashrate aiming to WU. After 25 min its amount reached 3.5 M.
All diffs are acceptable with no rejects and HW errors.
Amount of diffs are 0.5 and 1.0.
Good work, Feeleep!
But this coin is not profitable even with this results for me.

does miner hashrate equals what pool was saying?

Hmm.. So to understand correctly:

The miner which gets rejected on for example suprnova works well on coinmine now and reports the correct hashrate ? no more fake shares ? and it is fixed by a tweak at stratum on coinmine only

i can confirm that exactly yes - i did not download a new compile,  i used the same miner, i used it on a windows machine (only one i have) and its the first download from the topic:

works on Feeleeps SRC pool which is the Quark algo as i left it exactly and only changed the pool address .

did not work on any other Quark (or clone) pool, and gave back {share target error}  ~ < the error

[feeleep]

I saw some guys started to test it - any results?

Hi! Just registered especially for this theme because of its significance.
So, only one of many kernels are working, the quarkcoin one.
But it's working well!
On my test 7950 it shows 1.500 M avg hashrate with 5.000 M WU.
On the site side I see increasing value of my hashrate aiming to WU. After 25 min its amount reached 3.5 M.
All diffs are acceptable with no rejects and HW errors.
Amount of diffs are 0.5 and 1.0.
Good work, Feeleep!
But this coin is not profitable even with this results for me.

does miner hashrate equals what pool was saying?

Hmm.. So to understand correctly:

The miner which gets rejected on for example suprnova works well on coinmine now and reports the correct hashrate ? no more fake shares ? and it is fixed by a tweak at stratum on coinmine only

i can confirm that exactly yes - i did not download a new compile,  i used the same miner, i used it on a windows machine (only one i have) and its the first download from the topic:

works on Feeleeps SRC pool which is the Quark algo as i left it exactly and only changed the pool address .

did not work on any other Quark (or clone) pool, and gave back {share target error}  ~ < the error

Thanks guys for testing. There is still something not correct between cpuminer-gpuminer-stratum so my solution if following:
- GPU miners: diff1 = 0x000000ffff000000000000000000000000000000000000000000000000000000
- CPU miners: diff1 = 0x00000000ffff0000000000000000000000000000000000000000000000000000

but GPU miner share is 256 time less valuable than CPU share - dont ask me why but it works. If there is anyone who have an idea let me know please.

I will set up GPU stratum port for all my quark-based coins (Quarks, Securecoin, Particle, Frozen) later today

feeleep

[phm]

I guess I'll make share target multiplier configurable, then users will be able to select a different one if there are problems with a pool.