CCminer(SP-MOD) Modded GPU kernels.

[zTheWolfz]

Folding@EVGA http://forums.evga.com/tt.aspx?forumid=28 will get you more money quicker then any algo, $10. 00 for 2 days work for me with my hardware, takes 5 to 6 days with the same hardware to make $10 mining the top paying algo. ;-) I do folding at home for 2 days each month for the 10 EVGA Buck to go towards a new video card. Those EVGA buck bought me my new GTX960 GTX760 and 2x GTX660Ti and other video cards over the years. Sure you have to save them all year to really have anything since its a once a month thing and 10 buck max, but it still makes a nice Christmas gift for yourself.

http://folding.extremeoverclocking.com/user_summary.php?s=&u=610174

[sp_]

ok just tried release 78: works like crap (107kh/s the 750ti OC, other cards just not working) 

hehe. Cuda 7.5

[zTheWolfz]

ok just tried release 78: works like crap (107kh/s the 750ti OC, other cards just not working)  

hehe. Cuda 7.5

78 worked with my 960 but was slower, but I'm still on cuda 6.5 and no plans to upgrade anytime soon since 74 is still really quicker for me.

Edit: SP when do I get some of those private miners? I donated to you somewhere between .02 to .03 over time and about the same to many of the other Dev's.
I'm only making about half that now with more hardware & improved miners per day is only .005   was .01+ when I sent out those donations.

[tbearhere]

But You know very well that the reason why I complain, is because you claimed you improved the speed of lyra and when I checked out your release (was expecting actually to use some of your "faster" kernels) I

Look at the chart above. The first release with the lyra2v2 is your kernal. And then the speed is improved.


I have optimized your kernal on the gtx 950,960 and 970. The 980 is not a card for the farms because it is to expensive

The 980ti is same price or lower then 4 750ti and much more hashrate if some would implement it.
And get the mem clock up.

[chrysophylax]

just tested release 74 on my 980ti g1 and 980 reference standard clocks

sp ...

in the next few days ill test the big rig - even if the 6th card wont run ...

im still having massive issues - no matter what i do with the setup ... bios / drivers / yet to try windows ( and very reluctant - it means i have to support my mate EVERYTIME he has an issue with it - where linux is plug and run ) ...

so ill compile the latest - and run on test using the gigabyte 980ti g1 - and then the 980ti extreme ...

apologies i havent done this yet - but i have a lot of personal things i need to sort ...

#crysx

[chrysophylax]

Reading these convo's is almost as good as watching a movie.

It's hard for me sometimes to decide what to choose..

I would rather say a very boring soap opera...  

ok just tried release 78: works like crap (107kh/s the 750ti OC, other cards just not working) 
(how am I surprise... so many version half broken... )
didn't pay attention it was 74 which should be used (lets try again...)

btw watching a very good serie: Galavant

is that testing neoscrypt on that djm34? ...

im guessing it is with the 'kh' rate ...

#crysx

[chrysophylax]

But You know very well that the reason why I complain, is because you claimed you improved the speed of lyra and when I checked out your release (was expecting actually to use some of your "faster" kernels) I

Look at the chart above. The first release with the lyra2v2 is your kernal. And then the speed is improved.


I have optimized your kernal on the gtx 950,960 and 970. The 980 is not a card for the farms because it is to expensive

The 980ti is same price or lower then 4 750ti and much more hashrate if some would implement it.
And get the mem clock up.

agreed ...

in benchmark - some algos show that its the hashrate of 5 x gigabyte 750ti oc lp cards ...

so if the optimization are focused on the 980ti ( which is the predecessor of pascal - which ever way you look at it ) then it would make it a better option to mine with ...

but those cards are STILL very expensive ...

this machine im building for my mate in total cost about $7900aud so far ( without a windows license - regardless of windows 10 x64 ) ... BUT the hashrate if ALL 6 cards were running is about 50% - 75% of the gigabyte 750ti oc lp farm im running now - depending on which algo ...

i cant wait till im settled in the next two weeks - so i can focus on all this again ...

#crysx

[joblo]

But You know very well that the reason why I complain, is because you claimed you improved the speed of lyra and when I checked out your release (was expecting actually to use some of your "faster" kernels) I

Look at the chart above. The first release with the lyra2v2 is your kernal. And then the speed is improved.


I have optimized your kernal on the gtx 950,960 and 970. The 980 is not a card for the farms because it is to expensive

The 980ti is same price or lower then 4 750ti and much more hashrate if some would implement it.
And get the mem clock up.

agreed ...

in benchmark - some algos show that its the hashrate of 5 x gigabyte 750ti oc lp cards ...

so if the optimization are focused on the 980ti ( which is the predecessor of pascal - which ever way you look at it ) then it would make it a better option to mine with ...

but those cards are STILL very expensive ...

this machine im building for my mate in total cost about $7900aud so far ( without a windows license - regardless of windows 10 x64 ) ... BUT the hashrate if ALL 6 cards were running is about 50% - 75% of the gigabyte 750ti oc lp farm im running now - depending on which algo ...

i cant wait till im settled in the next two weeks - so i can focus on all this again ...

#crysx

In Canada the price of a 980ti is about 5x 750ti, needs to drop some.

Did you look back in the thread for the old 6th card problem discussion?

[bensam1231]

Djm34 is always complaining when somebody else is optimizing his opensource code. Me and Pallas optimized the neoscrypt kernal 20% faster, but it slower than the djm34 private kernal. I look forward to see the new code.

It is true that I can be a little protective over my work   (may-be a bit jealous too yeah well I am not flawless).

But You know very well that the reason why I complain, is because you claimed you improved the speed of lyra and when I checked out your release (was expecting actually to use some of your "faster" kernels) I discovered there was absolutely no speed difference.
This clearly pissed me off... Because it means you have been mostly asking for donations over my work without adding anything new...
and this is a little unfair to me and other devs since donation are a little scarce around (Hey Bensam123   want to add something here ?).

I usually try to not compete with other devs so that everyone gets a little something out of their work, and I expect the same behavior
toward me.

I assume you're using SP and Pallas's work to make yours faster as well? Are you also giving them mentions in the miner itself? I imagine you can't be that thick skinned as to not use their work to make yours better due to pride...

[chrysophylax]

But You know very well that the reason why I complain, is because you claimed you improved the speed of lyra and when I checked out your release (was expecting actually to use some of your "faster" kernels) I

Look at the chart above. The first release with the lyra2v2 is your kernal. And then the speed is improved.


I have optimized your kernal on the gtx 950,960 and 970. The 980 is not a card for the farms because it is to expensive

The 980ti is same price or lower then 4 750ti and much more hashrate if some would implement it.
And get the mem clock up.

agreed ...

in benchmark - some algos show that its the hashrate of 5 x gigabyte 750ti oc lp cards ...

so if the optimization are focused on the 980ti ( which is the predecessor of pascal - which ever way you look at it ) then it would make it a better option to mine with ...

but those cards are STILL very expensive ...

this machine im building for my mate in total cost about $7900aud so far ( without a windows license - regardless of windows 10 x64 ) ... BUT the hashrate if ALL 6 cards were running is about 50% - 75% of the gigabyte 750ti oc lp farm im running now - depending on which algo ...

i cant wait till im settled in the next two weeks - so i can focus on all this again ...

#crysx

In Canada the price of a 980ti is about 5x 750ti, needs to drop some.

Did you look back in the thread for the old 6th card problem discussion?

i did try to - but its a LOOONG process of doing so - when there are no real search words that are unique to this issue ... search '6' or '6 cards' or 'driver issues' read through manually ... sp's thread is very long ...

if anyone you know where about ( pagewise ) i can look - ill search around there ...

here - the price is a little under 1070aud for the gigabyte 980ti g1 - and a little under 1100aud for the 980ti extreme ... so its very pricey ...

he actually had 10000aud to spare and spend - but i couldnt take advantage of that and buy myself two cards too ... its just not in me ... though he is allowing me to mine for a week when its finished - so there we have it ...

im off to the office in a few minutes ...

tanx joblo ...

#crysx

[djm34]

Djm34 is always complaining when somebody else is optimizing his opensource code. Me and Pallas optimized the neoscrypt kernal 20% faster, but it slower than the djm34 private kernal. I look forward to see the new code.

It is true that I can be a little protective over my work  (may-be a bit jealous too yeah well I am not flawless).

But You know very well that the reason why I complain, is because you claimed you improved the speed of lyra and when I checked out your release (was expecting actually to use some of your "faster" kernels) I discovered there was absolutely no speed difference.
This clearly pissed me off... Because it means you have been mostly asking for donations over my work without adding anything new...
and this is a little unfair to me and other devs since donation are a little scarce around (Hey Bensam123  want to add something here ?).

I usually try to not compete with other devs so that everyone gets a little something out of their work, and I expect the same behavior
toward me.

I assume you're using SP and Pallas's work to make yours faster as well? Are you also giving them mentions in the miner itself? I imagine you can't be that thick skinned as to not use their work to make yours better due to pride...

ccminer mentioned everyone who has work on it.
Now that something is used or not is based on a specific algo part. sp partial unrolling of the blake function is used, but not the work of Pallas as my code predate his, and I think (didn't check tbh) concerns the same part.
It really has nothing with pride or anything else. When I write an algo, I rethink entirely the way it works... It is very likely that the part he optimised doesn't exist anymore in the form which allowed the optimisation

[pallas]

Folding@EVGA http://forums.evga.com/tt.aspx?forumid=28 will get you more money quicker then any algo, $10. 00 for 2 days work for me with my hardware, takes 5 to 6 days with the same hardware to make $10 mining the top paying algo. ;-) I do folding at home for 2 days each month for the 10 EVGA Buck to go towards a new video card. Those EVGA buck bought me my new GTX960 GTX760 and 2x GTX660Ti and other video cards over the years. Sure you have to save them all year to really have anything since its a once a month thing and 10 buck max, but it still makes a nice Christmas gift for yourself.

http://folding.extremeoverclocking.com/user_summary.php?s=&u=610174

Are you "merge mining" foldingcoin, curecoin etc.?

[pokeytex]

SP - any reason why your miner does not work with Lyrabar?  I have tried many different ways to get it to work and sometimes can get it to hash but very slowly.  Please advise.

[sp_]

SP - any reason why your miner does not work with Lyrabar?  I have tried many different ways to get it to work and sometimes can get it to hash but very slowly.  Please advise.

Lyrabar has switched to the lyra2v2 algo.

[sp_]

I added some more quark hash on compute 5.2 devices(@github) +200KHASH(g1 970)

[hashbrown9000]

yeah that evga forums link is dead. what's the real one?

[pallas]

I added aome more quark hash on compute 5.2 devices(@github) +200KHASH(g1 970)

Thanks! Maybe those jh and keccak improvements are good for X11 as well.

[Schleicher]

yeah that evga forums link is dead. what's the real one?

http://forums.evga.com/FoldingEVGA-f28.aspx
http://www.evga.com/folding/

[kama]

hey guys.what do you think ? can this affect nvidia card's performance ?

http://www.win-raid.com/t892f16-AMD-and-Nvidia-GOP-update-No-requests-DIY.html

[pokeytex]

SP - any reason why your miner does not work with Lyrabar?  I have tried many different ways to get it to work and sometimes can get it to hash but very slowly.  Please advise.

Lyrabar has switched to the lyra2v2 algo.

Yes - I have been trying to use your versions and it doesn't work.  Do I have something wrong?

ccminer.exe -a lyra2v2 -o 127.0.0.1:4047 -u  -p  --no-gbt --no-longpoll --no-stratum

I should also mention I have tried version 76, 77, and now 78.