Topic: [ANN][X11/X13] X11 (Darkcoin)/X13 (Marucoin) miner (based on sph-sgminer) - page 61. (Read 351555 times)

My rig at idle using 255W  (AMD6300 + 280x+7870+7850+5670 + 6 funs)

If you're using 230W at idle something is wrong. All of my rigs use 40-50W at idle with no vid cards, 100-120W with 3-5 cards.

I know.. but I won't get more for my 7950 than what I'd have to pay for a 750Ti. And 230 watts is like a tad bit more than running idle 

750ti are still better and they are receiving an additional boost(2.8MH/s) in 1 week or so

My RIG 2*280x Sapphire + 2*270x ASUS
[SPOILER]
{
"pools" : [
   {
      "name" : "TMB-1",
      "url" : "stratum+tcp://am01.eu.trademybit.com:4440",
      "user" : "GrandZ.13",
      "pass" : "159159",
      "pool-priority" : "0"
   },
   {
      "name" : "TMB-01",
      "url" : "stratum+tcp://am02.eu.trademybit.com:4440",
      "user" : "GrandZ.13",
      "pass" : "159159",
      "pool-priority" : "1"
   }
],
"api-allow" : "W:127.0.0.1,W:192.168.0.0/24",
"api-listen" : true,
"api-mcast-port" : "4028",
"api-port" : "4028",
"expiry" : "28",
"vector" : "1",
"failover-only" : true,
"kernel-path" : "/usr/local/bin",
"log" : "5",
"queue" : "1",
"scan-time" : "7",
"shares" : "0",
"gpu-threads" : "2,2,2,2",
"gpu-dyninterval" : "7",
"gpu-engine" : "1170,1200,1150,1150",
"gpu-platform" : "0",
"gpu-memclock" : "1500",
"gpu-memdiff" : "0,0,0,0",
"gpu-powertune" : "20,20,20,20",
"intensity" : "22,22,18,18",
"kernel" : "x11mod,x11mod,x11mod,x11mod",
"temp-hysteresis" : "3",
"worksize" : "128,128,256,256",
"lookup-gap" : "2,2,2,2",
"shaders" : "2048,2048,1280,1280",
"thread-concurrency" : "8192,8192,8192,8192",
"failover-switch-delay" : "10",
"no-pool-disable" : true,
"tcp-keepalive" : "30"
}
[/SPOILER]
http://s29.postimg.org/4n7prgeyb/X11_mod.jpg
SANKS for this mod!!!

I use -g 2, thread-concurrency 8192, -i 17 and can get 2.45MH, with core/memory 950/1250MHz. You should get more with your 1100MHz core.

My 7950 is modded with faster memory timing. There is a thread about that in litecointalk

I am currently on ~2.8 Mh/s @ 1050/1400, I=18 (19 or 20 gives me nothing extra), w=256

230watts total system power (i5), 59 degrees.

I was about to sell my card for a 750ti but mehhh....

I have updated my post. And here's the thing. With the old miner, both my cards used to use up about 260 Megabytes of video memory on average, but now the video memory usage on my HD 7950 is 2.7 GB. The more threads you will add, the more video memory it will require, and if the number of threads are too high (I used to run 8 threads on my HD 7950, now I am running 5) the miner will not work at all. If you are facing inconsistant hash rates, simply start lowering the thread count till you reach to a point where the hash rate isn't jumpy anymore. And yes, the pool dashboard isn't real time, so there will be a variance.

If I didn't lower the number of threads, will I still get the hashrate as reported in long run?

The miner seems to be reporting higher hash rates, however, from the pool's side, I see no improvements.

Edit: Oddly enough, now it's reporting higher than what I used to get with the older miner over time from the pool's side,

What I noticed is that the video memory usage of the cards have skyrocketed, the miner is now using eight to ten times more video memory than it used to.

Those of you who are facing inconsistent hash rates, try lowering the number of threads. The inconsistent hash rate occurs when the card starts hitting it's video memory limit. Be sure to leave 200 Mb of free memory at least (you can view the vram usage using Gpu-Z)

And for those of you who are still using Intensity, I think it's about time you people learn to use the Experimental intensity, and use more than two threads. It provides much better results while putting less stress on your hardware.

My settings for my Sapphire HD 7950 3GB Dual-X OC:

sgminer.exe -k x11mod -o stratum+tcp://your.pool.address:port -u username -p password --thread-concurrency 16384 --no-submit-stale -g 5 -X 6 --gpu-powertune 15 --lookup-gap 0 --gpu-engine 1100--queue 0 --scan-time 1 --expiry 30

This gets me 2.442 Mh/s (previously I used to get only 2 Mh/s with the old miner, and my clocks are pretty conservative. If you can push the core and memory more, you will yield better results)

And the settings I have for my Sapphire HD 7850 Dual-X OC 2GB:

sgminer.exe -k x11mod -o stratum+tcp://your.pool.address:port -u username -p password --thread-concurrency 4096 --lookup-gap 0 --no-submit-stale -g 2 -X 6 --gpu-engine 1100 --queue 0 --scan-time 7 --expiry 28

This gets me about 1.4 Mh/s, previously I used to get about 1.120 Mh/s at best.

I hope this helps.

So, got 2 rigs each with 3x r9 290 and the 3rd card in each one always reports lower hashrate with this miner (tried using all the versions posted here)
Here's how it shows:

 GPU 0:  71.0C  72%    | 3.662M/3.638Mh/s | R:  3.0% HW:0 WU:0.054/m I:19
 GPU 1:  65.0C  66%    | 3.613M/3.628Mh/s | R:  1.8% HW:0 WU:0.052/m I:19
 GPU 2:  66.0C  67%    | 3.222M/3.284Mh/s | R:  2.9% HW:0 WU:0.044/m I:19

It's identical in both rigs (first 2 GPUs are 3.6mh/s while 3rd one is 3.2mh/s). Tried swapping cards location, playing with configs, numerous restarts, but nothing helped.

When mining Scrypt, n-Factor or x-11 using girino's they're all stable (getting 3.6mh/s with girino's as well)

Any ideas why that is happening?

I used the B2 version... didn't saw never version ?

4x Asus r9 280x dc2 for a total hashrate of 12.66Mh/s

Power consumtion of 560W at wall (kill-a-wat measured). All GPUs undervolted using MSI Afterburner (two at 1,056V and two at 1,075V).

All cards at 1.070 Core clock and 1.500 memory clock

Temps between 44º and 47ºC (room temp 21ºC) with fixed 50% fan speed

My settings are:

"api-allow" : "W:127.0.0.1",
"api-listen" : true,
"expiry" : "1",
"failover-only" : true,
"queue" : "1",
"scan-time" : "1",
"intensity" : "18",
"vector" : "1",
"worksize" : "256",
"kernel" : "x11mod",
"thread-concurrency" : "8192",
"lookup-gap" : "2",
"failover-switch-delay" : "10",
"gpu-threads" : "2"

Which version of miner you are using, there is actually 3-4 version of this sph-sgminer. Use the latest one which fixed some bugs I think..

I too had too much problems with this version of miner, gpu's freezing, restarting, locking and so on. I had to use higher voltages than for scrypt to make it work stable but power draw is almost the same as original darkcoin kernel.

Wow ... 7950 Dual-X @1050/1250

before = 1.78 Mh/s
after = 2.55 Mh/s

I'll check the power draw when I'm home

It's called variance.

Anybody have low hash rate reported from pool website? My total rig is 12.56 mh/s but my pool only reported 10.25 mh/s

I see Diff and WU on his screenshots ) And high intensity for r9 280x (i use 16).
And HW errors (he said) is a effect of engine overboost (1100MHz for used undervoltaged card is too high with multiple kernels), high intensity and high tc...
Hint for everybody: results are better and stability much higher when a hashrate is divided equally between all threads on GPU, like this:

what does really mean ?