hey SP, i emailed you from [email protected]
Wanting your BTC address so I can send you a few days worth of what I made using your miner. I love it, its way faster
I will continue to send a days worth each week
Email me back there, take care, and keep up the good work
Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 1237. (Read 2347588 times)
Try the miner. The pools will pay you. And the algo has 0 rounds of blake. Low reject rate. (Jackpointcoin)
I send exe files not screenshots.
The most interesting code is in the first exe file. It is actually finding valid nounces even if the blake hash was replaced by something else (jackpot coin)
The jackpotcoin contains 5 scientific chained hash algos from the NIST contest.
keccak512
blake512
groestl512
jh512
skein512
If you replace 512 rounds of blake with 5 million rounds my miner would be around 1 million times faster than the competition.
This is where I come from:
http://www.youtube.com/watch?v=N73F2ezz8r0
The most interesting code is in the first exe file. It is actually finding valid nounces even if the blake hash was replaced by something else (jackpot coin)
The jackpotcoin contains 5 scientific chained hash algos from the NIST contest.
keccak512
blake512
groestl512
jh512
skein512
If you replace 512 rounds of blake with 5 million rounds my miner would be around 1 million times faster than the competition.
This is where I come from:
http://www.youtube.com/watch?v=N73F2ezz8r0
Hit 9MH/s on my new 290X, clocks 1110/1625.
That is very good. I hope you don't release it. 90% of the GPU mining rigs are from AMD.
I wil try to push the 980 above 10MHASH, and then keep it private if the speed get's bether.Release it? Right now, it's priceless; that would make it worthless :3
I think it can get better, though.
fair game!
Your miner is 50% faster than the opensourceminer. That's very good. Did you consider to do a Claymore? I heard he had over $100 000 in alt coins now.
Good job!
Going with generic prices, without factoring in power consumption and density related costs, 1 MH/s of X11 costs:
0.124BTC on the 750 Ti;
0.133BTC on the 970;
0.189BTC on the 980 and;
0.110BTC on the 290x based on the stats from this thread.
I'm curious as to how those numbers will change.
Going with generic prices, without factoring in power consumption and density related costs, 1 MH/s of X11 costs:
0.124BTC on the 750 Ti;
0.133BTC on the 970;
0.189BTC on the 980 and;
0.110BTC on the 290x based on the stats from this thread.
I'm curious as to how those numbers will change.
Hit 9MH/s on my new 290X, clocks 1110/1625.
That is very good. I hope you don't release it. 90% of the GPU mining rigs are from AMD.
I wil try to push the 980 above 10MHASH, and then keep it private if the speed get's bether.
sp_ , I'm interessted to test your ccminer fork.
But my rig are mainly running on Kopiemtu (linux : https://litecointalk.org/index.php?topic=16800.0).
It's not possible to get the source ? If not, can you provide à tgz with a compiled version on this system ?
Or do you only provide windows binaries ?
Thanks, I would like to see how my 750ti rig will hash with your ccminer
But my rig are mainly running on Kopiemtu (linux : https://litecointalk.org/index.php?topic=16800.0).
It's not possible to get the source ? If not, can you provide à tgz with a compiled version on this system ?
Or do you only provide windows binaries ?
Thanks, I would like to see how my 750ti rig will hash with your ccminer
Blasülz, my latest mod is faster. Work is in progress so no source is released yet. 2 beta exe files(windows) has been sendt out to a few testers. It does 8MHASH on x11 (980) with heavy overclock. The 750ti is above 3MHASH. and the 970 7MHASH.
The next version will be faster. I have managed to optimize all the Hash functions in the x11 and x13 and x15 with some help from the other ccminer coders.
The next version will be faster. I have managed to optimize all the Hash functions in the x11 and x13 and x15 with some help from the other ccminer coders.
How well does the computer still function when this is running? I use the regular ccminer with no issues to folks that use the computers and it gets about 2500 kh/s on a 750ti.
I know that computer function of the miner doesn't matter to a lot of folks and usually not to me either but there are a few that I prefer to still have usable computers. tia
I know that computer function of the miner doesn't matter to a lot of folks and usually not to me either but there are a few that I prefer to still have usable computers. tia
you need cpu graphics or you need to change temp limit, so it throttles the card where the sreen is connected to
cheers
x11: 2750 - 2850 khash (nvminer1.2u-d8: 2740 - 2780 khash)
x13: 2100 - 2180 khash (nvminer1.2u-d8: 2090 - 2110 khash)
x15: 1880 - 1910 khash (nvminer1.2u-d8: 1840 - 1860 khash)
750ti, Core: 1332 Mhz, RAM: 3425 Mhz
---------------------------------------------------------
x11: 6350 - 6410 khash
Gigabyte Gaming G1 970 (Core: 1521 Mhz, RAM: 3000 Mhz)
Was there an updated version already? Couldn't follow the last days.
How well does the computer still function when this is running? I use the regular ccminer with no issues to folks that use the computers and it gets about 2500 kh/s on a 750ti.
I know that computer function of the miner doesn't matter to a lot of folks and usually not to me either but there are a few that I prefer to still have usable computers. tia
I know that computer function of the miner doesn't matter to a lot of folks and usually not to me either but there are a few that I prefer to still have usable computers. tia
Managed to fix quark and Nist5 now. I had to rollback my assembly blake and the optimalisations I merged from Schleicher.
750ti standard clock
quark/nist5
4950/ 8000
750ti oc.
quark/nist5
5600/9000
But all algorithms seems to work now.
750ti standard clock
quark/nist5
4950/ 8000
750ti oc.
quark/nist5
5600/9000
But all algorithms seems to work now.
Amazing, when its ready send it my way and I'll post GTX970 stats
Managed to fix quark and Nist5 now. I had to rollback my assembly blake and the optimalisations I merged from Schleicher.
750ti standard clock
quark/nist5
4950/ 8000
750ti oc.
quark/nist5
5600/9000
But all algorithms seems to work now.
750ti standard clock
quark/nist5
4950/ 8000
750ti oc.
quark/nist5
5600/9000
But all algorithms seems to work now.
My groestl fixes seems to increase x11 with around 20 khash on the 750ti. around 60khash on the 980. The groestl is now only the 3rd slowest of the x11 hashes.
// Experimental Kernel for Kepler (Compute 3.5) devices
// code submitted by nVidia performance engineer Alexey Panteleev
https://github.com/cbuchner1/CudaMiner/blob/35984c723eb786d614158daca9b07ac20de8645d/nv_kernel2.cu
It's been done before, in OpenCL. See Maxcoin's CL file.// code submitted by nVidia performance engineer Alexey Panteleev
https://github.com/cbuchner1/CudaMiner/blob/35984c723eb786d614158daca9b07ac20de8645d/nv_kernel2.cu
(Compute 3.5) is outdated. All kernals needs to be rewritten for maxwell for optimal performance. preferably in Assembly language. This work takes months.
Today I have integrated the changes from Schleicher that he posted on the ccminer thread into my mod:
Quote from: Schleicher on October 19, 2014, 05:07:14 PM
I managed to increase the quark and nist5 speed a little bit.
Source code:
https://github.com/KlausT/ccminer
simd512 seems to be 10% faster.
My stock 750TI is approaching 2700 without overclock. and 3000 with oclock.
This is without the faster Keccak (created by nvidia) in cudaminer.
These are the optimized numbers so far.
Blake xxx (not done)
skein 1.5%
BMW 60%
jh512 4.5%
keccac 1%
cubehash: 7.5%
shavite: 3.6%
simd512: 9,2%
fuge: 4,70%
hamsi: 6.97%
shabal: 22%
wirlpool:1.87%
echo: 5.5%
luffa: 0.4%
3 coders have contributed to the new speedup.
The sourcecode will be checked into the blakecoin fork by Epsylon3
https://bitcointalksearch.org/topic/ann-ccminer-23-opensource-gpl-tpruvot-770064
Quote from: Schleicher on October 19, 2014, 05:07:14 PM
I managed to increase the quark and nist5 speed a little bit.
Source code:
https://github.com/KlausT/ccminer
simd512 seems to be 10% faster.
My stock 750TI is approaching 2700 without overclock. and 3000 with oclock.
This is without the faster Keccak (created by nvidia) in cudaminer.
These are the optimized numbers so far.
Blake xxx (not done)
skein 1.5%
BMW 60%
jh512 4.5%
keccac 1%
cubehash: 7.5%
shavite: 3.6%
simd512: 9,2%
fuge: 4,70%
hamsi: 6.97%
shabal: 22%
wirlpool:1.87%
echo: 5.5%
luffa: 0.4%
3 coders have contributed to the new speedup.
The sourcecode will be checked into the blakecoin fork by Epsylon3
https://bitcointalksearch.org/topic/ann-ccminer-23-opensource-gpl-tpruvot-770064
That uint2 keccak was done by mtrlt, if I'm not mistaken.
// code submitted by nVidia performance engineer Alexey Panteleev
https://github.com/cbuchner1/CudaMiner/blob/35984c723eb786d614158daca9b07ac20de8645d/nv_kernel2.cu
It's been done before, in OpenCL. See Maxcoin's CL file.
(but, I think Reorder used it a lot)
And actually it makes no difference on compute 3.0 and 3.5 (might explain why it wasn't used anymore)
Today I managed a few percent in groestl!
All the kernals in x11 have now been optimized, but there is more potential.
All the kernals in x11 have now been optimized, but there is more potential.
Today I have integrated the changes from Schleicher that he posted on the ccminer thread into my mod:
Quote from: Schleicher on October 19, 2014, 05:07:14 PM
I managed to increase the quark and nist5 speed a little bit.
Source code:
https://github.com/KlausT/ccminer
simd512 seems to be 10% faster.
My stock 750TI is approaching 2700 without overclock. and 3000 with oclock.
This is without the faster Keccak (created by nvidia) in cudaminer.
These are the optimized numbers so far.
Blake xxx (not done)
skein 1.5%
BMW 60%
jh512 4.5%
keccac 1%
cubehash: 7.5%
shavite: 3.6%
simd512: 9,2%
fuge: 4,70%
hamsi: 6.97%
shabal: 22%
wirlpool:1.87%
echo: 5.5%
luffa: 0.4%
3 coders have contributed to the new speedup.
The sourcecode will be checked into the blakecoin fork by Epsylon3
https://bitcointalksearch.org/topic/ann-ccminer-23-opensource-gpl-tpruvot-770064
Quote from: Schleicher on October 19, 2014, 05:07:14 PM
I managed to increase the quark and nist5 speed a little bit.
Source code:
https://github.com/KlausT/ccminer
simd512 seems to be 10% faster.
My stock 750TI is approaching 2700 without overclock. and 3000 with oclock.
This is without the faster Keccak (created by nvidia) in cudaminer.
These are the optimized numbers so far.
Blake xxx (not done)
skein 1.5%
BMW 60%
jh512 4.5%
keccac 1%
cubehash: 7.5%
shavite: 3.6%
simd512: 9,2%
fuge: 4,70%
hamsi: 6.97%
shabal: 22%
wirlpool:1.87%
echo: 5.5%
luffa: 0.4%
3 coders have contributed to the new speedup.
The sourcecode will be checked into the blakecoin fork by Epsylon3
https://bitcointalksearch.org/topic/ann-ccminer-23-opensource-gpl-tpruvot-770064
That uint2 keccak was done by mtrlt, if I'm not mistaken.
// code submitted by nVidia performance engineer Alexey Panteleev
https://github.com/cbuchner1/CudaMiner/blob/35984c723eb786d614158daca9b07ac20de8645d/nv_kernel2.cu
Have you investigated whatever is going on with jackpotcoin? I emailed you what was going on when I tried mining it. It seemed like a nice improvement, but the pools only saw about 1/8 of the reported hash in the cmd prompt.
In the first version the blake implementation was commented out for jackpoint coin. 1/8 of the reported hash. What are the odds for that? The chained cryptohash is missing one of the algorithms and still reports found nounces. Is this the killer Blake?
Jump to: