[VTC] Lyra2RE Vertcoin optimized miner, NOW WITH SOURCES!

[pallas]

A tip for the guys who use older drivers with wolf0's binaries.

Since this build is made for the latest drivers, I think you need to copy the latest opencl files(omega dll's) into the sgminer folder before you run the miner.

Or upgrade. But an upgrade could cause the other algos to run at 50% speed.

You should just need to copy the newer opencl.dll (or libopencl.so on linux) in order to make your own bins.

I forgot to say that if you want to use libopencl.so from a different driver on linux, you can't just copy it.
Better put it in a related directory (say "/usr/local/lib"), run "lddconfig" on that directory and check it with "ldd sgminer".

[pallas]

total unpaid balance: 0.00001106 BTC

wow lotta donation hashing! :-D
and only one user saying thank you.
but of that 90 users on nicehash lyra, surely a nice part is using my miner.

(this makes me think I should switch to nvidia, much more grateful users there)

ooooooo why am I saying this over again? I should know by now!!! :-D

[fenomenhaa]

total unpaid balance: 0.00001106 BTC

wow lotta donation hashing! :-D
and only one user saying thank you.
but of that 90 users on nicehash lyra, surely a nice part is using my miner.

(this makes me think I should switch to nvidia, much more grateful users there)

ooooooo why am I saying this over again? I should know by now!!! :-D

Hi Pallas,
Nowadays Amd gpus are not profitable,that's why peoples lose their interests to mine with amd cards,maybe cus of this.When i earn some bucks with your miner, i'll donate as usual ))) .I can't be sure but if you can make Amd gpus profitable again, people will focused Amd cards again and some of them be ready for donations(at least myself) )) .Thanks for your works and sharing M8...

[smolen]

(this makes me think I should switch to nvidia, much more grateful users there)

Or to operating your own pool

[MaxDZ8]

Nowadays Amd gpus are not profitable,that's why peoples lose their interests to mine with amd cards,maybe cus of this.

Maybe, or maybe not. AMD cards have demonstrated multiple times they work great when given proper kernels but apparently those guys are more interested in giving money to NV (which doesn't give a fuck to budget users and it's becoming borderline patent troll recently) than rewarding the few talented individuals effectively doing research on that topic.

[amateo613]

Congrats Pallas

Your Miner/Kernel combination allowed me to tap into my Lenovo intel graphics for another 100Kh. I appreciate it.
Do the other kernels work I tried them multiple times but I can not get them to work on the intel graphics but the Lyra works great.

Thanks again.

[thevictimofuktyranny]

What speed are people getting with this miner with the omega 14.12 drivers?

[pallas]

Congrats Pallas

Your Miner/Kernel combination allowed me to tap into my Lenovo intel graphics for another 100Kh. I appreciate it.
Do the other kernels work I tried them multiple times but I can not get them to work on the intel graphics but the Lyra works great.

Thanks again.

Glad it works on Intel GPUs as well! :-)
About the other kernels, you can try adding the missing .cl files into the kernel folder, getting them from the stock sgminer. Not guaranteed to work....

[pallas]

What speed are people getting with this miner with the omega 14.12 drivers?

I've been using omega drivers when developing it.

[rednoW]

Just for fun: managed to run it on my gtx 750.
It gives 650 khs instead of 880khs with sp_'s ccminer. Not bad but cpu usage is huge.

In order to make it work I needed to change long long constants in Lyra2cl to unsigned long as long long is not supported in nvidia opencl

__constant static const sph_u64 blake2b_IV[8] =
{
  0x6a09e667f3bcc908UL, 0xbb67ae8584caa73bUL,
  0x3c6ef372fe94f82bUL, 0xa54ff53a5f1d36f1UL,
  0x510e527fade682d1UL, 0x9b05688c2b3e6c1fUL,
  0x1f83d9abfb41bd6bUL, 0x5be0cd19137e2179UL
};

Then it was an error in keccak1600.cl with

inline void keccak_block(ulong *s) {

I added static
inline static void keccak_block(ulong *s) {

and the problems seems to vanish.

Then it was an error with 2d array init skein256.cl

__constant static const ulong ROT256[8][4] = {
   46, 36, 19, 37,
   33, 27, 14, 42,
   17, 49, 36, 39,
   44, 9, 54, 56,
   39, 30, 34, 24,
   13, 50, 10, 17,
   25, 29, 39, 43,
   8, 35, 56, 22,
};

I've added more {} to form this

__constant static const ulong ROT256[8][4] = {
   {46, 36, 19, 37},
   {33, 27, 14, 42},
   {17, 49, 36, 39},
   {44, 9, 54, 56},
   {39, 30, 34, 24},
   {13, 50, 10, 17},
   {25, 29, 39, 43},
   {8, 35, 56, 22}
};

And it finally got compiled.
It works and receives "accepted" from the pool )))

I used "intensity" : "14",   "worksize" : "512",  "gpu-threads" : "1" to run ...
gpu_z shows 100% load for gpu engine and memory, the same is with ccminer )))

[pallas]

Just for fun: managed to run it on my gtx 750.
It gives 650 khs instead of 880khs with sp_'s ccminer. Not bad but cpu usage is huge.

In order to make it work I needed to change long long constants in Lyra2cl to unsigned long as long long is not supported in nvidia opencl

__constant static const sph_u64 blake2b_IV[8] =
{
  0x6a09e667f3bcc908UL, 0xbb67ae8584caa73bUL,
  0x3c6ef372fe94f82bUL, 0xa54ff53a5f1d36f1UL,
  0x510e527fade682d1UL, 0x9b05688c2b3e6c1fUL,
  0x1f83d9abfb41bd6bUL, 0x5be0cd19137e2179UL
};

Then it was an error in keccak1600.cl with

inline void keccak_block(ulong *s) {

I added static
inline static void keccak_block(ulong *s) {

and the problems seems to vanish.

Then it was an error with 2d array init skein256.cl

__constant static const ulong ROT256[8][4] = {
   46, 36, 19, 37,
   33, 27, 14, 42,
   17, 49, 36, 39,
   44, 9, 54, 56,
   39, 30, 34, 24,
   13, 50, 10, 17,
   25, 29, 39, 43,
   8, 35, 56, 22,
};

I've added more {} to form this

__constant static const ulong ROT256[8][4] = {
   {46, 36, 19, 37},
   {33, 27, 14, 42},
   {17, 49, 36, 39},
   {44, 9, 54, 56},
   {39, 30, 34, 24},
   {13, 50, 10, 17},
   {25, 29, 39, 43},
   {8, 35, 56, 22}
};

And it finally got compiled.
It works and receives "accepted" from the pool )))

I used "intensity" : "14",   "worksize" : "512",  "gpu-threads" : "1" to run ...
gpu_z shows 100% load for gpu engine and memory, the same is with ccminer )))

Thanks very much for testing this!
Long long is not used on amd, there is an ifdef which defines sph_u64 differently on nvidia (legacy sph stuff I didn't care to polish yet).
I will update the two archives and the post, to add intel and nvidia compatibility, later.
High cpu usage can be avoided by setting GPU_USE_SYNC_OBJECTS=1, don't know if there is an equivalent for nvidia.

[Epsylon3]

in fact its just a coding copy paste typo from cpu or ccminer code...

OpenCL doesnt supports ULL, because UL is the same... 64bit

I guess the amd compiler "handles" it but its just wrong...

[pallas]

in fact its just a coding copy paste typo from cpu or ccminer code...

Originally from sph -sgminer, not a copy paste typo, just bad precompiler code ;-)

[Dogeshop_eu]

I'm having 800kh on my r 290 ( 3GB )  what's the problem?

[pallas]

I'm having 800kh on my r 290 ( 3GB )  what's the problem?

Which driver version are you using?
Which settings: clocks, intensity etc.?
Please paste the full commandline or conf file.

[Dogeshop_eu]

start file:

setx GPU_FORCE_64BIT_PTR 0
setx GPU_MAX_HEAP_SIZE 100
setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 100

sgminer.exe -c lyra2re.conf

pause

config:

{
"pools" : [
        {
                "name" : "NiceHash_Lyra2RE",
                "url" : "ca.simplemulti.com:3374",
                "user" : "VpzyncgawVk3kC7Yp5M8YHVcS97bYLvQCs",
                "pass" : "x",
                "algorithm" : "Lyra2RE",
                "intensity" : "18",
                "gpu-engine" : "935",
                "gpu-memclock" : "1500",
                "worksize" : "128",
                "gpu-threads" : "2"
        },
   
}


Drivers the latest  15.4 Beta

[pallas]

{
"pools" : [
        {
                "name" : "NiceHash_Lyra2RE",
                "url" : "ca.simplemulti.com:3374",
                "user" : "VpzyncgawVk3kC7Yp5M8YHVcS97bYLvQCs",
                "pass" : "x",
                "algorithm" : "Lyra2RE",
                "intensity" : "18",
                "gpu-engine" : "935",
                "gpu-memclock" : "1500",
                "worksize" : "128",
                "gpu-threads" : "2"
        },
   
}

Drivers the latest  15.4 Beta

try:
"intensity" : "15",
"gpu-engine" : "835",

I never tried compiling the opencl code using 15.4: it might need tweaking for that version.
You can try 14.12 or I can send you the precompiled bin (64 bit for hawaii).

[Dogeshop_eu]

What should I do with the bin file ( i'm newbie in mining ) please send mik_druid@poczta.fm

[Dogeshop_eu]

But i think it's probably drivers' fault

[pallas]

But i think it's probably drivers' fault

if you are on windows, you can keep your driver and compile the opencl on 14.12 just by putting opencl.dll in the miner's folder.