Allow to directly get api data in HTML5

Tested on Chrome... IE>=10 required, not tested

IE11 seems buggy on connection close...  todo

Signed-off-by: Tanguy Pruvot <tanguy.pruvot@gmail.com>

Tested on x11 which find sometimes 3 nonces in one call,
actually they are ignored because only the biggest was kept...

This commit doesnt fix that, but will allow to enhance shares rate later...

Added to most algos, checkhash function scans a big range
and can find multiple nonces at once if the difficulty is low.

Stop ignoring them, submit second one if found...

Clean the draft code for rc=2 implemented for blake and pentablake

btw... fix the reduced displayed hashrate when a nonce is found...

Signed-off-by: Tanguy Pruvot <tanguy.pruvot@gmail.com>

But use a define in AES to use or not device initial memcpy

I already tried to use everywhere direct device constants
and its not faster for big arrays (difference is small)

also change launch bounds to reduce spills (72 regs)

to check on windows too, could improve the perf... or not

Seems to be djm34 work, i recognize the code style ;)

Code was cleaned/indented and adapted to my fork...

Only usable on the test pool until 16 december 2014!

Use djm34 new uint2 method to get a +40% boost (115 to 153MH/s)

seems to resolve solo mining lock on share.
export also computed solo work diff in api (not perfect)

In high rate algos, throughput should be unsigned...
This fixes keccak, blake and doom problems

And change terminal color of debug lines, to be selectable in putty,
color code is not supported in windows but selection is ok there.

and disable multiple nonce on wallets,
a bloc cant be resolved twice ;)

Signed-off-by: Tanguy Pruvot <tanguy.pruvot@gmail.com>

before, only -b <ip:port>, <port> or 0 (disable) was allowed

easier to set -b 0.0.0.0 with default port

rejects in solo mode often means "generated block is stale" (see Debug.log)

Most of current wallets only reports "rejected" reason without explanations.

So prevent this common error,
you can turn off this gbt feature (height check) with --no-gbt option.

based on djm34 keccak 256-bit changes, and keep SM3.0 compat

affect most other algos too (quark, nist5, x13...)

Thanks to sp and djm34 for this fast uint64 storage alternative

hmm... this is weird, should be 256 but make bad shares on pool diff 2

Only enabled on first duplicate share

Not sure its the fastest way, but it works for offsets 0-63 + 64

Also note than asm SM 3.5+ doesn't support ROR with offset 64

quickly tested with a SM 3.0 binary...

todo: simd512 SM 2.x variant (shfl op), and groestl/myriad functions

priority param value is from 0 (default idle like before) to 5 (highest)

affinity could be enhanced to set multiple cores... todo

Signed-off-by: Tanguy Pruvot <tanguy.pruvot@gmail.com>

Some cards have 2 gpus on board...

even when no parameter is set (default set in each algo code)

do not impact perfs neither...

+ some cleanup... 15KH/s won (750Ti)

and clean the old variables, no more required

and fix windows build... missing __func define

This reverts commit a74df25.

This allow to "debug" easily in visual studio (without project changes)

Prepare version 1.5.3, update README

only allowed if --api-remote parameter or config key is set

and fix possible problem with urls containing user:password@

Signed-off-by: Tanguy Pruvot <tanguy.pruvot@gmail.com>

to keep current X11 output (sample: 2950.01 kH/s)

Yes, its a big commit, was waiting 1.6 to do that...
Sorry for your possible merge issues ;)

This enhances the hashrate when mining together with a cpu miner,
and doesn't slow down the machine, ccminer doesn't use a lot of cpu.

Note: this value doesn't require Admin rights on windows, unlike highest ones.

x11 perf is now close to x86 binary, but its not the case on all algos.

Add also a build.cmd script to remember how to build on windows cmdline

and convert ccminer.rc from utf-16 to Latin1 (doesnt want utf-8),
required for git diff/merge/rebase...

remove trailing spaces and move to x15 folder

add also cputest entry and do some cleanup...

bug reported by PVMining with a lot of cards (not seen here)

remains the cpu validation check to do...

throughput for this algo is divided by 128 to keep same kind of intensity values (default 18.0)

the comma presence check was wrong...

Main problem was the arrays allocations which should be made per cpu

Signed-off-by: Tanguy Pruvot <tanguy.pruvot@gmail.com>

and change algo names color to green...

will be exported in the api too...

the jansson error seems only seen in windows debug mode

did not import the extra final function, which should stay compatible
with the common cuda_check_hash()

uint4 copy + keccak cleanup, groestl: small uint4 opt

Signed-off-by: Tanguy Pruvot <tanguy.pruvot@gmail.com>

SKEIN512 + SHA256

Signed-off-by: Tanguy Pruvot <tanguy.pruvot@gmail.com>

Implementation based on klausT work.. a bit different

This code must be placed in a common .cu file,
cuda.cpp is not compiled with nvcc and doesnt allow cuda code...

could be also the source of the problem seen with CUDA 7

restored the code before sp/klaus changes for SM 3.0 devices...