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

[haggis]

Hey guys,

I just started mining Darkcoin and everything is fine. Significantly lower temps + power consumption

As my R290s are watercooled, the system is totaly noiseless. But by mining X11 I hear my GPUs "scratching" permanently. Normally, at scrypt, I sometimes hear this only for a short time at startup but never during the mining process.

Can someone confirm this? Is this something I should care about regarding the health of my cards? Would be nice to get rid of this (annoying) sound as I run the rig in my office

[mfqrs3]

Hey guys,

I just started mining Darkcoin and everything is fine. Significantly lower temps + power consumption

As my R290s are watercooled, the system is totaly noiseless. But by mining X11 I hear my GPUs "scratching" permanently. Normally, at scrypt, I sometimes hear this only for a short time at startup but never during the mining process.

Can someone confirm this? Is this something I should care about regarding the health of my cards? Would be nice to get rid of this (annoying) sound as I run the rig in my office

Yes, you are right. It sounds like inductors noise, whining, and it's only with this algo. And it doesn't sound good for our HW

[apple_talk]

running latest SPH-SGMINER .98 version for dark coin & getting no monitor.

rebuild so many times with updated ADL SDL .h files.

290 & 290X GPUs

when using sgminer 4.1.0 (in scrypt), no issues with hardware monitoring, temp, fan.

did you run:

autoreconf -fvi
./configure CFLAGS="-O2 whatever you depend on"
make

??

I also Ran test cmd
aticonfig --adapter=1 --od-getclocks && aticonfig --adapter=1 --od-gettemperature
This confirms hardware is working.


Used this

./autogen.sh
./configure --enable-opencl --disable-adl
make
make install


Also tried...
./autogen.sh
./configure CFLAGS="-O3" --enable-opencl --enable-adl
make
make install

Not CFLAGS="-O3 ...."

CFLAGS="-O2 ..." !!!!

Pls try exactly this, without anything more or less.

This all is only working if you copied the ADL-Header files from your ADL-Download into sgminer/ADL_SDK (Build-Directory).

then in your build-directory run:

autoreconf -fvi
./configure CFLAGS="-O2 -Wall -march=native"
make


that's it.

Tried
autoreconf -fvi
./configure CFLAGS="-O2 -Wall -march=native"
make


sgminer 4.1.0-98-ga683 - Started: [2014-04-03 15:05:52]
--------------------------------------------------------------------------------
(5s):1.447M (avg):2.011Mh/s | A:0  R:0  HW:0  WU:0.000/m
ST: 10  SS: 0  NB: 1  LW: 0  GF: 0  RF: 0
Connected to Pool 0 (longpoll) diff 424 as user vcoinrpc
Block: 4a9731f1...  Diff:424  Started: [15:05:51]  Best share: 0
--------------------------------------------------------------------------------
[P]ool management [G]PU management [ S ]ettings [D]isplay options [Q]uit
 GPU 0:                | 799.4K/997.3Kh/s | R:  0.0% HW:0 WU:0.000/m I:18
 GPU 1:                | 864.0K/1.098Mh/s | R:  0.0% HW:0 WU:0.000/m I:18
--------------------------------------------------------------------------------
[15:05:48] Started sgminer 4.1.0-98-ga683
[15:05:48] Loaded configuration file /opt/ifmi/cgminer.darkcoin.conf
[15:05:49] Kernel darkcoin is experimental.
[15:05:49] Kernel darkcoin is experimental.
[15:05:49] Kernel darkcoin is experimental.
[15:05:50] Kernel darkcoin is experimental.
[15:05:50] Kernel darkcoin is experimental.
[15:05:50] Kernel darkcoin is experimental.
[15:05:50] Kernel darkcoin is experimental.
[15:05:51] Kernel darkcoin is experimental.
[15:05:51] Probing for an alive pool
[15:05:51] Network diff set to 424
[15:05:51] No suitable long-poll found for http://192.168.1.99:12345
[15:05:57] API bind to port 4028 failed - trying again in 30sec

When running sgminer...
sgminer 4.1.153-63-g10ec - Started: [2014-04-03 19:02:50]
--------------------------------------------------------------------------------
(5s):1.677M (avg):1.598Mh/s | A:0  R:0  HW:0  WU:1674.3/m
ST: 1  SS: 0  NB: 2  LW: 42  GF: 0  RF: 0
Connected to sf.clevermining.com (stratum) diff 512 as user XXXXXXXXXXXXXXXXXXXXXXXXXXXX
Block: e47ee237...  Diff:48.6M  Started: [19:02:55]  Best share: 220
--------------------------------------------------------------------------------
[P]ool management [G]PU management [ S ]ettings [D]isplay options [Q]uit
GPU 0:  73.0C 2050RPM | 666.9K/909.4Kh/s | R:  0.0% HW:0 WU: 979.0/m I:18
GPU 1:  61.0C 2410RPM | 636.0K/869.4Kh/s | R:  0.0% HW:0 WU: 848.6/m I:18
--------------------------------------------------------------------------------
[19:02:43] Started sgminer 4.1.153-63-g10ec
[19:02:43] Loaded configuration file /opt/ifmi/cgminer.conf
[19:02:44] Initialising kernel ckolivas.cl with bitalign, unpatched BFI
[19:02:44] Initialising kernel ckolivas.cl with bitalign, unpatched BFI
[19:02:44] Probing for an alive pool
[19:02:49] sf.clevermining.com difficulty changed to 512
[19:02:50] Network diff set to 46.9M
[19:02:55] API running in IP access mode on port 4028 (11)
[19:02:55] Network diff set to 48.6M
[19:02:55] Stratum from sf.clevermining.com detected new block

[Pleskac]

well thaks for that that solved the location error but still i have no accepted shares no rejected and no hardware errors... my hashrate is working and LW is increasing... i really dont see problem anywhere it doesnt give me any errors now..

Code:

setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 100
sgminer -k groestlcoin -o stratum+tcp://grs.suprnova.cc:5544 -u nick.1 -p 1 -I 19 -w 256 --scan-time 1 --no-submit-stale -g 1 --gpu-engine 1150 --gpu-memclock 1500,1500,1500,1500 --thread-concurrency 16320 --lookup-gap 2 --gpu-powertune 10 --gpu-fan 85

[badman74]

well thaks for that that solved the location error but still i have no accepted shares no rejected and no hardware errors... my hashrate is working and LW is increasing... i really dont see problem anywhere it doesnt give me any errors now..

Code:

setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 100
sgminer -k groestlcoin -o stratum+tcp://grs.suprnova.cc:5544 -u nick.1 -p 1 -I 19 -w 256 --scan-time 1 --no-submit-stale -g 1 --gpu-engine 1150 --gpu-memclock 1500,1500,1500,1500 --thread-concurrency 16320 --lookup-gap 2 --gpu-powertune 10 --gpu-fan 85

if you want to use that pool i suggest using one of the pre diff patch versions as it looks like supernova pool is set up for the extra low diff shares which means they are actually calling fo super hard shares on the new miner
at the current diff it should take me at least 10 min to get a share with 17.5mh/s
net diff 85
share diff from supernova pool 3
.3 WU/m
edit: hmm my diff calc is different from yours but still should add up about the same

edit2: well it seems the IS something wrong with the diff calc....
i will try to figure it out when i get the chance

edit3: not the diff calc but there is something wrong... i know it was working before with the same setup, so not sure what could have changed

[csa1234]

This is my results for Darkcoin on 3 x R9 280x Sapphire TOXIC:


hosting imagenes

[ocminer]

well thaks for that that solved the location error but still i have no accepted shares no rejected and no hardware errors... my hashrate is working and LW is increasing... i really dont see problem anywhere it doesnt give me any errors now..

Code:

setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 100
sgminer -k groestlcoin -o stratum+tcp://grs.suprnova.cc:5544 -u nick.1 -p 1 -I 19 -w 256 --scan-time 1 --no-submit-stale -g 1 --gpu-engine 1150 --gpu-memclock 1500,1500,1500,1500 --thread-concurrency 16320 --lookup-gap 2 --gpu-powertune 10 --gpu-fan 85

if you want to use that pool i suggest using one of the pre diff patch versions as it looks like supernova pool is set up for the extra low diff shares which means they are actually calling fo super hard shares on the new miner
at the current diff it should take me at least 10 min to get a share with 17.5mh/s
net diff 85
share diff from supernova pool 3
.3 WU/m
edit: hmm my diff calc is different from yours but still should add up about the same

edit2: well it seems the IS something wrong with the diff calc....
i will try to figure it out when i get the chance

edit3: not the diff calc but there is something wrong... i know it was working before with the same setup, so not sure what could have changed

Hey !

Suprnova pool is patch against various versions of fake shares, therefore the miners has to accept the INITIAL diff AND also the ascending or descending variable diff's.

If Surpnova sees your client is "too fast" for a superlow share, you'll get a higher diff quite soon, so it should make no big problem with miners IF they miner is correct and now manipulated in any way.

[restless]

The fact is build from 26-03 works flawlessly... so something was changed after that date

[ocminer]

The fact is build from 26-03 works flawlessly... so something was changed after that date

Then I'd stick to that version or if I'm on a pool where the other version works, look exactly in the stats if the hashrates are plausible.

[fenghush]

Is it possible to squeeze more juice from the pure groestl, feels bit slow for a non-chained algo does it not?

[mfqrs3]

Is it possible to squeeze more juice from the pure groestl, feels bit slow for a non-chained algo does it not?

Yes I agree. As we saw with heavycoin X11 algo, every week 30-40% performance improvement. Devs should release the faster miners to public
Groestl is heavly unoptimised, it runs on 5870 as fast as 280x...

[phm]

Is it possible to squeeze more juice from the pure groestl, feels bit slow for a non-chained algo does it not?

Yes I agree. As we saw with heavycoin X11 algo, every week 30-40% performance improvement. Devs should release the faster miners to public
Groestl is heavly unoptimised, it runs on 5870 as fast as 280x...

How about optimizing it then? I'm waiting for git pull requests.

[phm]

i mean -k groestlcoin, no accepted share for me with this argument

Just tested 2 latest builds from badman, exe inside are from 30-03 and 31-03, builds:
sgminer 4.1.0-98-ga683
sgminer 4.1.0-111-g8bb946
No accepted shares for groestl on https://grs.suprnova.cc

Build from 26-03 works flawlessly - sgminer 4.1.0-105-g9dc6e

Perhaps this build is from the patched sources? I remember that sph-sgminer did not work with some GRS stratum pool and someone created a patch solving that problem. I had no time to review and integrate these changes yet, so they are not present in my source tree.

[tseug]

hey phm, any chance of including twecoin? as I mentioned in a earlier post it's sphlib based

[phm]

hey phm, any chance of including twecoin? as I mentioned in a earlier post it's sphlib based

I'm working on this.

[tseug]

hey phm, any chance of including twecoin? as I mentioned in a earlier post it's sphlib based

I'm working on this.

Excellent, good to hear. Thank you for the great work on this miner.

[restless]

Is it possible to squeeze more juice from the pure groestl, feels bit slow for a non-chained algo does it not?

I think qubit algo has more room for improvement.
ATM a 6970 is only 3x faster than quad-core i5. Comparing vs igotic's client for myriad-qubit

[janos666]

ok it works thanks but! i have no accepted shares so far ? hashrate is showing up to me but not on pool..

edit: says GPU_USE_SYNC_OBJECTS is not specified?

edit2: is it possible to mine 2 algo with one card?

ah lol you're doing a beginners mistake, you have to set the environment variables before you start the miners, otherwise you only get HW Errors.

do:

export GPU_MAX_ALLOC_PERCENT=100
export GPU_USE_SYNC_OBJECTS=1

BEFORE you start ./sgminer ....

You don't actually need "export GPU_USE_SYNC_OBJECTS=1" anymore. It was a workaround for a driver bug which got fixed a long time ago (you might mined Bitcoin with your GPU back then).
So, it bugs me that sgminer (and it's forks) instructs me to set this env.var.

[ocminer]

hey phm, any chance of including twecoin? as I mentioned in a earlier post it's sphlib based

I'm working on this.

I already have a test version for suprnova which is just missing the .cl as i have no time left atm if you don't want to start from the beginning i can send you this

[azhago]

i mean -k groestlcoin, no accepted share for me with this argument

Just tested 2 latest builds from badman, exe inside are from 30-03 and 31-03, builds:
sgminer 4.1.0-98-ga683
sgminer 4.1.0-111-g8bb946
No accepted shares for groestl on https://grs.suprnova.cc

Build from 26-03 works flawlessly - sgminer 4.1.0-105-g9dc6e


btw Anyone aware what happened to https://cpu-pool.net/grs/index.php ?
6hours ago its stratum stopped working. As strange as it may sound I was getting higher hashrates on it than on grs.suprnova (at least client indicated higher hashrates)

You need to get the version from dwarfpool, which work with pools.

Ther is a version for solo, and a modified one for pool (with any pool, not only dwarfpool) : http://dwarfpool.com/grs