[ANN][X11/X13] X11 (Darkcoin)/X13 (Marucoin) miner (based on sph-sgminer) - page 49.

ozzy1926

sr. member

Activity: 378

Merit: 250

Quote from: Elun on May 30, 2014, 05:49:39 AM

Quote from: ozzy1926 on May 30, 2014, 04:28:31 AM

Quote from: Elun on May 30, 2014, 04:02:28 AM

Quote from: dragonmike on May 30, 2014, 03:46:36 AM

Quote from: ozzy1926 on May 30, 2014, 01:22:22 AM

although watt meter says all cards working after 7-8 hurs miner freeze and pool hashrate is 0anybody having the same issue?

Yes.
It runs well for a couple of hours and suddenly hashrate goes ballistic (1Gh/sec) but nothing gets sent to the pool anymore. No idea what causes it.
EDIT: it would seem that the card is going sick.

Can i see your miner log?

my cards arent sick they continue after restarting miner and where is the miner log i can post

Described here: https://bitcointalksearch.org/topic/m.7006129
But it's not acceptable for you, because it works fine in hours. Is there something interesting in miner windows, when it hangs?

nothing happens miner window open but kh/s number and other numers freeze
btw my cards are 290

pallas

legendary

Activity: 2716

Merit: 1094

Black Belt Developer

Quote from: lasybear on May 30, 2014, 06:23:55 AM

Quote from: platinum4 on May 30, 2014, 06:08:44 AM

I'll say it once and only once, from personal experience.

Thread concurrency for Hawaii chipsets (290/290X) should be 14080 not 8192.

You'll be surprised, but thread-concurrency (and lookup gap too) is affected only for scrypt kernels )

ocl.c

Code:

        sprintf(CompilerOptions, "-I \"%s\" -I \"%s\" -I \"%skernel\" -I \".\" -D LOOKUP_GAP=%d -D CONCURRENT_THREADS=%d -D WORKSIZE=%d",
                        opt_kernel_path, sgminer_path, sgminer_path,
                        cgpu->lookup_gap, (unsigned int)cgpu->thread_concurrency, (int)clState->wsize);

Neither x11, nor x13 are not using external definition of CONCURRENT_THREADS.

nor they are using the allocated global buffer of that size.

lasybear

member

Activity: 113

Merit: 14

Quote from: platinum4 on May 30, 2014, 06:08:44 AM

I'll say it once and only once, from personal experience.

Thread concurrency for Hawaii chipsets (290/290X) should be 14080 not 8192.

You'll be surprised, but thread-concurrency (and lookup gap too) is affected only for scrypt kernels )

ocl.c

Code:

        sprintf(CompilerOptions, "-I \"%s\" -I \"%s\" -I \"%skernel\" -I \".\" -D LOOKUP_GAP=%d -D CONCURRENT_THREADS=%d -D WORKSIZE=%d",
                        opt_kernel_path, sgminer_path, sgminer_path,
                        cgpu->lookup_gap, (unsigned int)cgpu->thread_concurrency, (int)clState->wsize);

Neither x11, nor x13 are not using external definition of CONCURRENT_THREADS.

restless

legendary

Activity: 1151

Merit: 1001

Quote from: platinum4 on May 30, 2014, 06:08:44 AM

I'll say it once and only once, from personal experience.

Thread concurrency for Hawaii chipsets (290/290X) should be 14080 not 8192.

Why 14080 !?!?
290 has 2560 shaders... so tc should be I*2560, where I is whole number

platinum4

sr. member

Activity: 547

Merit: 250

I'll say it once and only once, from personal experience.

Thread concurrency for Hawaii chipsets (290/290X) should be 14080 not 8192.

platinum4

sr. member

Activity: 547

Merit: 250

Guys there's no problem when your GPUs get set to idle, that has to do with how certain profit-switching pools serve their stratum work.

Ideally, if you D/C from a pool and there's no work available from any pools, then the GPUs will be set to idle if have_ADL.

So you may experience this when you are hooked up to TMB, this was a patch I had requested from Kalroth's 3.7.3 scrypt cgminer fork, which is often used on multipools and profit-switching stratums.

Make sure you have ample failovers and you'll be good.

ELUN keep up the great work bro. x11mod and x13mod and all recent patches asked for have been implemented, Elun, lasy, and RRT5ON are on fire the past few days.

ellave

sr. member

Activity: 434

Merit: 250

What people getting for x13 on a 290, I have 3 x tri x & can only get 2.7 from them

this is what I settled on after trying various configs but seems low as I get 2.5 from my 280x

"worksize": "128",
"intensity": "18",
"thread-concurrency": "8192",
"gpu-threads" : "4",
"gpu-engine" : "1050",
"gpu-memclock" : "1400",
"gpu-powertune" : "18",
"api-port": "4028",
"api-listen": true,
"api-allow": "W:127.0.0.1",

"failover-only" : true,
"no-pool-disable": true,
"no-submit-stale" : true,

"queue": "0",

"kernel" : "x13mod",
"kernel-path": "/usr/local/bin"

GunPowder

full member

Activity: 168

Merit: 100

Quote from: Elun on May 30, 2014, 05:49:39 AM

I'm using trademybitpool with x11mod and it words fine. 17 hours, no hang. It's so strange. Do you use sick time extended version for x11?

Yes, I've compiled it from git a few days ago, x13 is included, so sick extension is included also, as the both were implemented in one time.

Elun

member

Activity: 117

Merit: 10

Quote from: ozzy1926 on May 30, 2014, 04:28:31 AM

Quote from: Elun on May 30, 2014, 04:02:28 AM

Quote from: dragonmike on May 30, 2014, 03:46:36 AM

Quote from: ozzy1926 on May 30, 2014, 01:22:22 AM

although watt meter says all cards working after 7-8 hurs miner freeze and pool hashrate is 0anybody having the same issue?

Yes.
It runs well for a couple of hours and suddenly hashrate goes ballistic (1Gh/sec) but nothing gets sent to the pool anymore. No idea what causes it.
EDIT: it would seem that the card is going sick.

Can i see your miner log?

my cards arent sick they continue after restarting miner and where is the miner log i can post

Described here: https://bitcointalksearch.org/topic/m.7006129
But it's not acceptable for you, because it works fine in hours. Is there something interesting in miner windows, when it hangs?

Quote from: GunPowder on May 30, 2014, 05:25:54 AM

Guys,
does anybody have experience with "x11mod" kernel and multipool? I have an issue the miner just stops after a while of mining. All cards in the farm stop in one moment. I suppose it depends to coin switching.
"x13mod" kernal does ok with that.
Any idea?

I'm using trademybitpool with x11mod and it words fine. 17 hours, no hang. It's so strange. Do you use sick time extended version for x11?

I have a question. Now i'm trying to figure out HW on 5xxx cards. I have 5770, but i can't run marucoin kernel for tests. It's just doesn't compile. Error -11: Building Program (clBuildProgram)
And marucoin.cl in kernel folder, it's compiles fine for 280x. Anyone have same card? If yes, can you compile marucoin kernel?

Sorry for my bad english.

GunPowder

full member

Activity: 168

Merit: 100

Guys,
does anybody have experience with "x11mod" kernel and multipool? I have an issue the miner just stops after a while of mining. All cards in the farm stop in one moment. I suppose it depends to coin switching.
"x13mod" kernal does ok with that.
Any idea?

ozzy1926

sr. member

Activity: 378

Merit: 250

Quote from: Elun on May 30, 2014, 04:02:28 AM

Quote from: dragonmike on May 30, 2014, 03:46:36 AM

Quote from: ozzy1926 on May 30, 2014, 01:22:22 AM

although watt meter says all cards working after 7-8 hurs miner freeze and pool hashrate is 0anybody having the same issue?

Yes.
It runs well for a couple of hours and suddenly hashrate goes ballistic (1Gh/sec) but nothing gets sent to the pool anymore. No idea what causes it.
EDIT: it would seem that the card is going sick.

Can i see your miner log?

my cards arent sick they continue after restarting miner and where is the miner log i can post

Elun

member

Activity: 117

Merit: 10

Quote from: dragonmike on May 30, 2014, 03:46:36 AM

Quote from: ozzy1926 on May 30, 2014, 01:22:22 AM

although watt meter says all cards working after 7-8 hurs miner freeze and pool hashrate is 0anybody having the same issue?

Yes.
It runs well for a couple of hours and suddenly hashrate goes ballistic (1Gh/sec) but nothing gets sent to the pool anymore. No idea what causes it.
EDIT: it would seem that the card is going sick.

Can i see your miner log?

pallas

legendary

Activity: 2716

Merit: 1094

Black Belt Developer

Quote from: Starlightbreaker on May 29, 2014, 09:40:08 PM

does "vectors" setting matter?

on old version of cgminer, memory speed & vectors settings affect the hashrate, but doesn't seem like there's any significant improvement from 1, 2, or 4.

i gained only ~0.075 mh/s changing it from 1 to 4.

I believe vectors matters only for sha256, and surely not for x11 or x13. Those have fixed 32 and 64 bit integers and do not make use of integer vectors.
The difference in hashrate you experienced is probably just compiler variance.
If you delete the bin files and relaunch the miner, you'll see they probably have a different size and different speed.
That's why I always try 3/4 times before settling on the fastest ones.

dragonmike

hero member

Activity: 1274

Merit: 556

Quote from: ozzy1926 on May 30, 2014, 01:22:22 AM

although watt meter says all cards working after 7-8 hurs miner freeze and pool hashrate is 0anybody having the same issue?

Yes.
It runs well for a couple of hours and suddenly hashrate goes ballistic (1Gh/sec) but nothing gets sent to the pool anymore. No idea what causes it.
EDIT: it would seem that the card is going sick.

bengx

full member

Activity: 202

Merit: 100

Quote from: restless on May 30, 2014, 02:30:26 AM

Quote from: SpeedDemon13 on May 29, 2014, 12:16:31 PM

Quote from: bullus on May 29, 2014, 12:11:40 PM

Quote from: Remember remember the 5th of November on May 29, 2014, 10:05:08 AM

Quote from: restless on May 29, 2014, 09:39:28 AM

Had someone succeeded on running htis mod on older gen Radeons?

I tried on 5850 and 6970
It compiles, runs and NO accepted shares. Just HW increasing

Previous version, which was giving 2-2.1MH on 290 runs fine on these radeons
Using Cat 14.4 because prev version needed these to increase performance (on 13.12 there was no gain)

Yes, I have a 5850 and 14.4 drivers, x11mod works just fine, just from some default settings and underclock gets me 1.07Mh/s. Accepted and all that.

I think he meant X13mod.

x13mod kernel is a no go at the moment for the VLIW cards, ie. 5xxx/6xxx series....

But first x13mod version (the messy one, 281KB) - works fine and gave good boost - 6970 went from 1M to 1.4M
Or you meant later improvements are GCN specific?

Guys, can someone post speed&settings with stock 290? I can't get near 2.5 on stock (950/1250) :S

PS: seems that -w 128 is the key for extra performance.

first x13mod version? the one that came with the marucoin miner?

restless

legendary

Activity: 1151

Merit: 1001

Quote from: SpeedDemon13 on May 29, 2014, 12:16:31 PM

Quote from: bullus on May 29, 2014, 12:11:40 PM

Quote from: Remember remember the 5th of November on May 29, 2014, 10:05:08 AM

Quote from: restless on May 29, 2014, 09:39:28 AM

Had someone succeeded on running htis mod on older gen Radeons?

I tried on 5850 and 6970
It compiles, runs and NO accepted shares. Just HW increasing

Previous version, which was giving 2-2.1MH on 290 runs fine on these radeons
Using Cat 14.4 because prev version needed these to increase performance (on 13.12 there was no gain)

Yes, I have a 5850 and 14.4 drivers, x11mod works just fine, just from some default settings and underclock gets me 1.07Mh/s. Accepted and all that.

I think he meant X13mod.

x13mod kernel is a no go at the moment for the VLIW cards, ie. 5xxx/6xxx series....

But first x13mod version (the messy one, 281KB) - works fine and gave good boost - 6970 went from 1M to 1.4M
Or you meant later improvements are GCN specific?

Guys, can someone post speed&settings with stock 290? I can't get near 2.5 on stock (950/1250) :S

PS: seems that -w 128 is the key for extra performance.

ozzy1926

sr. member

Activity: 378

Merit: 250

although watt meter says all cards working after 7-8 hurs miner freeze and pool hashrate is 0anybody having the same issue?

JPS2K5

full member

Activity: 266

Merit: 100

Oddness.... I had the b2 running for over 24 hours. I decided to fiddle with some settings. All I got was HW errors, reversed back to old settings, still HW errors. Deleted .bin files, still HW errors.
Went for latest version, same thing.....

Now all that seems to be running 'normally' is the Original release with I=15 Huh

skyyyy

member

Activity: 85

Merit: 10

thanks,dev
hi dev, could you please add aidencoin mod, it is minable on summer

Starlightbreaker

legendary

Activity: 1764

Merit: 1006

does "vectors" setting matter?

on old version of cgminer, memory speed & vectors settings affect the hashrate, but doesn't seem like there's any significant improvement from 1, 2, or 4.

i gained only ~0.075 mh/s changing it from 1 to 4.

Topic: [ANN][X11/X13] X11 (Darkcoin)/X13 (Marucoin) miner (based on sph-sgminer) - page 49. (Read 351531 times)