[ANN] sgminer v5 - optimized X11/X13/NeoScrypt/Lyra2RE/etc. kernel-switch miner

[scryptr]

NeoScrypt --

When i issue the command "git clone --recursive https://github.com/sgminer-dev/sgminer.git", the cloned "sgminer" directory does not have the "neoscrypt.*" files in the "kernel" or "algorithm" directories.

ystarnaud has added the "neoscrypt.cl" file from Wolf9466 a few hours ago.  The new "neoscrypt.*" files can be seen in the directories mentioned above via ystarnaud's github post.

How do I clone ystarnaud's latest edition?       --scryptr

try
git clone -b develop --recursive git://github.com/sgminer-dev/sgminer.git

Thanks for your informative reply, Badman.  The "sgminer" directory that I cloned had the latest neoscypt files, and compiled readily.  I am running Ubuntu 14.04.1 with AMD 14.9 drivers.  The compiled sgminer 5.1.0-dev-2-gf608 flew with worksize 64 and I=14.  NO HW ERRORS, they were a big problem when trying to use the master branch.  I did not have to change my drivers, either.       --scryptr

[MTJ151]

Anyone tried the new 14.11 drivers??

[badman74]

for anyone that is still having issues with HW errors on various algo's try my latest binary
i think i may have found the problem but would like others to test and see if it is fixed
also, arebyp kernel can now be used for nscrypt although it seems to put out less hashrate than zuikkis, though it produces less errors for me

[Zuikkis]

for anyone that is still having issues with HW errors on various algo's try my latest binary
i think i may have found the problem but would like others to test and see if it is fixed
also, arebyp kernel can now be used for nscrypt although it seems to put out less hashrate than zuikkis, though it produces less errors for me

Hi,

I also fixed it in my local build, is this the same you found? This fixed HW errors 100% for me.

sgminer.c, function test_nonce, near line 6873:

Code:

  // for Neoscrypt, the diff1targe value is in work->target
  if ((work->pool->algorithm.name, "neoscrypt")) {
    diff1targ = ((uint32_t *)work->target)[7];
  } else {
    diff1targ = work->pool->algorithm.diff1targ;
  }

Should be like this, just add !safe_cmp:

Code:

  // for Neoscrypt, the diff1targe value is in work->target
  if (!safe_cmp(work->pool->algorithm.name, "neoscrypt")) {
    diff1targ = ((uint32_t *)work->target)[7];
  } else {
    diff1targ = work->pool->algorithm.diff1targ;
  }

The original version is really a no-op, always true. So neoscrypt code was used for all algorithms.

[Zuikkis]

Oh, anyone have the new neoscrypt kernel binary for Pitcairn?

I have Catalyst 14.9, it really can't compile the kernel properly. The binaries above fixed it for Hawaii and Tahiti, but I have a few Pitcairns as well..

[damm315er]

Oh, anyone have the new neoscrypt kernel binary for Pitcairn?

I have Catalyst 14.9, it really can't compile the kernel properly. The binaries above fixed it for Hawaii and Tahiti, but I have a few Pitcairns as well..

Try what is linked here for the older cards.

https://forum.feathercoin.com/index.php?/topic/7780-neoscrypt-gpu-miner-public-beta-test/?p=70067

[mitache365]

can someone share batch file for 280x cards? thanks!

[oktay50000]

any luck something like this for x11 x13 x15

[badman74]

for anyone that is still having issues with HW errors on various algo's try my latest binary
i think i may have found the problem but would like others to test and see if it is fixed
also, arebyp kernel can now be used for nscrypt although it seems to put out less hashrate than zuikkis, though it produces less errors for me

Hi,

I also fixed it in my local build, is this the same you found? This fixed HW errors 100% for me.

sgminer.c, function test_nonce, near line 6873:

Code:

  // for Neoscrypt, the diff1targe value is in work->target
  if ((work->pool->algorithm.name, "neoscrypt")) {
    diff1targ = ((uint32_t *)work->target)[7];
  } else {
    diff1targ = work->pool->algorithm.diff1targ;
  }

Should be like this, just add !safe_cmp:

Code:

  // for Neoscrypt, the diff1targe value is in work->target
  if (!safe_cmp(work->pool->algorithm.name, "neoscrypt")) {
    diff1targ = ((uint32_t *)work->target)[7];
  } else {
    diff1targ = work->pool->algorithm.diff1targ;
  }

The original version is really a no-op, always true. So neoscrypt code was used for all algorithms.

hmm that wasn't what i changed but i will do that tonight

[Prelude]

For some reason, since I added neoscrypt to my config, my miner isn't auto switching to x11, x13, or x15 anymore when they are more profitable. Is there something wrong below?

Code:

{
"pools" : [
        {
        	"name" : "WestHash_X11_auto",
        	"url" : "stratum+tcp://stratum.westhash.com:4336",
          	"user" : "1MDRjWZxttXRDYz3aAf1GjUY9JwscgRnFL.1",
         	"pass" : "f0=0;f2=0;f3=4;f4=3;f5=0;f6=2;f7=0;f8=1",
          	"xintensity": "144",
		"worksize" : "128",
		"gpu-threads" : "2",
          	"kernel" : "darkcoin-mod"
	},
	{
        	"name" : "WestHash_X13_auto",
          	"url" : "stratum+tcp://stratum.westhash.com:4337",
         	"user" : "1MDRjWZxttXRDYz3aAf1GjUY9JwscgRnFL.1",
         	"pass" : "f0=0;f2=0;f3=4;f4=3;f5=0;f6=2;f7=0;f8=1",
         	"xintensity": "144",
		"worksize" : "128",
		"gpu-threads" : "2",
         	"kernel" : "marucoin-mod"
	},
	{
         	"name" : "WestHash_X15_auto",
         	"url" : "stratum+tcp://stratum.westhash.com:4339",
         	"user" : "1MDRjWZxttXRDYz3aAf1GjUY9JwscgRnFL.1",
         	"pass" : "f0=0;f2=0;f3=4;f4=3;f5=0;f6=2;f7=0;f8=1",
         	"xintensity": "144",
		"worksize" : "128",
		"gpu-threads" : "2",
         	"kernel" : "bitblock"
	},
	{
         	"name" : "WestHash_NeoScrypt_auto",
         	"url" : "stratum+tcp://stratum.westhash.com:4341",
         	"user" : "1MDRjWZxttXRDYz3aAf1GjUY9JwscgRnFL.1",
         	"pass" : "d=16;f0=0;f2=0;f3=4;f4=3;f5=0;f6=2;f7=0;f8=1",
		"nfactor" : "10",
		"intensity": "13",
		"worksize" : "64",
		"gpu-threads" : "2",
         	"kernel" : "neoscrypt"
	},
	{
         	"name" : "WestHash_X11",
         	"url" : "stratum+tcp://stratum.westhash.com:3336",
         	"user" : "1MDRjWZxttXRDYz3aAf1GjUY9JwscgRnFL.1",
         	"pass" : "d=0.02",
         	"nfactor" : "10",
		"xintensity": "144",
		"worksize" : "128",
		"gpu-threads" : "2",
         	"kernel" : "darkcoin-mod"
	}
]
,
"lookup-gap" : "2",
"thread-concurrency" : "8192",
"api-mcast-port" : "4028",
"api-port" : "4028",
"expiry" : "28",
"failover-switch-delay" : "60",
"queue" : "0",
"scan-time" : "7",
"tcp-keepalive" : "30",
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/sgminer-5-dev-neoscrypt-windows/kernel"
}

[oktay50000]

For some reason, since I added neoscrypt to my config, my miner isn't auto switching to x11, x13, or x15 anymore when they are more profitable. Is there something wrong below?

Code:

{
"pools" : [
        {
        	"name" : "WestHash_X11_auto",
        	"url" : "stratum+tcp://stratum.westhash.com:4336",
          	"user" : "1MDRjWZxttXRDYz3aAf1GjUY9JwscgRnFL.1",
         	"pass" : "f0=0;f2=0;f3=4;f4=3;f5=0;f6=2;f7=0;f8=1",
          	"xintensity": "144",
		"worksize" : "128",
		"gpu-threads" : "2",
          	"kernel" : "darkcoin-mod"
	},
	{
        	"name" : "WestHash_X13_auto",
          	"url" : "stratum+tcp://stratum.westhash.com:4337",
         	"user" : "1MDRjWZxttXRDYz3aAf1GjUY9JwscgRnFL.1",
         	"pass" : "f0=0;f2=0;f3=4;f4=3;f5=0;f6=2;f7=0;f8=1",
         	"xintensity": "144",
		"worksize" : "128",
		"gpu-threads" : "2",
         	"kernel" : "marucoin-mod"
	},
	{
         	"name" : "WestHash_X15_auto",
         	"url" : "stratum+tcp://stratum.westhash.com:4339",
         	"user" : "1MDRjWZxttXRDYz3aAf1GjUY9JwscgRnFL.1",
         	"pass" : "f0=0;f2=0;f3=4;f4=3;f5=0;f6=2;f7=0;f8=1",
         	"xintensity": "144",
		"worksize" : "128",
		"gpu-threads" : "2",
         	"kernel" : "bitblock"
	},
	{
         	"name" : "WestHash_NeoScrypt_auto",
         	"url" : "stratum+tcp://stratum.westhash.com:4341",
         	"user" : "1MDRjWZxttXRDYz3aAf1GjUY9JwscgRnFL.1",
         	"pass" : "d=16;f0=0;f2=0;f3=4;f4=3;f5=0;f6=2;f7=0;f8=1",
		"nfactor" : "10",
		"intensity": "13",
		"worksize" : "64",
		"gpu-threads" : "2",
         	"kernel" : "neoscrypt"
	},
	{
         	"name" : "WestHash_X11",
         	"url" : "stratum+tcp://stratum.westhash.com:3336",
         	"user" : "1MDRjWZxttXRDYz3aAf1GjUY9JwscgRnFL.1",
         	"pass" : "d=0.02",
         	"nfactor" : "10",
		"xintensity": "144",
		"worksize" : "128",
		"gpu-threads" : "2",
         	"kernel" : "darkcoin-mod"
	}
]
,
"lookup-gap" : "2",
"thread-concurrency" : "8192",
"api-mcast-port" : "4028",
"api-port" : "4028",
"expiry" : "28",
"failover-switch-delay" : "60",
"queue" : "0",
"scan-time" : "7",
"tcp-keepalive" : "30",
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/sgminer-5-dev-neoscrypt-windows/kernel"
}

you have to set f8=0.15 in your password mate

[Zuikkis]

Oh, anyone have the new neoscrypt kernel binary for Pitcairn?

I have Catalyst 14.9, it really can't compile the kernel properly. The binaries above fixed it for Hawaii and Tahiti, but I have a few Pitcairns as well..

I fixed this myself, simply by installing 14.6.. It is actually available from AMD web page, just not publicly listed anymore! But easy to find with google.

14.6 is so much better than 14.9.. X11 hashrate went from 4.0Mh/s to 4.4MH/s, similar increase in other coins as well.. Just delete all old binaries.

280x Scrypt and Nscrypt bins are far better when compiled with catalyst 13.11, luckily I still had those stored.

[nickdino]

Hi yall, i have 5x280x by sapphire, its hasn't been running for more than a month. I was using guiminer and pointing to coinshift but the cards can't connect anymore. So now i want to get them running again, best thing would probably be to get my rig on another profit switching multipool, im considering multipool.us at this point and ofcourse i'll need a miner that can handle the switching to different algorithms. Could you guys please link me to the best one out there for my needs?  It has to be a very stable miner and preferrably with gui, thx!!

[nicehash]

Hi yall, i have 5x280x by sapphire, its hasn't been running for more than a month. I was using guiminer and pointing to coinshift but the cards can't connect anymore. So now i want to get them running again, best thing would probably be to get my rig on another profit switching multipool, im considering multipool.us at this point and ofcourse i'll need a miner that can handle the switching to different algorithms. Could you guys please link me to the best one out there for my needs?  It has to be a very stable miner and preferrably with gui, thx!!

If this is a dedicated rig simply get PiMP on an USB stick, activate NiceHash/WestHash multi-algo profile and you're ready to go (with a simple GUI )

http://www.getpimp.org/downloads/pimp-amd/pimp-amd-1-6-0.html

http://www.getpimp.org/community/blog/165-nicehash.html

Many other options are listed here: https://www.nicehash.com/software/

[Anobix]

Hi yall, i have 5x280x by sapphire, its hasn't been running for more than a month. I was using guiminer and pointing to coinshift but the cards can't connect anymore. So now i want to get them running again, best thing would probably be to get my rig on another profit switching multipool, im considering multipool.us at this point and ofcourse i'll need a miner that can handle the switching to different algorithms. Could you guys please link me to the best one out there for my needs?  It has to be a very stable miner and preferrably with gui, thx!!

If this is a dedicated rig simply get PiMP on an USB stick, activate NiceHash/WestHash multi-algo profile and you're ready to go (with a simple GUI )

http://www.getpimp.org/downloads/pimp-amd/pimp-amd-1-6-0.html

http://www.getpimp.org/community/blog/165-nicehash.html

Many other options are listed here: https://www.nicehash.com/software/

Does that work with Neoscrypt though?

edit: apparently not yet
http://www.getpimp.org/features.html

[badman74]

Hi yall, i have 5x280x by sapphire, its hasn't been running for more than a month. I was using guiminer and pointing to coinshift but the cards can't connect anymore. So now i want to get them running again, best thing would probably be to get my rig on another profit switching multipool, im considering multipool.us at this point and ofcourse i'll need a miner that can handle the switching to different algorithms. Could you guys please link me to the best one out there for my needs?  It has to be a very stable miner and preferrably with gui, thx!!

If this is a dedicated rig simply get PiMP on an USB stick, activate NiceHash/WestHash multi-algo profile and you're ready to go (with a simple GUI )

http://www.getpimp.org/downloads/pimp-amd/pimp-amd-1-6-0.html

http://www.getpimp.org/community/blog/165-nicehash.html

Many other options are listed here: https://www.nicehash.com/software/

Does that work with Neoscrypt though?

edit: apparently not yet
http://www.getpimp.org/features.html

I will look into this but I think if I clone my master branch to my develop then you could use get my miner from miner menu and it would have neoscrypt even though they haven't released it

[Slix]

for anyone that is still having issues with HW errors on various algo's try my latest binary
i think i may have found the problem but would like others to test and see if it is fixed
also, arebyp kernel can now be used for nscrypt although it seems to put out less hashrate than zuikkis, though it produces less errors for me

Hi,

I also fixed it in my local build, is this the same you found? This fixed HW errors 100% for me.

sgminer.c, function test_nonce, near line 6873:

Code:

  // for Neoscrypt, the diff1targe value is in work->target
  if ((work->pool->algorithm.name, "neoscrypt")) {
    diff1targ = ((uint32_t *)work->target)[7];
  } else {
    diff1targ = work->pool->algorithm.diff1targ;
  }

Should be like this, just add !safe_cmp:

Code:

  // for Neoscrypt, the diff1targe value is in work->target
  if (!safe_cmp(work->pool->algorithm.name, "neoscrypt")) {
    diff1targ = ((uint32_t *)work->target)[7];
  } else {
    diff1targ = work->pool->algorithm.diff1targ;
  }

The original version is really a no-op, always true. So neoscrypt code was used for all algorithms.

Thank you so much for pointing this out. (Although it would have been nice to get a pull request or a notice on github lol...)

I must have pushed the wrong file up originally and when debugging the issue I was using corrected code. No wonder I didn't get HW errors. Chasing a bug that isn't there is can be quite a headache...

Anyway, I pushed the fix to the develop branch. Can anybody try it and confirm that the HW errors are gone?

- ystarnaud

[kotarius]

Oh, anyone have the new neoscrypt kernel binary for Pitcairn?

I have Catalyst 14.9, it really can't compile the kernel properly. The binaries above fixed it for Hawaii and Tahiti, but I have a few Pitcairns as well..

I fixed this myself, simply by installing 14.6.. It is actually available from AMD web page, just not publicly listed anymore! But easy to find with google.

14.6 is so much better than 14.9.. X11 hashrate went from 4.0Mh/s to 4.4MH/s, similar increase in other coins as well.. Just delete all old binaries.

280x Scrypt and Nscrypt bins are far better when compiled with catalyst 13.11, luckily I still had those stored.

[1:58] sling00: is what zuikkis says true about 13.11 building better bins than 14.x for neoscrypt? if so i may need to setup a bin repo for neoscrypt
[1:58] kotarius: sling00:  you got a link to histatement
[1:58] ystarnaud: in  my experience no
[1:58] ystarnaud: thats not true
[1:58] ystarnaud: 13.11 is slow
[2:00] kotarius: maybe he means its faster relative to what he has been using
[2:00] kotarius: we would need to know his gpus, hashrates, before and after on 13.11 and 14.6

[Zuikkis]

Not neoscrypt, regular scrypt and scrypt-n!

I said "NScrypt" incorrectly when I ment scrypt nfactor, sorry.

I have been away from altcoin business too long.

But for scrypt, the difference is obvious.. 13.11 bin gives about 800kh/s on my 280x, and 700kh/s if compiled under 14.6, from the same source.

Always backup your bins when upgrading Catalyst, then you can try each version. Only compile time matters, there usually is no big difference what version Catalyst you are using at runtime..

[Slix]

Not neoscrypt, regular scrypt and scrypt-n!

I said "NScrypt" incorrectly when I ment scrypt nfactor, sorry.

I have been away from altcoin business too long.

But for scrypt, the difference is obvious.. 13.11 bin gives about 800kh/s on my 280x, and 700kh/s if compiled under 14.6, from the same source.

Always backup your bins when upgrading Catalyst, then you can try each version. Only compile time matters, there usually is no big difference what version Catalyst you are using at runtime..

It seems that the compiler changed a lot in 14.x. This might account for the difference. The scrypt/scrypt-n kernels should probably be revised and optimized for the new compiler.