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

[Thirtybird]

mining ( and building the farm ) is by far the most intensive build i have done in a very long time - and that includes rolling out server and desktop systems for american express those many years ago ...

stable - not quite ...

though i do agree with you that if its only doing something as 'simple' ( and i dont mean that any of it simple - but compared to the stresses that the equipment gets put under in mining ) like data services or staorage and iis - most os's can do that ...

as far as rock solid stability goes with mining - windows is a flake ...

tried it - done it - dumped it ... next! ... linux ... and the farm is still rock solid ...

not as oc'd as the windows based environments are - but that will change as time progresses ...

the issue is support ... when linux gets the support that it deserves - then there wont be much else in its way ...

#crysx

Mining exclusively on Windows since 2012 - maybe I've just been lucky for 3+ years... Or, it may just be that supporting MS products has paid the bills for almost 20 years

[1714753786]

1714753786

[1714753786]

1714753786

[chrysophylax]

mining ( and building the farm ) is by far the most intensive build i have done in a very long time - and that includes rolling out server and desktop systems for american express those many years ago ...

stable - not quite ...

though i do agree with you that if its only doing something as 'simple' ( and i dont mean that any of it simple - but compared to the stresses that the equipment gets put under in mining ) like data services or staorage and iis - most os's can do that ...

as far as rock solid stability goes with mining - windows is a flake ...

tried it - done it - dumped it ... next! ... linux ... and the farm is still rock solid ...

not as oc'd as the windows based environments are - but that will change as time progresses ...

the issue is support ... when linux gets the support that it deserves - then there wont be much else in its way ...

#crysx

Mining exclusively on Windows since 2012 - maybe I've just been lucky for 3+ years... Or, it may just be that supporting MS products has paid the bills for almost 20 years

hehehe ... true - true ...

i have to admit - microsoft have helped put a lot of money in mine over the last 27 years also ...

ahhh support ... i really dont miss it ... that much ...

#crysx

[chrysophylax]

Hoping they release soon Quark miner optimized for 290x...

maybe if we all pitch in - wolf 'may' have a nice optimized one if he is interested ...

#crysx

[sp_]

I think it's time to make a sgminer (sp-mod). Open source.
For the last 9 months I have been sucessfully increasing the hashrate in all hashing algorithms for the NVIDIA GPU's. Time to move on to the AMD's.

Wolf0 a private quark kernel that does 18MHASH@quark on the 290x, and the opensource version is only doing 2.5MHASH.

I think I can push the opensource quark to 10MHASH pretty easy.

Who wants to donate some beers?

[Rux]

I think it's time to make a sgminer (sp-mod). Open source.
For the last 9 months I have been sucessfully increasing the hashrate in all hashing algorithms for the NVIDIA GPU's. Time to move on to the AMD's.

Wolf0 a private quark kernel that does 18MHASH@quark on the 290x, and the opensource version is only doing 2.5MHASH.

I think I can push the opensource quark to 10MHASH pretty easy.

Who wants to donate some beers?

atm AMD @ quark

270   = 6 MH/S
280x = 13 MH/S
290   = 16 MH/S

and for 290x, well i don't own it

[sp_]

I think it's time to make a sgminer (sp-mod). Open source.
For the last 9 months I have been sucessfully increasing the hashrate in all hashing algorithms for the NVIDIA GPU's. Time to move on to the AMD's.
Wolf0 a private quark kernel that does 18MHASH@quark on the 290x, and the opensource version is only doing 2.5MHASH.
I think I can push the opensource quark to 10MHASH pretty easy.
Who wants to donate some beers?

atm AMD @ quark
270   = 6 MH/S
280x = 13 MH/S
290   = 16 MH/S
and for 290x, well i don't own it

I know there are faster kernals, but they are not open source.

[Eliovp]

I think it's time to make a sgminer (sp-mod). Open source.
For the last 9 months I have been sucessfully increasing the hashrate in all hashing algorithms for the NVIDIA GPU's. Time to move on to the AMD's.

Wolf0 a private quark kernel that does 18MHASH@quark on the 290x, and the opensource version is only doing 2.5MHASH.

I think I can push the opensource quark to 10MHASH pretty easy.

Who wants to donate some beers?

A nice initiatif from you sp_ but the "optimized Windows" version is also in a way opensource and faster then 10Mhash. Everyone is using it. A few, die hard *nix fans aren't.

Anyhow, nice of you to want to get involved.

[sp_]

The free quark miner (windows executable) is most likely a Trojan virus and have hidden mining inside it. Opensource is the way to go..
Secures the blockchains, secure the coins.

[pallas]

The free quark miner (windows executable) is most likely a Trojan virus and have hidden mining inside it. Opensource is the way to go..
Secures the blockchains, secure the coins.

True, but I fear people won't support an opensource miner which is slower than a closed source, readily available one.
Personally I'd prefer if Wolf0 would release his miner (the old one), so you don't have to do the same work over again. But of course it's up to him ;-)
I myself made part of the work for x11 and it takes a lot of time indeed.

[sp_]

I know. The nvidia boosts are small but they are many. Some of the tricks I have used can also be done in the opencl code.. F.eks remove a couple of rounds of echo and shavite and precalc them..
The point is that I have already done it in cuda.. Rewriting to opencl will take time, but I am not starting from scratch.. I might have some changes that wolf0 hasn't thought of. The final miner could be faster than the leaked bins....

[chrysophylax]

I know. The nvidia boosts are small but they are many. Some of the tricks I have used can also be done in the opencl code.. F.eks remove a couple of rounds of echo and shavite and precalc them..
The point is that I have already done it in cuda.. Rewriting to opencl will take time, but I am not starting from scratch.. I might have some changes that wolf0 hasn't thought of. The final miner could be faster than the leaked bins....

i would like to see that - especially in x11 ... and would definitely back you on that sp ...

nvidia and amd ... fully optimized AND fully open-sourced ...

now that would make for a nice mix - and would definitely be worth more than a couple of beers ...

especially if you could get it all into one miner - one compile ... x11 - quark - spreadx11 - and so forth ...

#crysx

[sp_]

The Russian(?) virus kernals are too slow...

Quark:

Gigabyte winddforce gtx 970 standard clocks (running opensource sp-mod version 51): 16.2MHASH (1664 shaders) 140WATT
Overclocked radeon 290 (2560 shaders) 16MHASH 300watt++

[Oscilson]

The Russian(?) virus kernals are too slow...

Quark:

Gigabyte winddforce gtx 970 standard clocks (running opensource sp-mod version 51): 16.2MHASH (1664 shaders) 140WATT
Overclocked radeon 290 (2560 shaders) 16MHASH 300watt++

Looking forward to your more efficient miners.

[anatolikostis]

The Russian(?) virus kernals are too slow...

Quark:

Gigabyte winddforce gtx 970 standard clocks (running opensource sp-mod version 51): 16.2MHASH (1664 shaders) 140WATT
Overclocked radeon 290 (2560 shaders) 16MHASH 300watt++

It`s injected with @double-shares "option" because of the same creator - kachur aka kpnt24 - we have faced with monero @double-shares "option" miner.

[jimlite]

The Russian(?) virus kernals are too slow...

Quark:

Gigabyte winddforce gtx 970 standard clocks (running opensource sp-mod version 51): 16.2MHASH (1664 shaders) 140WATT
Overclocked radeon 290 (2560 shaders) 16MHASH 300watt++

It`s injected with @double-shares "option" because of the same creator - kachur aka kpnt24 - we have faced with monero @double-shares "option" miner.

I wouldn't say that without any proof. I have been running it for a week and my miner and work at pools all show equivalent work. Sure it is possible that
it could be skimming an insignificant amount that I  can't detect, but without proof you are just making accusations.

[chrysophylax]

The Russian(?) virus kernals are too slow...

Quark:

Gigabyte winddforce gtx 970 standard clocks (running opensource sp-mod version 51): 16.2MHASH (1664 shaders) 140WATT
Overclocked radeon 290 (2560 shaders) 16MHASH 300watt++

It`s injected with @double-shares "option" because of the same creator - kachur aka kpnt24 - we have faced with monero @double-shares "option" miner.

I wouldn't say that without any proof. I have been running it for a week and my miner and work at pools all show equivalent work. Sure it is possible that
it could be skimming an insignificant amount that I  can't detect, but without proof you are just making accusations.

here is the difference ...

we know nothing really about the full workings of the miner due to it being closed ...

that being said - we dont really know a great deal about the developer himself ...

so all this speculation about whether it skims or it doesnt really is irrelevant - especially if you are running the miner without ANY prior research into the miner AND the dev AND dont take precautions to protect your systems from any malicious code as well ( the 'just in case' factor ) ...

the difference is simple in this respect ...

devs like sp - djm34 - wolf0 ( amongst a few others ) are blunt and open ... they are respected community members and when they develop - they develop with no ambiguity ... there are no questions whether the miner or optimizations they develop are questionable in any way - because they TELL the community if it is open-source or closed-source AND what the miner does and doesnt do ... whether they sell their work or not - there are no questions ... unlike what we are going through at the moment with this miner ...

this brings about a level of TRUST that you dont get with these other devs that release 'their' apps and bring nothing else really to the table ...

this is why we can afford to wait and see and NOT go with the miners we have doubts and unanswered questions ... and there is no way we ( ACE ) would risk our farm ( linux or not ) with a miner that is released from a developer that we cant trust - due to his lack of transparency ...

if YOU want to run it - its at your own risk ... either way its YOUR decision ...

simple ...

#crysx

[arvita112]

how much should speed 290 in quark?

[lbr]

how much should speed 290 in quark?

12.4@900/1250

[jtan1ph]

What sgminer version did you use?

Thanks.

[Oscilson]

That guy has very high speed, one rig has over 600MH/s, how could that be? FPGA?

https://www.westhash.com/index.jsp?p=miners&a=12&addr=1MPnrpSFBEvzDqpFpv2Fit4ktRVzkBVfdQ