[ANN] dstm's ZCash / Equihash Nvidia Miner v0.6.2 (Linux / Windows)

[remix4life]

Sound like a driver crash, odd that the miner doesn't crash either, you could try to kill all binaries, and do a nvidia drivers reset, then restart the miner, may work, your solution work too are you using latest version with the --noconnect ?

Me i noticed another odd thing, if you use the miner multiple times throught the day, the hashrate lower gradually like 1% each time it's started, if you reboot the rig it's back up to the normal hashrate, i never noticed this with EWBF, it's kinda odd.

Yep, latest zm with --noconnect, latest nvidia and cuda driver.

For now ewbf's still a win for me, despite a little slower than zm, but I don't have to worry either pools going down, or have to restart the miner every X hours. I like zm with active dev and hope with constant update, the stability and features is up there or better then with ewbf.
I'm, still requesting pool failover, by the way 

I'm getting 390 vs 375 with my 1070 (60%, +175 core, +200 Mem)
But with dev fee the added 15 sol/s turns into a net +7.5 sol
Would be nice to have the dev fee to drop to 1%
Great work

I've got 478-489 sol/s Gigabyte Aorus GTX 1070 with micron memory, 70% power, +100 core, +700 memory

Has anyone else noticed that when you drop the power draw on your cards, it doesn't adjust the Sol/W readings in the miner? I took mine from 85% down to 67% and after a few minutes they were still showing 3.30 Sol/W.

Seems like driver bug? I noticed after add more card, Sol/W reading shows "infinity" in my rig. I have to reinstall nvidia and cuda driver then the Sol/W reading back to normal.

[1714118257]

1714118257

[1714118257]

1714118257

[1714118257]

1714118257

[induktor]

after long tests and usage my opinion about dstm for GTX1080ti is here - it is the same as ewbf
why? here is why :
hashrate for one gtx 1080ti aprox  770sol, so 4 x 760 = 3040 sol/s
while with same settings it is 2960 sol on ewbf.
so you it is 2.7% win for dstm.  but, there is a dev fee 2% and ewbf we can use without devfee - that means that actual difference is less then 1%
and taking into account, that ewbf is more stable for my rig, the choice is obvious

however! for 2 x gtx1070 on dstm i'm getting 960 sol/s and 920 sol/s with ewbf 
so it is 4.3% difference, minus 2% dev fee and we comes to 2.3% benefit.

I can accept that bonus for gtx1070

I second that, I see exactly the same production difference.

Only my numbers are different because you are pushing the cards hard in power limit, and I am aiming at top efficiency, but regardless the result is the same, my numbers says 2.2 % difference bettween EWBF and DSTM miner, so taking into the account the 2% fee the production is the same.

at least on the 1080Ti's  i am testing on the 1080s (non-Tis) I'll post back tomorrow with results, but it seems that production is the same, i'm also don't notice much CPU usage right now, it looks lower than previous versions.

BTW: zm miner tends to spike power usage (it's not constant, it goes up and down) the average endup being the same as EWBF but EWBF miner uses constant power, no spikes of any kind (i have an AC current meter bar graph (vu-meter like) to measure this behavior)

so in the end no much difference, but as someone said here dstm is active developing it, so I'm more inclined towards using ZM miner for the time being.

cheers
Indkt.

[Makak4R]

Has anyone else noticed that when you drop the power draw on your cards, it doesn't adjust the Sol/W readings in the miner? I took mine from 85% down to 67% and after a few minutes they were still showing 3.30 Sol/W.

I guess your driver had an error or glitch! that's why dstm kinda frozen?
maybe that's happening because of unstable overclocking settings
try to lower it a little and check how it will be going.

[dstm]

Okey, i updated to version 1080ti:

When start it gota 750-760sols
after few hours it gota stable on 740~ sols its okey. But what is mean by contime? If i lose connection can it be due high oc settings?

contime denotes the time passed since last connect/reconnect.
ZM reconnects if the OS signals a closed connection to it.

[dstm]

Bceм пpивeт. Ктo мoжeт пoдcкaзaть пoчeмy нe нaчиcляютcя zec зa мaйнинг? Oнo зaчиcляeтcя oт кaкoй тo oпpeдeлeннoй cyммы или пo вpeмeни? Cтoялo нoчь, peзyльтaт пo 0. Bидeoкapтa gtx 1050 2gb. B paзгoнe выдaeт 164 sol. Кoшeлeк Zec пpoпиcaл в start.bat

Знaчит нe пpaвильнo бaтник cocтaвил.

I have this error

2017-11-16 11:54:22|gpu_id 1 2 000000052AB60400 unknown error
2017-11-16 11:54:22|gpu_id 3 2 000000055F600000 unknown error
2017-11-16 11:54:22|gpu_id 4 0 0000000579C00000 unknown error
2017-11-16 11:54:22|gpu_id 0 0 000000050DEB0200 unknown error
2017-11-16 11:54:22|gpu_id 2 6 00000005450420C0 unknown error
2017-11-16 11:54:22|gpu 2 unresponsive - check overclocking
2017-11-16 11:54:22|cudaMemcpy 1 failed
2017-11-16 11:54:22|gpu 4 unresponsive - check overclocking
2017-11-16 11:54:22|cudaMemcpy 1 failed
2017-11-16 11:54:22|gpu 1 unresponsive - check overclocking
2017-11-16 11:54:22|cudaMemcpy 1 failed
2017-11-16 11:54:22|gpu 0 unresponsive - check overclocking
2017-11-16 11:54:22|gpu 3 unresponsive - check overclocking
2017-11-16 11:54:22|cudaMemcpy 1 failed
2017-11-16 11:54:22|cudaMemcpy 1 failed

Hello, i got 2x10rigs (all 1070), and your miner work awesome on the first 10 rigs (+7% hash vs EWBF), The other 10 rigs have a different GPU brand and same hashpower (all 1070) but i can't get your miner to work, i experience the exact same error as bruce81, even with base clock/power on cards, i also get "memory errors" log line, but thoses rigs have worked overclocked flawlessly for months on EWBF (and still does) (All 20rigs have the same gpu, drivers, OS, SSD, ram, and so on, only the brand of the GPU change (10x rigs with EVGA's 1070, 10x rigs with KFA2 1070's)

An idea what it could be ? Im willing to do some testing or advanced logging.

I was wondering if you could also add something to tweak the intensity? Maybe i need to lower it a bit so that the miner doesn't explode, dunno

Regards.

This kind of errors usually happen if the GPU crashes and fails to transfer data over the pcie bus.
Does it happen right after starting zm or randomly after some time?
Are there really no other differences like CPU, mainboard, power supply etc.?

[dstm]

You can probably get something similar with setting up a batch script that after it exit the miner because of disconnect (use the --noreconnect miner switch) it restart the miner on a different pool. Have to be sure the miner exit on pool error. But should work

Hey that's a good idea, but I don't think I can do that in my rig,

I've got a strange experience using zm after running 15 hours give or take, I received an e-mail from flypool, said that my miner was offline. Then I checked my rig, Internet connection ok, zm still running and mining but not sending any shares to the pool (only reporting my hashrate, temp, etc) therefore the pool sent an e-mail.

I tried to close the miner and restart it again, everything going back to normal. Now I have to set my batch file to auto restart zm every 12 hours, been running good for about a week. I tried to make a loop like you suggested but it didn't work, the miner didn't auto restarted like it used to be, it just opened new cmd window with the new backup pool address but not mining because the gpus somewhat still minning in the first cmd. 

It might take some time till the OS signals a closed connection to ZM.

[dstm]

Has anyone else noticed that when you drop the power draw on your cards, it doesn't adjust the Sol/W readings in the miner? I took mine from 85% down to 67% and after a few minutes they were still showing 3.30 Sol/W.

https://bitcointalk.org/index.php?topic=2021765.msg24542311#msg24542311

[bruce81]

Bceм пpивeт. Ктo мoжeт пoдcкaзaть пoчeмy нe нaчиcляютcя zec зa мaйнинг? Oнo зaчиcляeтcя oт кaкoй тo oпpeдeлeннoй cyммы или пo вpeмeни? Cтoялo нoчь, peзyльтaт пo 0. Bидeoкapтa gtx 1050 2gb. B paзгoнe выдaeт 164 sol. Кoшeлeк Zec пpoпиcaл в start.bat

Знaчит нe пpaвильнo бaтник cocтaвил.

I have this error

2017-11-16 11:54:22|gpu_id 1 2 000000052AB60400 unknown error
2017-11-16 11:54:22|gpu_id 3 2 000000055F600000 unknown error
2017-11-16 11:54:22|gpu_id 4 0 0000000579C00000 unknown error
2017-11-16 11:54:22|gpu_id 0 0 000000050DEB0200 unknown error
2017-11-16 11:54:22|gpu_id 2 6 00000005450420C0 unknown error
2017-11-16 11:54:22|gpu 2 unresponsive - check overclocking
2017-11-16 11:54:22|cudaMemcpy 1 failed
2017-11-16 11:54:22|gpu 4 unresponsive - check overclocking
2017-11-16 11:54:22|cudaMemcpy 1 failed
2017-11-16 11:54:22|gpu 1 unresponsive - check overclocking
2017-11-16 11:54:22|cudaMemcpy 1 failed
2017-11-16 11:54:22|gpu 0 unresponsive - check overclocking
2017-11-16 11:54:22|gpu 3 unresponsive - check overclocking
2017-11-16 11:54:22|cudaMemcpy 1 failed
2017-11-16 11:54:22|cudaMemcpy 1 failed

Hello, i got 2x10rigs (all 1070), and your miner work awesome on the first 10 rigs (+7% hash vs EWBF), The other 10 rigs have a different GPU brand and same hashpower (all 1070) but i can't get your miner to work, i experience the exact same error as bruce81, even with base clock/power on cards, i also get "memory errors" log line, but thoses rigs have worked overclocked flawlessly for months on EWBF (and still does) (All 20rigs have the same gpu, drivers, OS, SSD, ram, and so on, only the brand of the GPU change (10x rigs with EVGA's 1070, 10x rigs with KFA2 1070's)

An idea what it could be ? Im willing to do some testing or advanced logging.

I was wondering if you could also add something to tweak the intensity? Maybe i need to lower it a bit so that the miner doesn't explode, dunno

Regards.

This kind of errors usually happen if the GPU crashes and fails to transfer data over the pcie bus.
Does it happen right after starting zm or randomly after some time?
Are there really no other differences like CPU, mainboard, power supply etc.?

No, its randomly. After restart rig, one day works. Let's see more

[dstm]

Hello dstm, may I suggest add failover pools feature like in ewbf and cgminer, thanks.

This was requested several times, will add.

[santan]

maybe after every 20th === line, add 2 summary lines describing power and max diff share submitted.....perhaps within
================================================
Power:    GPU#0(147W), GPU#1(151W), GPU#2(145W), GPU#3(150W)
Max Diff: GPU#0(11.73K), GPU#1(8.82K), GPU#2(2.01M), GPU#3(78.46K)
================================================
Or whatever else you want to...........

[coinfoundry]

Hello dstm, may I suggest add failover pools feature like in ewbf and cgminer, thanks.

This was requested several times, will add.

That's good news. Thanks for listening!

[Gans888]

Hello dstm, may I suggest add failover pools feature like in ewbf and cgminer, thanks.

This was requested several times, will add.

Hi! Do you not increase miner speed in version 0.5.5 compared 0.5.3?

[m1n1ngP4d4w4n]

This kind of errors usually happen if the GPU crashes and fails to transfer data over the pcie bus.
Does it happen right after starting zm or randomly after some time?
Are there really no other differences like CPU, mainboard, power supply etc.?

Nope no difference, even using the same PSU's, it happens very fast after the start of the miner, but i can run other miners or EWBF for days without issues

[PickAxe131]

I have this error

2017-11-16 11:54:22|gpu_id 1 2 000000052AB60400 unknown error
2017-11-16 11:54:22|gpu_id 3 2 000000055F600000 unknown error
2017-11-16 11:54:22|gpu_id 4 0 0000000579C00000 unknown error
2017-11-16 11:54:22|gpu_id 0 0 000000050DEB0200 unknown error
2017-11-16 11:54:22|gpu_id 2 6 00000005450420C0 unknown error
2017-11-16 11:54:22|gpu 2 unresponsive - check overclocking
2017-11-16 11:54:22|cudaMemcpy 1 failed
2017-11-16 11:54:22|gpu 4 unresponsive - check overclocking
2017-11-16 11:54:22|cudaMemcpy 1 failed
2017-11-16 11:54:22|gpu 1 unresponsive - check overclocking
2017-11-16 11:54:22|cudaMemcpy 1 failed
2017-11-16 11:54:22|gpu 0 unresponsive - check overclocking
2017-11-16 11:54:22|gpu 3 unresponsive - check overclocking
2017-11-16 11:54:22|cudaMemcpy 1 failed
2017-11-16 11:54:22|cudaMemcpy 1 failed

I get the exactly same error with 0.5.5. All 4 rigs crashes between 1minute to 30minutes. Switched back to 0.5.3 and it works flawless again. I use 1080ti's

Edit: Its on Windows. Tested on my main computer, crashes alot on gtx 1070 too.

[m1n1ngP4d4w4n]

I get the exactly same error with 0.5.5. All 4 rigs crashes between 1minute to 30minutes. Switched back to 0.5.3 and it works flawless again. I use 1080ti's

Edit: Its on Windows. Tested on my main computer, crashes alot on gtx 1070 too.

I never tried 0.5.3, will try on those rigs that have troubles thx for the tip

Edit : nope, tried on 3 rigs, even ZM 0.5.3 crash after a few seconds.

[Vispilio]

this miner is almost there, but still needs hash rate improvements with every upcoming version, only then can it finally surpass the EWBF miner which has stopped updating...

When can we expect a hash rate increase in the upcoming versions ?..

[slackert]

There was a windows update that stops the miner from running. Back to EWBF until this is resolved.

[Vispilio]

yeah also many other trivial errors like: "NVIDIA can't initialize, gpu is lost" that stops the miner from running,

maybe for all these minor issues you can add a functionality like "if .exe is running but nothing is being mined for x minutes, relaunch .exe file"

[crocozino]

this miner is almost there, but still needs hash rate improvements with every upcoming version, only then can it finally surpass the EWBF miner which has stopped updating...

When can we expect a hash rate increase in the upcoming versions ?..

agree, we need really some decent optimizations!
especially for 1080 and 1080ti, they way underpowered now..
possible some important parts of miner's code has to be rewritten in asm?

[randywald]

I like this miner.

Cards: 4x EVGA 1080 Ti FTW3
Efficiency: ~4.6 Sols/W

If you run this at 2.8 sols/W you must have very cheap electricity costs.