hsrminer - Nvidia mining software for various algos by palgin&alexkap

[kostas554]

Looks interesting, will give it a try.  What other algos are you targeting for addition?

[Monberg75]

Hi, got a little problem here..

My HSR miner is stopping once a while with this message.

--
Cuda error in func 'neoscrypt_cpu_hash_1060' at line 1420 : unspecified launch failure.
Cuda error in func 'neoscrypt_cpu_hash_1080' at line 1509 : unspecified launch failure.
--

http://img.techpowerup.org/180125/p1kalmig2k256.jpg

Any clue on what i should do?

Best regards:
Monberg75

[NetopyrMan]

Hi, got a little problem here..

My HSR miner is stopping once a while with this message.

--
Cuda error in func 'neoscrypt_cpu_hash_1060' at line 1420 : unspecified launch failure.
Cuda error in func 'neoscrypt_cpu_hash_1080' at line 1509 : unspecified launch failure.
--

http://img.techpowerup.org/180125/p1kalmig2k256.jpg

Any clue on what i should do?

Best regards:
Monberg75

:loop
hsrminer command
goto loop

[Monberg75]

How does that solve the problem?

[MagicSmoker]

How does that solve the problem?

It automatically restarts the miner.

Your problem is likely from overclocking too much. This miner is very finicky about that. Try backing off core clock or increasing core voltage *if* it has been lowered.

[maggari]

Miner working fine, though according to pool data (actual hashrate), seems to be even lower than KlausT or similar minus donations. No matter what those pretty numbers tell in miners. 

Well, about 36hours and I have screenshots to give.


Might also be my R9 Fury rig having it's periods and ruining the results, it's mining into same wallet, should restrict it into separate one.

[Elder III]

If anyone has problems with 100% rejected shares on a GTX 1060, try running the test version of the miner that is on the first page of this thread. We had 2 rigs that would not work (100% rejected shares) until we tried out the test version.  Those 2 rigs have been working fine for going on 2 full days now using the test miner.

[Monberg75]

How does that solve the problem?

It automatically restarts the miner.

Your problem is likely from overclocking too much. This miner is very finicky about that. Try backing off core clock or increasing core voltage *if* it has been lowered.

I'm not running any overclock. Standard all the way.

Already got a script to restart the program, but in the end it all locks up.
I'll try to run an instance for every card to se if I can solve the problem. Maybe it's because I'm using 3 different cards in same miner. Then I'll try to separate every type to a specific miner.

[stevascha]

Hi, got a little problem here..

My HSR miner is stopping once a while with this message.

--
Cuda error in func 'neoscrypt_cpu_hash_1060' at line 1420 : unspecified launch failure.
Cuda error in func 'neoscrypt_cpu_hash_1080' at line 1509 : unspecified launch failure.
--

http://img.techpowerup.org/180125/p1kalmig2k256.jpg

Any clue on what i should do?

Best regards:
Monberg75

yeah me too, but when use another ccmin (ex klaust) it don't show ur message

[Monberg75]

Hi, got a little problem here..

My HSR miner is stopping once a while with this message.

--
Cuda error in func 'neoscrypt_cpu_hash_1060' at line 1420 : unspecified launch failure.
Cuda error in func 'neoscrypt_cpu_hash_1080' at line 1509 : unspecified launch failure.
--

http://img.techpowerup.org/180125/p1kalmig2k256.jpg

Any clue on what i should do?

Best regards:
Monberg75

yeah me too, but when use another ccmin (ex klaust) it don't show ur message

I think I'll be switching to Nemos instead.. I have to try something else, at least til i solve this problem with HSR.

[sugi88999]

my HSR miner stop after a few hour, and it happened several time. i use it on windows 10. how to solve the problem tq

[al1n]

my HSR miner stop after a few hour, and it happened several time. i use it on windows 10. how to solve the problem tq

I have the same issue on win 10 too
I solved it by automatically closing and restarting the process every hour

[Dropsh0t]

When Linux?

Hy кoгдa, нaкoнeц, coбepeтe пoд линyкc?

source on githab, find the one who does under Linux, will you be a miner
иcxoдники нa гитxaбe, нaйди тoгo ктo дeлaeт пoд линyкc, бyдeт тeбe мaйнep

There's no sources. Release only. It's closed code miner with fee.

[jugger1028]

For you people that keep crashing, lower your overclock settings, HSRMiner is not as stable as CCMiner when it comes to overclocking so lower it. The cuda errors are overclock errors. Also, make sure you have AT LEAST the same amount of virtual memory set up as you have memory on the card.. 5 1080ti cards = 11*5 = 55gb of virtual memory minimum. You shouldn't need to set it to restart every hour, this is stable IF you don't push the limits.

[Monberg75]

For you people that keep crashing, lower your overclock settings, HSRMiner is not as stable as CCMiner when it comes to overclocking so lower it. The cuda errors are overclock errors. Also, make sure you have AT LEAST the same amount of virtual memory set up as you have memory on the card.. 5 1080ti cards = 11*5 = 55gb of virtual memory minimum. You shouldn't need to set it to restart every hour, this is stable IF you don't push the limits.

How do i set the virtual memory.

I don't have any overclock, just standard settings, so the virtual memory might be my problem.

[MagicSmoker]

For you people that keep crashing, lower your overclock settings, HSRMiner is not as stable as CCMiner when it comes to overclocking so lower it. The cuda errors are overclock errors. Also, make sure you have AT LEAST the same amount of virtual memory set up as you have memory on the card.. 5 1080ti cards = 11*5 = 55gb of virtual memory minimum. You shouldn't need to set it to restart every hour, this is stable IF you don't push the limits.

How do i set the virtual memory.

I don't have any overclock, just standard settings, so the virtual memory might be my problem.

Then you may have to *underclock* your card for it to work with hsrminer; I did with one GTX 1080 I have that comes with a factory overclock (more specifically, I had to lower the core clock by setting -125 in MSI AB).

Virtual memory is almost certainly not the issue unless you have the minimum amount of system RAM (e.g. - 4GB for Windows 10) and multiple GPUs.

[jugger1028]

lol, virtually memory is most definitely a very big factor. Mining rigs usually have 4-8gb of ram and if you have even 1 or 2 cards with low virtual memory you will most definitely get crashes and errors. I run 8 1080ti's, I need 88gb of virtual memory minimum. My system hasn't crashed in weeks and my oc settings are 90%tdp, +100 +600 (some cores run +150 on my hybrids). No crashes.

Go to system, advanced settings (on left side) then advanced tab, then set it there.. Or just google changing virtual memory to get a more thorough guide.

[MagicSmoker]

lol, virtually memory is most definitely a very big factor....

Yeah, well, I did say that virtual memory isn't likely an issue unless minimum system memory and/or several GPUs are present.

Regardless, I've only seen the Cuda error from hsrminer when I try to manually run too high an intensity; I either have way more system memory than is necessary for both Windows and the GPU, or else a fixed pagefile with a minimum size of the sum of all the GPU VRAM, so running out of memory, virtual or physical, isn't an issue.

Now could those Cuda errors be caused by insufficient physical + virtual memory? I dunno, I'm not a programmer, but it certainly seems possible; it's just that I have seen similar errors and insufficient memory wasn't the cause.

[dragonmike]

For you people that keep crashing, lower your overclock settings, HSRMiner is not as stable as CCMiner when it comes to overclocking so lower it. The cuda errors are overclock errors. Also, make sure you have AT LEAST the same amount of virtual memory set up as you have memory on the card.. 5 1080ti cards = 11*5 = 55gb of virtual memory minimum. You shouldn't need to set it to restart every hour, this is stable IF you don't push the limits.

How do i set the virtual memory.

I don't have any overclock, just standard settings, so the virtual memory might be my problem.

Try this, it's guaranteed to work. Every time: http://bfy.tw/GGF3

And please do let us know how you get on! Your feedback is important to us!

[MarcusDe]

Your mem clock is extremely low for this type of memory. Set it to 5400 at least, also delete the intesity setting. HSR miner will do the job for you and set it better. Also if you have such low temp, maybe it is not good idea to lower power limit especially for neoscrypt.

Letting hsrminer set intensity has resulted in a mysterious drop in hashrate by 100-150 kH/s after 1-3 hours of operation, and this is with two separate GTX 1080 on two very different machines. But in the spirit of science I'll try your suggestions. I noticed you don't mention anything about core clock, though, and that seems to be the source of woes for me.

Right now my Asus GTX 1080 ROG Strix edition is hashing at 1208 kH/s while drawing 166W at 85% PL, -125 core (varying between 1595 to 1670) and +300 mem (4811). (while typing this the hashrate dropped to 1176 kH/s... sigh).

My older EVGA GTX 1080 FE is hashing at 1218 kH/s while drawing 146W at 85% PL, +125 core (1633 to 1658) and +300 mem (4811 again).

In both cases setting -i 4 resulted in about 100 kH/s lower hashrate but with no real change in stability, at least for the FE card; I didn't bother repeating the same experiment with the Asus.

Hi,

Funny that lowering core is showing those results -- confirmed on my ASUS 1080. ccminer gives 990 on same settings.

Any tips for 1060?