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

[Amph]

i don't think you need to specify the same settings everytime, if you use only one type of gpu

I have 2x290nonx & 270x

yeah in your case you must set different settings, but i was talking about Amgclk65 conf

[1714870455]

1714870455

[1714870455]

1714870455

[1714870455]

1714870455

[svdinu]

Hi guys, thanks for your work. It took a while to read all the posts, but I learnt a lot. I have just broke the 4MH - X11 barrier for R9 290 (ASUS R9 290-DC2OC).

Here are the settings for those interested:

[snip]

I have 2 cards, both running at 1050/1500MHz set with Afterburner (power limit +50). I had to take out the engine and memory settings from the config file as the memory frequency was jumping around between 150 and max setting. Very stressful on the system. Now everything is smooth. I kept only the temp control as I like the strategy.

GPU threads have been increased as to flood the memory as much as possible with mining work (5 seems to be the max limit for my system). That was a tip on the forum as well, thanks the author. Cards memory is now loaded on >80% and my RAM as well.

The miner is Elun's b2 version from the OP, drivers 14.3beta, Win 8.1.

Out of my 2 cards one is water cooled and I guess it can go higher. I'll upgrade the wc sytem in a couple of weeks (have to raise more coins and then go for 5MH/card. I hope I can get there.

When I started reading this post, 2 weeks ago, I had 2.5MH/card, then raised it to 3.6 by following you advises and now 4-4.1. First jump had no impact on energy consumption. Now, I estimate a less than 5% increase, which, for me, is acceptable.

Thanks again guys for your work.

My card memory speed keeps jumping around as well, but I do not have any gpu settings in my .conf file. I'll have to give those drivers a shot I guess.
[Edit:gonna try the latest build]

Can you please check the voltage spikes as well, before switching the driver? Mine varies between 1.125 - 1.180. I have no clue if it's normal or not. Thanks.

[Amgclk65]

i don't think you need to specify the same settings everytime, if you use only one type of gpu

I have 2x290nonx & 270x

yeah in your case you must set different settings, but i was talking about Amgclk65 conf

Running scrypt doesn't cause a issue. Ill give it a shot though.

[restless]

I have 2 cards, both running at 1050/1500MHz set with Afterburner (power limit +50). I had to take out the engine and memory settings from the config file as the memory frequency was jumping around between 150 and max setting. Very stressful on the system. Now everything is smooth. I kept only the temp control as I like the strategy.

My 290 also is jumping memory down to 150 and back, do you have any explanation what might be the reason for this behavior?!

[goodluck0319]

I have 2 cards, both running at 1050/1500MHz set with Afterburner (power limit +50). I had to take out the engine and memory settings from the config file as the memory frequency was jumping around between 150 and max setting. Very stressful on the system. Now everything is smooth. I kept only the temp control as I like the strategy.

My 290 also is jumping memory down to 150 and back, do you have any explanation what might be the reason for this behavior?!

mine also like that. what does it really mean?

[Starscream]

I have 2 cards, both running at 1050/1500MHz set with Afterburner (power limit +50). I had to take out the engine and memory settings from the config file as the memory frequency was jumping around between 150 and max setting. Very stressful on the system. Now everything is smooth. I kept only the temp control as I like the strategy.

My 290 also is jumping memory down to 150 and back, do you have any explanation what might be the reason for this behavior?!

Yes, powertune - when it's not configured properly mem-clock will throttle up and down (150-your mem speed).

I also had this issue and after looking for a solution for over a week I managed to fix it by enabling 'Unofficial overclocking mode without PowerPlay support'. I assume of-course that you are using MSI afterburner.
Uninstalling MSI AB might also fix it, or alternatively installing it if you haven't got it.

I made a topic about it a few days ago but no1 responded there; I had an image uploaded there, does it look similar to that? https://bitcointalk.org/index.php?topic=628425.msg6989373#msg6989373

Anyway, now I got a straight line representing my mem clock speed.

[platinum4]

Just ran across this issue and confirmed it on Hawaii chipsets as well.

memclock will jump frequently from 1500 -> 150 -> 1500 and there is a noticeable degradation that shows up in one of the card's hashrates in a multi-card rig.

Any thoughts, kernel smashers?

[platinum4]

I have 2 cards, both running at 1050/1500MHz set with Afterburner (power limit +50). I had to take out the engine and memory settings from the config file as the memory frequency was jumping around between 150 and max setting. Very stressful on the system. Now everything is smooth. I kept only the temp control as I like the strategy.

My 290 also is jumping memory down to 150 and back, do you have any explanation what might be the reason for this behavior?!

Yes, powertune - when it's not configured properly mem-clock will throttle up and down (150-your mem speed).

I also had this issue and after looking for a solution for over a week I managed to fix it by enabling 'Unofficial overclocking mode without PowerPlay support'. I assume of-course that you are using MSI afterburner.
Uninstalling MSI AB might also fix it, or alternatively installing it if you haven't got it.

I made a topic about it a few days ago but no1 responded there; I had an image uploaded there, does it look similar to that? https://bitcointalk.org/index.php?topic=628425.msg6989373#msg6989373

Anyway, now I got a straight line representing my mem clock speed.

Confirmed, followed the steps for MSI AB and now have a bit more hash at my disposal, with constant memclock at 1500.

Thanks for the advice, we'll see if hash degrades in the next 24hrs.

[Wolf_Pack]

The longest I have been able to get my rig to run is 12 hours. I don't know what's causing it, but it's freezing up and I have to do a hard reboot which isn't convenient since I normally use team viewer. I posted my settings at https://bitcointalk.org/index.php?topic=623409.msg7036775#msg7036775 if anyone can look at them and give me some suggestions. If the miner was just crashing I wouldn't care so much because that's easy to restart. It's just the hard reboots that are killing me.

lower your clocks. don't think you'll need 1500 mem anyways.

most of my cards run 1000/1200 @1.1v. Tweak the settings nicely, and they'll go as high as 2.5 mh/s

Thanks. I'll give it a shot. I was getting 2.2 mh/s on my 270 and 3.2 on my 280x. Hopefully I can get about the same with some stability.

[NRD]

I have 2 cards, both running at 1050/1500MHz set with Afterburner (power limit +50). I had to take out the engine and memory settings from the config file as the memory frequency was jumping around between 150 and max setting. Very stressful on the system. Now everything is smooth. I kept only the temp control as I like the strategy.

My 290 also is jumping memory down to 150 and back, do you have any explanation what might be the reason for this behavior?!

Yes, powertune - when it's not configured properly mem-clock will throttle up and down (150-your mem speed).

I also had this issue and after looking for a solution for over a week I managed to fix it by enabling 'Unofficial overclocking mode without PowerPlay support'. I assume of-course that you are using MSI afterburner.
Uninstalling MSI AB might also fix it, or alternatively installing it if you haven't got it.

I made a topic about it a few days ago but no1 responded there; I had an image uploaded there, does it look similar to that? https://bitcointalk.org/index.php?topic=628425.msg6989373#msg6989373

Anyway, now I got a straight line representing my mem clock speed.

Wow thanks for that, it's gotten me up to the 4.05-4.1mh/s level with my 290 tri-x's using my previous config but with 5 threads now instead of 1.

I tried other drivers but 12.13 seems to work best for me on Windows 7. Using the B2_sick version.

[Starscream]

I have 2 cards, both running at 1050/1500MHz set with Afterburner (power limit +50). I had to take out the engine and memory settings from the config file as the memory frequency was jumping around between 150 and max setting. Very stressful on the system. Now everything is smooth. I kept only the temp control as I like the strategy.

My 290 also is jumping memory down to 150 and back, do you have any explanation what might be the reason for this behavior?!

Yes, powertune - when it's not configured properly mem-clock will throttle up and down (150-your mem speed).

I also had this issue and after looking for a solution for over a week I managed to fix it by enabling 'Unofficial overclocking mode without PowerPlay support'. I assume of-course that you are using MSI afterburner.
Uninstalling MSI AB might also fix it, or alternatively installing it if you haven't got it.

I made a topic about it a few days ago but no1 responded there; I had an image uploaded there, does it look similar to that? https://bitcointalk.org/index.php?topic=628425.msg6989373#msg6989373

Anyway, now I got a straight line representing my mem clock speed.

Confirmed, followed the steps for MSI AB and now have a bit more hash at my disposal, with constant memclock at 1500.

Thanks for the advice, we'll see if hash degrades in the next 24hrs.

After I fixed the throttling issue my cards ran for at least 48 hours without an issue and with steady hash. I have since bumped up the memclock to 1500 (few hours ago) and so far the hash dropped once and got back to normal after restart. So since the restart 2.5 hours ago they been running fine.
Anyway, with 1250 memclock it was running smoothly after the fix. Will probably go back to 1250 if the hash crashes again.

[Starscream]

I have 2 cards, both running at 1050/1500MHz set with Afterburner (power limit +50). I had to take out the engine and memory settings from the config file as the memory frequency was jumping around between 150 and max setting. Very stressful on the system. Now everything is smooth. I kept only the temp control as I like the strategy.

My 290 also is jumping memory down to 150 and back, do you have any explanation what might be the reason for this behavior?!

Yes, powertune - when it's not configured properly mem-clock will throttle up and down (150-your mem speed).

I also had this issue and after looking for a solution for over a week I managed to fix it by enabling 'Unofficial overclocking mode without PowerPlay support'. I assume of-course that you are using MSI afterburner.
Uninstalling MSI AB might also fix it, or alternatively installing it if you haven't got it.

I made a topic about it a few days ago but no1 responded there; I had an image uploaded there, does it look similar to that? https://bitcointalk.org/index.php?topic=628425.msg6989373#msg6989373

Anyway, now I got a straight line representing my mem clock speed.

Wow thanks for that, it's gotten me up to the 4.05-4.1mh/s level with my 290 tri-x's using my previous config but with 5 threads now instead of 1.

I tried other drivers but 12.13 seems to work best for me on Windows 7. Using the B2_sick version.

You welcome. This issue bothered me for almost a month now

I can't get my R9 290 to run stable with 5 threads.. Than again 2 threads on 14.4 drivers are getting me 3.98MH/s so it isn't so bad I suppose for ref cards.

[NRD]

I have 2 cards, both running at 1050/1500MHz set with Afterburner (power limit +50). I had to take out the engine and memory settings from the config file as the memory frequency was jumping around between 150 and max setting. Very stressful on the system. Now everything is smooth. I kept only the temp control as I like the strategy.

My 290 also is jumping memory down to 150 and back, do you have any explanation what might be the reason for this behavior?!

Yes, powertune - when it's not configured properly mem-clock will throttle up and down (150-your mem speed).

I also had this issue and after looking for a solution for over a week I managed to fix it by enabling 'Unofficial overclocking mode without PowerPlay support'. I assume of-course that you are using MSI afterburner.
Uninstalling MSI AB might also fix it, or alternatively installing it if you haven't got it.

I made a topic about it a few days ago but no1 responded there; I had an image uploaded there, does it look similar to that? https://bitcointalk.org/index.php?topic=628425.msg6989373#msg6989373

Anyway, now I got a straight line representing my mem clock speed.

Wow thanks for that, it's gotten me up to the 4.05-4.1mh/s level with my 290 tri-x's using my previous config but with 5 threads now instead of 1.

I tried other drivers but 12.13 seems to work best for me on Windows 7. Using the B2_sick version.

You welcome. This issue bothered me for almost a month now

I can't get my R9 290 to run stable with 5 threads.. Than again 2 threads on 14.4 drivers are getting me 3.98MH/s so it isn't so bad I suppose for ref cards.

I'm running with 1 thread atm and getting the same speed, but rock solid stable speed of 4.029-4.031mh/s so you could try that. Here's what I'm running atm.

Code:

setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
export GPU_FORCE_64BIT_PTR=1
##export GPU_MAX_HEAP_SIZE=50
del *.bin
timeout /t 4
sgminer.exe -k x11mod --rawintensity 1306624,1306624,1306624,33280 -g 1,1,1,2 --shaders 2560,2560,2560,1280 -w 256,256,256,64 --thread-concurrency 24577,24577,24577,10241 --no-submit-stale -c sgminer.conf

[csa1234]

i have a question... how many MARUCOINS you can mine per day per MH?

[litecoinbeast]

Just sharing my 4 x 270x + 1 x 270 settings for x11mod (basically as per OP settings): aver 10.14 or ~ 2.03 / card (needed to decrease the settings on 270 for some reason as would get driver failures on that card ). Using Win7 x64, driver version 13.251.0.0, CCC completely removed/disabled. Using Trixx atm to manage mem and engine tweaks.

Code:

"gpu-powertune" : "20",
"intensity" : "16",
"kernel" : "x11mod",
"vectors" : "1",
"worksize" : "128",
"gpu-threads" : "2",
"thread-concurrency" : "8192",
"lookup-gap" : "2",
"gpu-engine" : "1100",
"gpu-memclock" : "1500",
"api-mcast-port" : "4028",
"failover-only" : true,
"api-port" : "4028",
"expiry" : "120",
"log" : "5",
"device" : "0,1,2,3,4",
"no-pool-disable" : false,
"no-submit-stale" : true,
"queue" : "1",
"scan-time" : "30",
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin"

Was only getting about 1.3 before using this release with modify kernels so cheers mate. Sure could increase the engine up to 1150 but seem to get a few more rejects.

[Amgclk65]

Just sharing my 4 x 270x + 1 x 270 settings for x11mod (basically as per OP settings): aver 10.14 or ~ 2.03 / card (needed to decrease the settings on 270 for some reason as would get driver failures on that card ). Using Win7 x64, driver version 13.251.0.0, CCC completely removed/disabled. Using Trixx atm to manage mem and engine tweaks.

Code:

"gpu-powertune" : "20",
"intensity" : "16",
"kernel" : "x11mod",
"vectors" : "1",
"worksize" : "128",
"gpu-threads" : "2",
"thread-concurrency" : "8192",
"lookup-gap" : "2",
"gpu-engine" : "1100",
"gpu-memclock" : "1500",
"api-mcast-port" : "4028",
"failover-only" : true,
"api-port" : "4028",
"expiry" : "120",
"log" : "5",
"device" : "0,1,2,3,4",
"no-pool-disable" : false,
"no-submit-stale" : true,
"queue" : "1",
"scan-time" : "30",
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin"

Was only getting about 1.3 before using this release with modify kernels so cheers mate. Sure could increase the engine up to 1150 but seem to get a few more rejects.

Thanks for sharing. Ill try it with my 6 270 rig. But whats CCC?
I think its catalyst control center ?

[svdinu]

I have 2 cards, both running at 1050/1500MHz set with Afterburner (power limit +50). I had to take out the engine and memory settings from the config file as the memory frequency was jumping around between 150 and max setting. Very stressful on the system. Now everything is smooth. I kept only the temp control as I like the strategy.

My 290 also is jumping memory down to 150 and back, do you have any explanation what might be the reason for this behavior?!

Yes, powertune - when it's not configured properly mem-clock will throttle up and down (150-your mem speed).

I also had this issue and after looking for a solution for over a week I managed to fix it by enabling 'Unofficial overclocking mode without PowerPlay support'. I assume of-course that you are using MSI afterburner.
Uninstalling MSI AB might also fix it, or alternatively installing it if you haven't got it.

I made a topic about it a few days ago but no1 responded there; I had an image uploaded there, does it look similar to that? https://bitcointalk.org/index.php?topic=628425.msg6989373#msg6989373

Anyway, now I got a straight line representing my mem clock speed.

Yes, it looked like in your picture. However, I did not use to keep Afterburner running. Installed it, changed settings and took it off. Just noticed yesterday this strange memory behaviour when trying to look for a sweet spot for my cards frequencies (used this: http://www.reddit.com/r/litecoinmining/comments/1uv67j/cgbenchmark/) and I got errors on memory setting.

After more than 8.5 hours of only mining, my system looks like that:

- hashes somewhere between 4.1 and 4.2:

https://www.dropbox.com/s/k9e9yhd0gaal13u/Hashes%20June%201.png

- pretty stable gpus, but a strange slow RAM increase:

https://www.dropbox.com/s/cts1eqlwjgsuy7n/GPUsMonitor%20June%201.png

- more stable received pool hash:

https://www.dropbox.com/s/sj6u3mwe8c4eqwm/PoolHashJune1.png

- still a restless system (overall power consumption), which was not the case on cgminer 4.7.3 times:

https://www.dropbox.com/s/1mnawjy5k9r8nxy/PowerconsumptionJune1.png

I have no idea what's going on. One thing is certain, stabilizing the memory frequency got me a hash increase (3.6 -> 4.1).

Just to sum up the system config: 2 x ASUS R9290DC2OC, AMD 14.3beta, Win 8.1.

Unfortunately I'm just a user so it's hard for me to figure out what induces stress on my PC or how big this stress is so I can estimate the risks of breaking it.

[svdinu]

I have 2 cards, both running at 1050/1500MHz set with Afterburner (power limit +50). I had to take out the engine and memory settings from the config file as the memory frequency was jumping around between 150 and max setting. Very stressful on the system. Now everything is smooth. I kept only the temp control as I like the strategy.

My 290 also is jumping memory down to 150 and back, do you have any explanation what might be the reason for this behavior?!

Yes, powertune - when it's not configured properly mem-clock will throttle up and down (150-your mem speed).

I also had this issue and after looking for a solution for over a week I managed to fix it by enabling 'Unofficial overclocking mode without PowerPlay support'. I assume of-course that you are using MSI afterburner.
Uninstalling MSI AB might also fix it, or alternatively installing it if you haven't got it.

I made a topic about it a few days ago but no1 responded there; I had an image uploaded there, does it look similar to that? https://bitcointalk.org/index.php?topic=628425.msg6989373#msg6989373

Anyway, now I got a straight line representing my mem clock speed.

Wow thanks for that, it's gotten me up to the 4.05-4.1mh/s level with my 290 tri-x's using my previous config but with 5 threads now instead of 1.

I tried other drivers but 12.13 seems to work best for me on Windows 7. Using the B2_sick version.

You welcome. This issue bothered me for almost a month now

I can't get my R9 290 to run stable with 5 threads.. Than again 2 threads on 14.4 drivers are getting me 3.98MH/s so it isn't so bad I suppose for ref cards.

I'm running with 1 thread atm and getting the same speed, but rock solid stable speed of 4.029-4.031mh/s so you could try that. Here's what I'm running atm.

Code:

setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
export GPU_FORCE_64BIT_PTR=1
##export GPU_MAX_HEAP_SIZE=50
del *.bin
timeout /t 4
sgminer.exe -k x11mod --rawintensity 1306624,1306624,1306624,33280 -g 1,1,1,2 --shaders 2560,2560,2560,1280 -w 256,256,256,64 --thread-concurrency 24577,24577,24577,10241 --no-submit-stale -c sgminer.conf

People on forums are claiming gpu threads help keep the reject rate low. Currently I'm below or max at pool level, which gives me an edge.

[ozzy1926]

I have 2 cards, both running at 1050/1500MHz set with Afterburner (power limit +50). I had to take out the engine and memory settings from the config file as the memory frequency was jumping around between 150 and max setting. Very stressful on the system. Now everything is smooth. I kept only the temp control as I like the strategy.

My 290 also is jumping memory down to 150 and back, do you have any explanation what might be the reason for this behavior?!

Yes, powertune - when it's not configured properly mem-clock will throttle up and down (150-your mem speed).

I also had this issue and after looking for a solution for over a week I managed to fix it by enabling 'Unofficial overclocking mode without PowerPlay support'. I assume of-course that you are using MSI afterburner.
Uninstalling MSI AB might also fix it, or alternatively installing it if you haven't got it.

I made a topic about it a few days ago but no1 responded there; I had an image uploaded there, does it look similar to that? https://bitcointalk.org/index.php?topic=628425.msg6989373#msg6989373

Anyway, now I got a straight line representing my mem clock speed.

Yes, it looked like in your picture. However, I did not use to keep Afterburner running. Installed it, changed settings and took it off. Just noticed yesterday this strange memory behaviour when trying to look for a sweet spot for my cards frequencies (used this: http://www.reddit.com/r/litecoinmining/comments/1uv67j/cgbenchmark/) and I got errors on memory setting.

After more than 8.5 hours of only mining, my system looks like that:

- hashes somewhere between 4.1 and 4.2:

https://www.dropbox.com/s/k9e9yhd0gaal13u/Hashes%20June%201.png

- pretty stable gpus, but a strange slow RAM increase:

https://www.dropbox.com/s/cts1eqlwjgsuy7n/GPUsMonitor%20June%201.png

- more stable received pool hash:

https://www.dropbox.com/s/sj6u3mwe8c4eqwm/PoolHashJune1.png

- still a restless system (overall power consumption), which was not the case on cgminer 4.7.3 times:

https://www.dropbox.com/s/1mnawjy5k9r8nxy/PowerconsumptionJune1.png

I have no idea what's going on. One thing is certain, stabilizing the memory frequency got me a hash increase (3.6 -> 4.1).

Just to sum up the system config: 2 x ASUS R9290DC2OC, AMD 14.3beta, Win 8.1.

Unfortunately I'm just a user so it's hard for me to figure out what induces stress on my PC or how big this stress is so I can estimate the risks of breaking it.

whats your setting for 4.2mh/s?

[svdinu]

whats your setting for 4.2mh/s?
[/quote]

"no-restart" : true,
"failover-only" : true,
"failover-switch-delay" : "60",
"api-listen" : true,
"api-port" : "4028",
"api-allow" : "W:127.0.0.1,R:0/0",
"xintensity" : "700",
"worksize" : "128,128",
"kernel" : "x11mod,x11mod",
"auto-fan" : true,
"temp-cutoff" : "81",
"temp-overheat" : "76",
"temp-target" : "71",
"temp-hysteresis" : "3",
"expiry" : "1",
"gpu-dyninterval" : "7",
"queue" : "0",
"scan-time" : "1",
"shaders" : "2560",
"thread-concurrency" : "25600",
"gpu-threads" : "5",
"lookup-gap" : "2"

NRD has some nice settings too, on page 41.