CCminer(SP-MOD) Modded GPU kernels.

[pallas]

I have pull requested the fix for animecoin on util.cpp, let's hope it goes to the right fork this time ;-)

[antonio8]

Ether trying to sync at leat 16hours.Is it normal or smthg wrong? )

If you used the genesis_block.json in the geth then it should only take 10-15 minutes.

Once synced it always scrolls so you might be synced but not realizing it.

Keep an eye on the geth for the block number and check an explorer to see if they match.

EDIT: I see scryptr already mentioned most of this.

[scryptr]

GENOIL--

Genoil will be posting an SP_ modded Windows Ethminer binary sometime today.  He announced this on his thread, page 7, linked a few posts back.        --scryptr

EDIT:  It is posted!  The link is on the bottom of page 7.  CPU usage is waaaaay down.       --scryptr

[chrysophylax]

I have pull requested the fix for animecoin on util.cpp, let's hope it goes to the right fork this time ;-)

nice ...

so that would mean that we wouldnt need to edit and can compile direct from git? ...

if sp includes it that is - is that right? ...

#crysx

[pallas]

I have pull requested the fix for animecoin on util.cpp, let's hope it goes to the right fork this time ;-)

nice ...

so that would mean that we wouldnt need to edit and can compile direct from git? ...

if sp includes it that is - is that right? ...

#crysx

Yes, it compiles for me now; and SP_ has already merged the pull request, so you can go on git pulling ;-)

[chrysophylax]

I have pull requested the fix for animecoin on util.cpp, let's hope it goes to the right fork this time ;-)

nice ...

so that would mean that we wouldnt need to edit and can compile direct from git? ...

if sp includes it that is - is that right? ...

#crysx

Yes, it compiles for me now; and SP_ has already merged the pull request, so you can go on git pulling ;-)

well - i can only do it remotely at the moment ...

not as satisfying as pulling at the office ...

git that is - pulling git ... ahem ...

think ill just clone and start over ... hehehe ...

tanx ...

#crysx

[pallas]

I have pull requested the fix for animecoin on util.cpp, let's hope it goes to the right fork this time ;-)

nice ...

so that would mean that we wouldnt need to edit and can compile direct from git? ...

if sp includes it that is - is that right? ...

#crysx

Yes, it compiles for me now; and SP_ has already merged the pull request, so you can go on git pulling ;-)

well - i can only do it remotely at the moment ...

not as satisfying as pulling at the office ...

git that is - pulling git ... ahem ...

think ill just clone and start over ... hehehe ...

tanx ...

#crysx

if you modified some files yourself and it refuses to do "git pull", just run "git checkout [file]" to revert it to the original state ;-)

[Genoil]

edit: I get "illegal memory access" with the shared memory method

What architecture? The benchmark (-U -M) and difficulty (-U -S) tests run fine on the shared memory code.
I don't do real mining  .

To switch between shared on shuffled , you need to  modify SHUFFLE_MIN_VER (on top of the .cu file) from 350 to i.e. 600 (fictive Compute 6.0).
  

[NiceHashSupport]

sp, and all other mining developers, I suggest you to watch development of this: https://bitcointalk.org/index.php?topic=895020.msg12164065#msg12164065

There has to be a census reached about Lyra2REv2 diff multiplicator.

[antonio8]

@Genoil

I just posted in your thread.

[Genoil]

@anyone who wants to poke around in the Ethereum CUDA miner source, i've implemented sp_'s mods and also cleaned it up a bit.

Windows build instructions:

- get and install cmake and Visual Studio 12 2013 (Express)
- download/git clone source
- run cpp-ethereum/extdep/getstuff.bat
- cd cpp-ethereum
- mkdir build
- cd build
- cmake -DBUNDLE=miner -DETHASHCU=1 [-DCOMPUTE=XX] -G "Visual Studio 12 2013 Win64" ..
- launch generated .sln
- set ethminer as startup project
- build/run

[-DCOMPUTE=XX] is optional. i.e for GTX 9x0 (Compute 5.2) specificy -DCOMPUTE52. Compiles faster.

[djm34]

sp, and all other mining developers, I suggest you to watch development of this: https://bitcointalk.org/index.php?topic=895020.msg12164065#msg12164065

There has to be a census reached about Lyra2REv2 diff multiplicator.

ok, here the new release, for this one you don't need to use "-diff 2" anymore: https://github.com/djm34/ccminer-lyra/releases/tag/v0.5
sp_ please update your release as well (as far as I know all the pool now use a difficulty of 256, in case one isn't using it --diff 0.5 will be necessary  )

edit: careful if you are using give-me-coin, they still use a difficulty of 128, you'll need to use diff of 0.5
would have been nice if yopu had search for a consensus among the pool, before posting that 

[djm34]

@anyone who wants to poke around in the Ethereum CUDA miner source, i've implemented sp_'s mods and also cleaned it up a bit.

Windows build instructions:

- get and install cmake and Visual Studio 12 2013 (Express)
- download/git clone source
- run cpp-ethereum/extdep/getstuff.bat
- cd cpp-ethereum
- mkdir build
- cd build
- cmake -DBUNDLE=miner -DETHASHCU=1 [-DCOMPUTE=XX] -G "Visual Studio 12 2013 Win64" ..
- launch generated .sln
- set ethminer as startup project
- build/run

[-DCOMPUTE=XX] is optional. i.e for GTX 9x0 (Compute 5.2) specificy -DCOMPUTE52. Compiles faster.

from what I checked, using uint2 was a bit slower (for once) than using uint64_t in keccak

[sp_]

sp, and all other mining developers, I suggest you to watch development of this: https://bitcointalk.org/index.php?topic=895020.msg12164065#msg12164065
There has to be a census reached about Lyra2REv2 diff multiplicator.

ok, here the new release, for this one you don't need to use "-diff 2" anymore: https://github.com/djm34/ccminer-lyra/releases/tag/v0.5
sp_ please update your release as well (as far as I know all the pool now use a difficulty of 256, in case one isn't using it --diff 0.5 will be necessary  )
edit: careful if you are using give-me-coin, they still use a difficulty of 128, you'll need to use diff of 0.5
would have been nice if yopu had search for a consensus among the pool, before posting that 

Ok. I have changed it now.. 100% more hash if you mine at nicehash.  nice he-he

[sp_]

Today's share sent
djm34: 047171eabd9db41dea5d4f8aece3647bd6990118c09c4ef0c056933e6920f5a7
pallas: ff450e633bec606b921855a307a4fbad1f37c1afa23e69ba426184fe7b2cae3f
sp_: 13acb0a16bb2c0252f04a864c24e59ba8e92cbc0080809476c6c15ad6bc2bfaf
Thanks. Keep going

Thanks

[sp_]

@anyone who wants to poke around in the Ethereum CUDA miner source, i've implemented sp_'s mods and also cleaned it up a bit.

from what I checked, using uint2 was a bit slower (for once) than using uint64_t in keccak

No it isn't, but I think you have to seperate the loops into two and remove all the constant inputs, load from memory differenty, play with the pragma urolls, alter the launch bounds etc..

But keccak was just 13% of the total algorithm so the effort will probobly just gain a few percent.

[djm34]

@anyone who wants to poke around in the Ethereum CUDA miner source, i've implemented sp_'s mods and also cleaned it up a bit.

from what I checked, using uint2 was a bit slower (for once) than using uint64_t in keccak

No it isn't, but I think you have to seperate the loops into two and remove all the constant inputs, load from memory differenty, play with the pragma urolls, alter the launch bounds etc..

But keccak was just 13% of the total algorithm so the effort will probobly just gain a few percent.

I tried yesterday it was slower on my 980

[dominuspro]

@anyone who wants to poke around in the Ethereum CUDA miner source, i've implemented sp_'s mods and also cleaned it up a bit.

from what I checked, using uint2 was a bit slower (for once) than using uint64_t in keccak

No it isn't, but I think you have to seperate the loops into two and remove all the constant inputs, load from memory differenty, play with the pragma urolls, alter the launch bounds etc..

But keccak was just 13% of the total algorithm so the effort will probobly just gain a few percent.

I tried yesterday it was slower on my 980

I can confirm it beeing a little bit slower on 970 and 960.

[sp_]

yes, but it will be faster with a small rewrite.

Seperate the loops into two(keccakfirst,. keccak last) and remove all the constant work (less instructions),
With the #pragma unroll and __forceinline__ I hoped that the the compiler would remove the constant work but it doesn't understand that rol(constant,constant) etc can be precalculated.. Tune, compile and study the ptx assembly code. Less instructions is normally bether..

Load from memory differenty(remove the uint2 cast) use maxwells  vector instuctions, play with the pragma unrolls(but make sure that you don't exceed the instructioncache), alter the launch bounds to tune the registers etc..

[hashbrown9000]

--diff 2 never worked for me on nicehash. it always reported 1/5 of my total hash.  which diff should i be using now?