[ANN][X11/X13] X11 (Darkcoin)/X13 (Marucoin) miner (based on sph-sgminer)

[ekoice]

my settings for 6x 280x toxics, i hope to help someone.

"intensity" : "15",
"worksize" : "256",
"kernel" : "x11mod",
"lookup-gap" : "2",
"thread-concurrency" : "8192",
"shaders" : "2048",
"gpu-threads" : "2",
"gpu-engine" : "1180",
"gpu-fan" : "50-100",
"gpu-memclock" : "1500",
"gpu-powertune" : "20",
"temp-cutoff" : "85",
"temp-overheat" : "80",
"temp-target" : "70",
"api-listen" : true,
"api-mcast-port" : "4028",
"api-port" : "4028",
"auto-fan" : true,
"expiry" : "20",
"failover-only" : true,
"failover-switch-delay" : "10",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"log" : "5",
"no-pool-disable" : true,
"queue" : "1",
"scan-time" : "10",
"tcp-keepalive" : "30",
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin",
"api-allow" : "W:127.0.0.1,W:192.168.1.0/24"
}

doing 3.3 mh/s each on pimp 1.4.2-amd-13.12

maybe should i try with shaders 0 ?

[JPS2K5]

I have the latest. Upped engine to 950 now, still looks stable. 2.4Mh/s+

gpu-engine up to 1000, still good. 2.5 Mh/s+

try mem 1500

Up to 2.57 Mh/s steady 

2.67 Mh/s now @ 1050/1500... oh boy 

[platinum4]

Can confirm that the 2% girino version is markedly different for Hawaii chipsets.

If somebody could just make the girino version with failover added back into its source code, we would all be happy.

[MIN3R]

I'm rocking various (Sapphire, XFX, Gigabyte) 270 & 270X and am getting 2.05+MH/s!  Settings:

Code:

"gpu-threads" : "4",
"gpu-fan" : "75-100",
"gpu-engine" : "1200",
"gpu-memclock" : "1200",
"gpu-powertune" : "20",
"kernel" : "x11mod",
"temp-target" : "75",
"temp-overheat" : "88",
"worksize" : "64",
"shaders" : "1280",
"thread-concurrency" : "6401",
"rawintensity" : "32768"

With previous software I was using 2 threads with other settings than above, but with the optimized x11mod  and after a LOT of testing the past couple days that's the "best" I have so far. I can push it faster but that seems to be pretty stable across all 5 of my machines (27 GPUs, > 50MH/s). If anyone has any better settings for 270's please pass along

[lasybear]

Does anyone wants to test updated version (intensity > 15)?
Changes may bumps (or not! - test only due to i haven't R9 290x) hashrates only on R9 290x. Tested R9 280x and R9 270x - no more hash rate while increasing intensity in compare to previous version.
Requirements:
0. I'll provide only zipped sources dropbox link.
1. R9 290x
2. Sufficient skills to compile in your environment.

[NRD]

Correct me if I'm wrong, but the difference between this miner and the 2% fee one is that this is built on sph-sgminer and the 2%fee version is built on sgminer 4.1 is it not?

[Remember remember the 5th of November]

Does anyone want to test updated version (intensity > 15)?
Changes may bumps (or not! - test only due to i haven't R9 290x) hashrates only on R9 290x. Tested R9 280x and R9 270x - no more hash rate while increasing intensity in compare to previous version.
Requirements:
0. I'll provide only zipped sources dropbox link.
1. R9 290x
2. Sufficient skills to compile in your environment.

Did you figure it out?

[NRD]

Does anyone want to test updated version (intensity > 15)?
Changes may bumps (or not! - test only due to i haven't R9 290x) hashrates only on R9 290x. Tested R9 280x and R9 270x - no more hash rate while increasing intensity in compare to previous version.
Requirements:
0. I'll provide only zipped sources dropbox link.
1. R9 290x
2. Sufficient skills to compile in your environment.

Thanks for your work Lasybear(Edit: and thanks RRT5ON as well!), I can't help with compiling to test it atm but I'm sure there will be a few along soon enough. What have you changed in this version? Is it just a change to the intensity limit?

[DriverDan]

Does anyone wants to test updated version (intensity > 15)?
Changes may bumps (or not! - test only due to i haven't R9 290x) hashrates only on R9 290x. Tested R9 280x and R9 270x - no more hash rate while increasing intensity in compare to previous version.
Requirements:
0. I'll provide only zipped sources dropbox link.
1. R9 290x
2. Sufficient skills to compile in your environment.

Why don't you create a branch on github and let people test it? That's half the point of open source software. Anyone can help and anyone can contribute.

[lasybear]

Does anyone wants to test updated version (intensity > 15)?
Changes may bumps (or not! - test only due to i haven't R9 290x) hashrates only on R9 290x. Tested R9 280x and R9 270x - no more hash rate while increasing intensity in compare to previous version.
Requirements:
0. I'll provide only zipped sources dropbox link.
1. R9 290x
2. Sufficient skills to compile in your environment.

Why don't you create a branch on github and let people test it? That's half the point of open source software. Anyone can help and anyone can contribute.

Please wait a bit. I've only one positive report:

Code:

GPU 0: 63.0C 1107RPM | 3.438M/3.584Mh/s | R: 0.0% HW:0 WU:0.054/m I:18

This is for R9 280x, but been tested intensity till 21 without crashes.

[platinum4]

Damn this is frustrating; just want to have girino's version with failover implemented by default and unhidden and the quota gutted out.

Shouldn't be too hard to isolate where that code in the source is, right?

[Gamer67]

Does anyone wants to test updated version (intensity > 15)?
Changes may bumps (or not! - test only due to i haven't R9 290x) hashrates only on R9 290x. Tested R9 280x and R9 270x - no more hash rate while increasing intensity in compare to previous version.
Requirements:
0. I'll provide only zipped sources dropbox link.
1. R9 290x
2. Sufficient skills to compile in your environment.

Why don't you create a branch on github and let people test it? That's half the point of open source software. Anyone can help and anyone can contribute.

Please wait a bit. I've only one positive report:

Code:

GPU 0: 63.0C 1107RPM | 3.438M/3.584Mh/s | R: 0.0% HW:0 WU:0.054/m I:18

Looks like some progress. I hope it gets solved soon. I have a few 290x's that could use it.

[lasybear]

Damn this is frustrating; just want to have girino's version with failover implemented by default and unhidden and the quota gutted out.

Shouldn't be too hard to isolate where that code in the source is, right?

If you have sources from girino, please provide it to community.
Because it is not his sources.

[carafleur]

Hi all,
These are the interesting settings I use for my 6x290 rig :

Code:

"intensity" : "15",
"worksize" : "128",
"kernel" : "x11mod",
"lookup-gap" : "0",
"thread-concurrency" : "20481",
"shaders" : "2816,2560,2560,2560,2560,2560",
"gpu-threads" : "4,4,4,4,8,8",
"gpu-engine" : "938-1029,945-1019,945-1019,945-1019,1000-1029,1000-1029",
"gpu-fan" : "20-60,20-60,40-70,40-70,40-70,45-80",
"gpu-memclock" : "1500,1400,1400,1400,1400,1400",
"auto-fan" : true,
"temp-target" : "80",

and the work in progress after 3h33 of mining :

Code:

ALL (60s):20.78M (avg):20.56Mh/s | A:60  R:0  HW:0  WU:0.289/m
ASUS 290X  : GPU  0:  78.0C  20%    | 3.417M/3.377Mh/s | R:  1.6% HW:0 WU:0.047/m I:15
ASUS 290   : GPU  1:  78.0C  27%    | 3.338M/3.316Mh/s | R:  0.4% HW:0 WU:0.047/m I:15
ASUS 290   : GPU  2:  79.0C  51%    | 3.415M/3.392Mh/s | R:  0.0% HW:0 WU:0.047/m I:15
ASUS 290   : GPU  3:  77.0C  52%    | 3.422M/3.392Mh/s | R:  0.0% HW:0 WU:0.048/m I:15
SPH TRI-X  : GPU  4:  77.0C  49%    | 3.511M/3.504Mh/s | R:  0.3% HW:0 WU:0.050/m I:15
SPH TRI-X  : GPU  5:  77.0C  59%    | 3.676M/3.641Mh/s | R:  0.3% HW:0 WU:0.051/m I:15

2 days ago I did not have more than 2.6MH per card 

OS : xubuntu 14.4 -64bit

Kill-A-Watt shows 1150W +/- 5%

I tried to set 8 gpu-threads on each card, but sgminer will not take it : it seems not to support more than 32 "virtual gpus" (core dumped at launch)
It seems to me that gpu/memclock does not make a big difference. I may set it back to default settings to lower the temps

Also, I found out that the last card detected always shows a higher hashrate than the others. No matter which one it is ! X11 mining seems like sorcery to me, I wihs I still had my voodoo cards =)

Thanks so much for your work beloved devs 
Pls share a DRK address so I can send you some (lasybear and the V4Vandetta fan )

[Bawb3]

Does anyone wants to test updated version (intensity > 15)?
Changes may bumps (or not! - test only due to i haven't R9 290x) hashrates only on R9 290x. Tested R9 280x and R9 270x - no more hash rate while increasing intensity in compare to previous version.
Requirements:
0. I'll provide only zipped sources dropbox link.
1. R9 290x
2. Sufficient skills to compile in your environment.

I'd be happy to test the windows binary if someone could compile it. I don't have my Ubuntu machine set up anymore. That drive got repurposed for a mining rig :-P.

[platinum4]

I have 8x R9 290X am willing to test a compiled Windows build with newer sources from lasybear, PM me.

[Bawb3]

I'm rocking various (Sapphire, XFX, Gigabyte) 270 & 270X and am getting 2.05+MH/s!  Settings:

Code:

"gpu-threads" : "4",
"gpu-fan" : "75-100",
"gpu-engine" : "1200",
"gpu-memclock" : "1200",
"gpu-powertune" : "20",
"kernel" : "x11mod",
"temp-target" : "75",
"temp-overheat" : "88",
"worksize" : "64",
"shaders" : "1280",
"thread-concurrency" : "6401",
"rawintensity" : "32768"

With previous software I was using 2 threads with other settings than above, but with the optimized x11mod  and after a LOT of testing the past couple days that's the "best" I have so far. I can push it faster but that seems to be pretty stable across all 5 of my machines (27 GPUs, > 50MH/s). If anyone has any better settings for 270's please pass along

Looks good. I'll try it out. What driver version are you running?

[Remember remember the 5th of November]

Alright, please test this version for the intensity fix https://www.dropbox.com/s/skv74wv378h5qnl/sgminerx11.rar. Thanks to leasybear for pointing out where the problem was, then I went to the 2% miner and looked up some values he was using and incorporated those in the miner, again, thanks to lasybear.

[lasybear]

Thanks so much for your work beloved devs 
Pls share a DRK address so I can send you some (lasybear and the V4Vandetta fan )

Thanks )
DRK XxMb95zfrxw8Csh9hSX3T98McDGZqk3Qrm

[platinum4]

Damn thanks both of y'all for helping out over the 3-day weekend.  Okay, I have the new build fired up

I have -w 128 -I 19 like on 2% no AMD driver crashes on R9 290X rigs.

Hashrate increase from 2.844 (sph) to 3.986 (2%) to ~4.0 (x11mod)

The difficulty readouts are still showing as Accepted -0.000/-0.000 is there a divide-by error being overlooked?

Confirmed this yields good results for Hawaii chipsets with TC 14080 I 19 or TC 24082 I 18