OpenCL GPU miner
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 
ystarnaud 81d9c9e574 Events framework 10 years ago
ADL_SDK doc: wrong link in ADL_SDK readme.txt. 11 years ago
algorithm Added Neoscrypt with Wolf9466 improvements. 10 years ago
ccan VS2010 build: Fixed compatibility with non-MSVS compilers. 11 years ago
doc Events framework 10 years ago
kernel Added Neoscrypt with Wolf9466 improvements. 10 years ago
lib Compile CPU mining for win32 and win64 12 years ago
m4 Mingw suseconds_t and sigaction fixes. 14 years ago
man added sgminer.1 manpage file, draft 11 years ago
ocl Refactored kernel source loading (missing / delimiter). Added more debug messages. 11 years ago
sph Add shabal.c and whirlpool.c to source list 11 years ago
submodules Updated to MSVS 2013 10 years ago
tools tools: typo in update_changelog.sh. 11 years ago
winbuild Events framework 10 years ago
.gitignore VS2010 build: move windows config.h gitignore rule to winbuild/.gitignore 11 years ago
.gitmodules Added Jansson as submodule. 11 years ago
API.class API.java allow partial reads 12 years ago
API.java basic copyright statement in API.java 12 years ago
AUTHORS.md Updated AUTHORS.md 11 years ago
COPYING Update licensing to GPL V3. 13 years ago
ChangeLog update changelog for v5 release 10 years ago
MCast.class misc: Replace all remaining instances of 'cgminer' with 'sgminer'. 11 years ago
MCast.java misc: Replace all remaining instances of 'cgminer' with 'sgminer'. 11 years ago
Makefile.am Events framework 10 years ago
NEWS.md bump version to 5.0.0, update changelog 10 years ago
README.md Add debug-log documentation and enable it by default 11 years ago
adl.c Updated to MSVS 2013 10 years ago
adl.h More options added 11 years ago
adl_functions.h Support building in cygwin. 11 years ago
algorithm.c Added Neoscrypt with Wolf9466 improvements. 10 years ago
algorithm.h Added Neoscrypt with Wolf9466 improvements. 10 years ago
api-example.c Added Jansson as submodule. 11 years ago
api-example.php misc: Replace all remaining instances of 'cgminer' with 'sgminer'. 11 years ago
api-example.py misc: Replace all remaining instances of 'cgminer' with 'sgminer'. 11 years ago
api.c Fix for UI/API intensity changes 10 years ago
api.h Fix for UI/API intensity changes 10 years ago
arg-nonnull.h Added previously missing gnulib files. 14 years ago
autogen.sh Don't run configure automatically. 11 years ago
bench_block.h misc: Replace all remaining instances of 'cgminer' with 'sgminer'. 11 years ago
c++defs.h Added previously missing gnulib files. 14 years ago
compat.h Added missing permission flags to Windows build #333 11 years ago
config_parser.c Events framework 10 years ago
config_parser.h Fix for UI/API intensity changes 10 years ago
configure.ac Updated to MSVS 2013 10 years ago
driver-opencl.c Fix for UI/API intensity changes 10 years ago
driver-opencl.h Fix worksize per GPU 10 years ago
elist.h VS2010 build: commit elist.h and relevant changes in sgminer.c and util.c 11 years ago
events.c Events framework 10 years ago
events.h Events framework 10 years ago
example.conf Updated example.conf 11 years ago
findnonce.c convert tabs to spaces in most commonly modified files 11 years ago
findnonce.h Add MaxCoin support 11 years ago
hexdump.c VS2010 build: Prepare hexdump.c 11 years ago
logging.c Fix for UI/API intensity changes 10 years ago
logging.h Fix for UI/API intensity changes 10 years ago
miner.h Fix for UI/API intensity changes 10 years ago
miner.php misc: Replace all remaining instances of 'cgminer' with 'sgminer'. 11 years ago
mknsis.sh Windows build tweaks. 14 years ago
ocl.c Added Neoscrypt with Wolf9466 improvements. 10 years ago
ocl.h convert tabs to spaces in most commonly modified files 11 years ago
pool.c config: write pool name and description to config. 11 years ago
pool.h pool: do not use defines, just inline strings. 11 years ago
sgminer.c Events framework 10 years ago
sha2.c Carve out the unused portions of sha2 implementation. 11 years ago
sha2.h Carve out the unused portions of sha2 implementation. 11 years ago
uthash.h Update uthash to latest. 11 years ago
util.c After cancelling a thread (unless it's detached, but mining threads aren't), it is always necessary to join the thread so the system will release resources. 11 years ago
util.h After cancelling a thread (unless it's detached, but mining threads aren't), it is always necessary to join the thread so the system will release resources. 11 years ago
warn-on-use.h Added previously missing gnulib files. 14 years ago

README.md

sgminer

Introduction

This is a multi-threaded multi-pool GPU miner with ATI GPU monitoring, (over)clocking and fanspeed support for scrypt-based cryptocurrency. It is based on cgminer by Con Kolivas (ckolivas), which is in turn based on cpuminer by Jeff Garzik (jgarzik).

releases: https://github.com/sgminer-dev/sgminer/releases

git tree: https://github.com/sgminer-dev/sgminer

bugtracker: https://github.com/sgminer-dev/sgminer/issues

irc: #sgminer and #sgminer-dev on freenode

mailing lists: https://sourceforge.net/p/sgminer/mailman/

License: GPLv3. See COPYING for details.

Documentation

Documentation is available in directory doc. It is organised by topics:

  • API for the RPC API specification;
  • configuration.md for (largely incomplete) detailed information on all configuration options;
  • FAQ.md for frequently asked questions;
  • GPU for semi-obsolete information on GPU configuration options and mining SHA256d-based coins;
  • kernel.md for OpenCL kernel-related information, including development procedure;
  • MINING.md for how to find the right balance in GPU configuration to mine Scrypt-based coins efficiently;
  • windows-build.txt for information on how to build on Windows.

Note that most of the documentation is outdated or incomplete. If you want to contribute, fork this repository, update as needed, and submit a pull request.

Building

Dependencies

Mandatory:

Optional:

  • curses dev library - libncurses5-dev on Debian or libpdcurses on WIN32, for text user interface
  • AMD ADL SDK - version 6, required for ATI GPU monitoring & clocking

If building from git:

  • autoconf
  • automake

sgminer-specific configuration options:

--disable-adl           Override detection and disable building with adl
--disable-adl-checks
--without-curses        Do not compile support for curses TUI

Debian Example

apt-get install libcurl4-openssl-dev pkg-config libtool libncurses5-dev

AMD APP SDK and AMD ADL SDK must be downloaded from the amd websites.

*nix build instructions

If needed, place include headers (*.h files) from ADL_SDK_*<VERSION>*.zip in sgminer/ADL_SDK.

Then:

git submodule init
git submodule update
autoreconf -i
CFLAGS="-O2 -Wall -march=native" ./configure <options>
make

To compile a version that can be used accross machines, remove -march=native.

To compile a debug version, replace -O2 with -ggdb.

Systemwide installation is optional. You may run sgminer from the build directory directly, or make install if you wish to install sgminer to a system location or a location you specified with --prefix.

Windows build instructions

See doc/windows-build.txt for MinGW compilation and cross-compiation, doc/cygwin-build.txt for building using Cygwin, or use the provided winbuild Microsoft Visual Studio project (tested on MSVS2010), with instructions in winbuild/README.txt.

Basic Usage

WARNING: documentation below this point has not been updated since the fork.

After saving configuration from the menu, you do not need to give sgminer any arguments and it will load your configuration.

Any configuration file may also contain a single

"include" : "filename"

to recursively include another configuration file.

Writing the configuration will save all settings from all files in the output.

Single pool:

sgminer -o http://pool:port -u username -p password

Multiple pools:

sgminer -o http://pool1:port -u pool1username -p pool1password -o http://pool2:port -u pool2usernmae -p pool2password

Single pool with a standard http proxy, regular desktop:

sgminer -o "http:proxy:port|http://pool:port" -u username -p password

Single pool with a socks5 proxy, regular desktop:

sgminer -o "socks5:proxy:port|http://pool:port" -u username -p password

Single pool with stratum protocol support:

sgminer -o stratum+tcp://pool:port -u username -p password

The list of proxy types are: http: standard http 1.1 proxy http0: http 1.0 proxy socks4: socks4 proxy socks5: socks5 proxy socks4a: socks4a proxy socks5h: socks5 proxy using a hostname

If you compile sgminer with a version of CURL before 7.19.4 then some of the above will not be available. All are available since CURL version 7.19.4.

If you specify the --socks-proxy option to sgminer, it will only be applied to all pools that don't specify their own proxy setting like above.

For more advanced usage , run sgminer --help.

See doc/GPU for more information regarding GPU mining and doc/SCRYPT for more information regarding Scrypt mining.

Runtime usage

The following options are available while running with a single keypress:

[P]ool management [G]PU management [S]ettings [D]isplay options [Q]uit

P gives you:

Current pool management strategy: Failover [F]ailover only disabled [A]dd pool [R]emove pool [D]isable pool [E]nable pool [C]hange management strategy [S]witch pool [I]nformation

S gives you:

[Q]ueue: 1 [S]cantime: 60 [E]xpiry: 120 [W]rite config file [C]gminer restart

D gives you:

[N]ormal [C]lear [S]ilent mode (disable all output) [D]ebug:off [P]er-device:off [Q]uiet:off [V]erbose:off [R]PC debug:off [W]orkTime details:off co[M]pact: off [L]og interval:5

Q quits the application.

G gives you something like:

GPU 0: [124.2 / 191.3 Mh/s] [A:77 R:33 HW:0 U:1.73/m WU 1.73/m] Temp: 67.0 C Fan Speed: 35% (2500 RPM) Engine Clock: 960 MHz Memory Clock: 480 Mhz Vddc: 1.200 V Activity: 93% Powertune: 0% Last initialised: [2011-09-06 12:03:56] Thread 0: 62.4 Mh/s Enabled ALIVE Thread 1: 60.2 Mh/s Enabled ALIVE

[E]nable [D]isable [R]estart GPU [C]hange settings Or press any other key to continue

The running log shows output like this:

[2012-10-12 18:02:20] Accepted f0c05469 Diff 1/1 GPU 0 pool 1 [2012-10-12 18:02:22] Accepted 218ac982 Diff 7/1 GPU 1 pool 1 [2012-10-12 18:02:23] Accepted d8300795 Diff 1/1 GPU 3 pool 1 [2012-10-12 18:02:24] Accepted 122c1ff1 Diff 14/1 GPU 1 pool 1

The 8 byte hex value are the 2nd 8 bytes of the share being submitted to the pool. The 2 diff values are the actual difficulty target that share reached followed by the difficulty target the pool is currently asking for.

The output line shows the following: (5s):1713.6 (avg):1707.8 Mh/s | A:729 R:8 HW:0 WU:22.53/m

Each column is as follows: 5s: A 5 second exponentially decaying average hash rate avg: An all time average hash rate A: The total difficulty of Accepted shares R: The total difficulty of Rejected shares HW: The number of HardWare errors WU: The Work Utility defined as the number of diff1 shares work / minute (accepted or rejected).

GPU 1: 73.5C 2551RPM | 427.3/443.0Mh/s | A:8 R:0 HW:0 WU:4.39/m

Each column is as follows: Temperature (if supported) Fanspeed (if supported) A 5 second exponentially decaying average hash rate An all time average hash rate The total difficulty of accepted shares The total difficulty of rejected shares The number of hardware erorrs The work utility defined as the number of diff1 shares work / minute

The sgminer status line shows: ST: 1 SS: 0 NB: 1 LW: 8 GF: 1 RF: 1

ST is STaged work items (ready to use). SS is Stale Shares discarded (detected and not submitted so don't count as rejects) NB is New Blocks detected on the network LW is Locally generated Work items GF is Getwork Fail Occasions (server slow to provide work) RF is Remote Fail occasions (server slow to accept work)

The block display shows: Block: 0074c5e482e34a506d2a051a... Started: [17:17:22] Best share: 2.71K

This shows a short stretch of the current block, when the new block started, and the all time best difficulty share you've found since starting sgminer this time.

Multipool

Failover strategies

A number of different strategies for dealing with multipool setups are available. Each has their advantages and disadvantages so multiple strategies are available by user choice, as per the following list:

Failover

The default strategy is failover. This means that if you input a number of pools, it will try to use them as a priority list, moving away from the 1st to the 2nd, 2nd to 3rd and so on. If any of the earlier pools recover, it will move back to the higher priority ones.

Round robin

This strategy only moves from one pool to the next when the current one falls idle and makes no attempt to move otherwise.

Rotate

This strategy moves at user-defined intervals from one active pool to the next, skipping pools that are idle.

Load balance

This strategy sends work to all the pools on a quota basis. By default, all pools are allocated equal quotas unless specified with --quota. This apportioning of work is based on work handed out, not shares returned so is independent of difficulty targets or rejected shares. While a pool is disabled or dead, its quota is dropped until it is re-enabled. Quotas are forward looking, so if the quota is changed on the fly, it only affects future work. If all pools are set to zero quota or all pools with quota are dead, it will fall back to a failover mode. See quota below for more information.

The failover-only flag has special meaning in combination with load-balance mode and it will distribute quota back to priority pool 0 from any pools that are unable to provide work for any reason so as to maintain quota ratios between the rest of the pools.

Balance

This strategy monitors the amount of difficulty 1 shares solved for each pool and uses it to try to end up doing the same amount of work for all pools.

Quotas

The load-balance multipool strategy works off a quota based scheduler. The quotas handed out by default are equal, but the user is allowed to specify any arbitrary ratio of quotas. For example, if all the quota values add up to 100, each quota value will be a percentage, but if 2 pools are specified and pool0 is given a quota of 1 and pool1 is given a quota of 9, pool0 will get 10% of the work and pool1 will get 90%. Quotas can be changed on the fly by the API, and do not act retrospectively. Setting a quota to zero will effectively disable that pool unless all other pools are disabled or dead. In that scenario, load-balance falls back to regular failover priority-based strategy. While a pool is dead, it loses its quota and no attempt is made to catch up when it comes back to life.

To specify quotas on the command line, pools should be specified with a semicolon separated --quota(or -U) entry instead of --url. Pools specified with --url are given a nominal quota value of 1 and entries can be mixed.

For example: --url poola:porta -u usernamea -p passa --quota "2;poolb:portb" -u usernameb -p passb Will give poola 1/3 of the work and poolb 2/3 of the work.

Writing configuration files with quotas is likewise supported. To use the above quotas in a configuration file they would be specified thus:

"pools" : [
    {
            "url" : "poola:porta",
            "user" : "usernamea",
            "pass" : "passa"
    },
    {
            "quota" : "2;poolb:portb",
            "user" : "usernameb",
            "pass" : "passb"
    }
]

Extra File Configuration

If you want to store a number of pools in your configuration file, but don't always want them automatically enabled at start up (or restart), then the "state" option with a value of "disabled" can be used:

"pools" : [
    {
            "url" : "poola:porta",
            "user" : "usernamea",
            "pass" : "passa"
    },
    {
            "quota" : "2;poolb:portb",
            "user" : "usernameb",
            "pass" : "passb",
            "state" : "disabled"
    }
]

It is then trivial to change the "state" setting to "enabled" in the configuration file at anytime and then restart the miner (see below). You can enable the pool whilst the miner is still running ('p' followed by 'e' followed by pool number) - but the pool will still be disabled on restart if the config file is not changed.

"state" can also be set to "hidden". This allows the json file to contain a large number of pools, of which some could be automatically culled at start up. This makes it easy to swap pools in and out of the runtime selection, without having a large list of pools cluttering up the display.

"pools" : [
    {
            "poolname" : "Main Pool",
            "url" : "poola:porta",
            "user" : "usernamea",
            "pass" : "passa",
            "state" : "disabled"
    },
    {
            "poolname" : "Joe's Weekend Pool",
            "quota" : "2;poolb:portb",
            "user" : "usernameb",
            "pass" : "passb",
            "state" : "hidden"
    }
]

These options are considered experimental and therefore will NOT be created when the 'Write config file' option is used ('s' followed by 'w').

A restart of the miner ('s' followed by 'c') will reload the config file and any changes that may have been made.

Logging

sgminer will log to stderr if it detects stderr is being redirected to a file. To enable logging simply append 2>logfile.txt to your command line and logfile.txt will contain all debug output unless you set debug-log to false, in which case it will only contain output at the log level you specified (notice by default).

There is also the -m option on Linux which will spawn a command of your choice and pipe the output directly to that command.

The WorkTime details 'debug' option adds details on the end of each line displayed for Accepted or Rejected work done. An example would be:

<-00000059.ed4834a3 M:X D:1.0 G:17:02:38:0.405 C:1.855 (2.995) W:3.440 (0.000) S:0.461 R:17:02:47

The first 2 hex codes are the previous block hash, the rest are reported in seconds unless stated otherwise: The previous hash is followed by the getwork mode used M:X where X is one of P:Pool, T:Test Pool, L:LP or B:Benchmark, then D:d.ddd is the difficulty required to get a share from the work, then G:hh:mm:ss:n.nnn, which is when the getwork or LP was sent to the pool and the n.nnn is how long it took to reply, followed by 'O' on it's own if it is an original getwork, or 'C:n.nnn' if it was a clone with n.nnn stating how long after the work was recieved that it was cloned, (m.mmm) is how long from when the original work was received until work started, W:n.nnn is how long the work took to process until it was ready to submit, (m.mmm) is how long from ready to submit to actually doing the submit, this is usually 0.000 unless there was a problem with submitting the work, S:n.nnn is how long it took to submit the completed work and await the reply, R:hh:mm:ss is the actual time the work submit reply was received

If you start sgminer with the --sharelog option, you can get detailed information for each share found. The argument to the option may be "-" for standard output (not advisable with the ncurses UI), any valid positive number for that file descriptor, or a filename.

To log share data to a file named "share.log", you can use either: ./sgminer --sharelog 50 -o xxx -u yyy -p zzz 50>share.log ./sgminer --sharelog share.log -o xxx -u yyy -p zzz

For every share found, data will be logged in a CSV (Comma Separated Value) format: timestamp,disposition,target,pool,dev,thr,sharehash,sharedata For example (this is wrapped, but it's all on one line for real): 1335313090,reject, ffffffffffffffffffffffffffffffffffffffffffffffffffffffff00000000, http://localhost:8337,GPU0,0, 6f983c918f3299b58febf95ec4d0c7094ed634bc13754553ec34fc3800000000, 00000001a0980aff4ce4a96d53f4b89a2d5f0e765c978640fe24372a000001c5 000000004a4366808f81d44f26df3d69d7dc4b3473385930462d9ab707b50498 f681634a4f1f63d01a0cd43fb338000000000080000000000000000000000000 0000000000000000000000000000000000000000000000000000000080020000