CCminer(SP-MOD) Modded GPU kernels.

[pallas]

its from me except you changed the code...

in mine i only use :
ccminer.cpp:1669:               if (have_stratum && work.data[0] == 0 && !opt_benchmark) {

loopcnt > 0 do the reverse of what you want

I don't know about loopcount, it was already there, I just added my new flag.
I've removed it, testing now...

EDIT: works fine, looks like this:
if (have_stratum && (work.data[0] == 0 || network_fail_flag))

[djm34]

its from me except you changed the code...

in mine i only use :
ccminer.cpp:1669:               if (have_stratum && work.data[0] == 0 && !opt_benchmark) {

loopcnt > 0 do the reverse of what you want

I don't know about loopcount, it was already there, I just added my new flag.
I've removed it, testing now...

EDIT: works fine, looks like this:
if (have_stratum && (work.data[0] == 0 || network_fail_flag))

the whole ccminer.cpp needs some cleaning no matter what (not mentioning that sp and tpruvot versions aren't using the same)

[fenomenhaa]

@sp  still waiting for your miner 750ti:7.05 mh/s thanks for spreadminer(didn't test yet,bioscoin goes good)  )) And as i said before i'm ready for donations  for DJM34 changes )) 

[pallas]

today I got this two times while mining quark:

11443 Floating point exception(core dumped)

[sp_]

submitted another speedup (quark/x11) removed 0.5% of the instructions in bmw512

[sp_]

I had a look at the code; here is a quick fix to avoid mining when network fails:

diff --git a/ccminer.cpp b/ccminer.cpp
index c13d617..6c11757 100644
--- a/ccminer.cpp
+++ b/ccminer.cpp
@@ -215,6 +215,7 @@ bool autotune = true;
 bool opt_autotune = true;
 
 bool abort_flag = false;
+bool network_fail_flag = false;
 char *jane_params = NULL;
 
 char *rpc_user = NULL;
@@ -1356,7 +1357,7 @@ static void *miner_thread(void *userdata)
                pthread_mutex_unlock(&g_work_lock);
 
                /* prevent gpu scans before a job is received */
-               if (have_stratum && loopcnt>0 && work.data[0] == 0)
+               if (have_stratum && ((loopcnt>0 && work.data[0] == 0) || network_fail_flag))
                {
                        sleep(1);
                        continue;       
@@ -1903,6 +1904,7 @@ static void *stratum_thread(void *userdata)
                                if (!opt_benchmark)
                                        applog(LOG_ERR, "...retry after %d seconds", opt_fail_pause);
                                sleep(opt_fail_pause);
+                               network_fail_flag = true;
                        }
                }
 
@@ -1913,6 +1915,7 @@ static void *stratum_thread(void *userdata)
                        g_work_time = time(NULL);
                        if (stratum.job.clean)
                        {
+                               network_fail_flag = false;
                                if (!opt_quiet)
                                        applog(LOG_BLUE, "%s %s block %d", short_url, algo_names[opt_algo],
                                                stratum.job.height);

I tested it by creating a REJECT iptable rule to the pool and it worked fine, even though it submitted a lot of expired jobs on resume (maybe the jobs should be cleaned, but it doesn't hurt....).
On a side note, I think the cpu-miner.c file is unused (maybe others as well?).

Thanks. I have added it now.

[bensam1231]

So I've been browsing around graphics cards, it's become apparent some manufacturers will use more power for the same clocks of the same make/model of card as another manufacturer.

Say a Asus 970 clocked at the same speed as a Gigabyte 970, the Gigabyte will use more power.

Is there anyway to figure out what card uses more power without actually testing it yourself? I noticed the power connectors are not the same across all cards in the same make and model. However, I have noticed that this isn't always a direct correlation to the power draw either. This can vary wildly too.

For instance some of the 960s have a six pin, others have a eight pin, and still others have two six pins. I don't know if this means a double six pin will pull infinitely more power then a single six pin, but it would hint at it. It could be the single six pins just don't have the right amount of power connectors for the maximum draw as well.

[djm34]

So I've been browsing around graphics cards, it's become apparent some manufacturers will use more power for the same clocks of the same make/model of card as another manufacturer.

Say a Asus 970 clocked at the same speed as a Gigabyte 970, the Gigabyte will use more power.

Is there anyway to figure out what card uses more power without actually testing it yourself? I noticed the power connectors are not the same across all cards in the same make and model. However, I have noticed that this isn't always a direct correlation to the power draw either. This can vary wildly too.

For instance some of the 960s have a six pin, others have a eight pin, and still others have two six pins. I don't know if this means a double six pin will pull infinitely more power then a single six pin, but it would hint at it. It could be the single six pins just don't have the right amount of power connectors for the maximum draw as well.

any cards which is superclocked or factory OC will obviously use more power.
If you want to compare check on nvidia website the original design and compare to what propose the manufacturer (and check on the manufacturer as many online just pick the first spec tech they find.

but again, it is probably easier to downclock yourself over a superclocked card than doing the inverse as the bios has been properly tune, so I still think it is better to buy a card designed to use more electricity than the inverse (wouldn't buy a Kingpin edition though...)

[bensam1231]

So I've been browsing around graphics cards, it's become apparent some manufacturers will use more power for the same clocks of the same make/model of card as another manufacturer.

Say a Asus 970 clocked at the same speed as a Gigabyte 970, the Gigabyte will use more power.

Is there anyway to figure out what card uses more power without actually testing it yourself? I noticed the power connectors are not the same across all cards in the same make and model. However, I have noticed that this isn't always a direct correlation to the power draw either. This can vary wildly too.

For instance some of the 960s have a six pin, others have a eight pin, and still others have two six pins. I don't know if this means a double six pin will pull infinitely more power then a single six pin, but it would hint at it. It could be the single six pins just don't have the right amount of power connectors for the maximum draw as well.

any cards which is superclocked or factory OC will obviously use more power.
If you want to compare check on nvidia website the original design and compare to what propose the manufacturer (and check on the manufacturer as many online just pick the first spec tech they find.

but again, it is probably easier to downclock yourself over a superclocked card than doing the inverse as the bios has been properly tune, so I still think it is better to buy a card designed to use more electricity than the inverse (wouldn't buy a Kingpin edition though...)

I'm not looking at overclocking potential, rather efficiency right off the bat. I don't really see any sort of indications of power usage looking at the specs besides 'factory' power usage, not necessarily what the bios is setup for.

[bathrobehero]

So I've been browsing around graphics cards, it's become apparent some manufacturers will use more power for the same clocks of the same make/model of card as another manufacturer.

Say a Asus 970 clocked at the same speed as a Gigabyte 970, the Gigabyte will use more power.

Is there anyway to figure out what card uses more power without actually testing it yourself? I noticed the power connectors are not the same across all cards in the same make and model. However, I have noticed that this isn't always a direct correlation to the power draw either. This can vary wildly too.

For instance some of the 960s have a six pin, others have a eight pin, and still others have two six pins. I don't know if this means a double six pin will pull infinitely more power then a single six pin, but it would hint at it. It could be the single six pins just don't have the right amount of power connectors for the maximum draw as well.

Look into the BIOS of the card from here or do your own tests to see efficiency.

Other than that, PCIE-E 1X can output 10 watts, PCI-E16x can output 75W, 6-pin can do 75W and 8-pin can do 150W.

[michaelbippom]

Hello!  I am currently using the OSX builds of ccminer 1.2 from John Chapman and its working great.  However I want to use this fork for the better hashrate.  I am hoping someone here can help me compile the latest sp for of ccminer on OSX 10.10.  I am getting 2 errors when I do "sudo make". 

First let me explain what process I did:

1.) Install Xcode 6.4, command line tools, homebrew
2.) update homebrew and install dependencies: coreutils autoconf automake jansson libgcrypt libgpg-error libtool libusb pkg-config yasm curl
3.) git clone https://github.com/sp-hash/ccminer.git
4.) ./autogen.sh
5.) ./configure
6.) sudo make

This results in this:

ccminer.cpp:464:26: error: use of undeclared identifier 'GetCurrentProcess'
                SetProcessAffinityMask(GetCurrentProcess(), mask);
                                       ^
ccminer.cpp:466:25: error: use of undeclared identifier 'GetCurrentThread'
                SetThreadAffinityMask(GetCurrentThread(), mask);
                                      ^

I cannot for the life of me get this fixed.  I tried commenting out those lines of code in the source because it looked like it pertained only to windows, but that just results in more "undeclared identifier" errors for other lines of code.

I AM WILLING TO COMPENSATE SOMEONE SOME BTC IF THEY CAN HELP ME GET THIS BUILT AND WORKING!
Any help is greatly appreciated.

Here is the entire output of the "sudo make" command

/Applications/Xcode.app/Contents/Developer/usr/bin/make  all-recursive
Making all in compat
make[3]: Nothing to be done for `all-am'.
gcc -DHAVE_CONFIG_H -I.  -I/usr/local/include   -pthread -fno-strict-aliasing  -I/usr/local/cuda/include  -DSCRYPT_KECCAK512 -DSCRYPT_CHACHA -DSCRYPT_CHOOSE_COMPILETIME   -g -O2 -Wall -MT ccminer-crc32.o -MD -MP -MF .deps/ccminer-crc32.Tpo -c -o ccminer-crc32.o `test -f 'crc32.c' || echo './'`crc32.c
mv -f .deps/ccminer-crc32.Tpo .deps/ccminer-crc32.Po
gcc -DHAVE_CONFIG_H -I.  -I/usr/local/include   -pthread -fno-strict-aliasing  -I/usr/local/cuda/include  -DSCRYPT_KECCAK512 -DSCRYPT_CHACHA -DSCRYPT_CHOOSE_COMPILETIME   -g -O2 -Wall -MT ccminer-hefty1.o -MD -MP -MF .deps/ccminer-hefty1.Tpo -c -o ccminer-hefty1.o `test -f 'hefty1.c' || echo './'`hefty1.c
mv -f .deps/ccminer-hefty1.Tpo .deps/ccminer-hefty1.Po
g++ -DHAVE_CONFIG_H -I.  -I/usr/local/include   -pthread -fno-strict-aliasing  -I/usr/local/cuda/include  -DSCRYPT_KECCAK512 -DSCRYPT_CHACHA -DSCRYPT_CHOOSE_COMPILETIME   -g -O2 -MT ccminer-ccminer.o -MD -MP -MF .deps/ccminer-ccminer.Tpo -c -o ccminer-ccminer.o `test -f 'ccminer.cpp' || echo './'`ccminer.cpp
ccminer.cpp:464:26: error: use of undeclared identifier 'GetCurrentProcess'
                SetProcessAffinityMask(GetCurrentProcess(), mask);
                                       ^
ccminer.cpp:466:25: error: use of undeclared identifier 'GetCurrentThread'
                SetThreadAffinityMask(GetCurrentThread(), mask);
                                      ^
ccminer.cpp:1105:4: warning: 'SHA256' is deprecated: first deprecated in OS X
      10.7 [-Wdeprecated-declarations]
                        SHA256((uchar*)sctx->job.coinbase, sctx->job.coi...
                        ^
/usr/include/openssl/sha.h:150:16: note: 'SHA256' has been explicitly marked
      deprecated here
unsigned char *SHA256(const unsigned char *d, size_t n,unsigned char *md...
               ^
ccminer.cpp:1121:74: warning: for loop has empty body [-Wempty-body]
  ...(i = 0; i < (int)sctx->xnonce2_size && !++sctx->job.xnonce2; i++);
                                                                         ^
ccminer.cpp:1121:74: note: put the semicolon on a separate line to silence this
      warning
ccminer.cpp:1360:21: warning: '&&' within '||' [-Wlogical-op-parentheses]
                if ((have_stratum && work.data[0] == 0 || network_fail_f...
                     ~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~ ~~
ccminer.cpp:1360:21: note: place parentheses around the '&&' expression to
      silence this warning
                if ((have_stratum && work.data[0] == 0 || network_fail_f...
                                  ^
                     (                                )
ccminer.cpp:1377:12: warning: 18 enumeration values not handled in switch:
      'ALGO_ANIME', 'ALGO_BITC', 'ALGO_DEEP'... [-Wswitch]
                        switch (opt_algo) {
                                ^
ccminer.cpp:1634:12: warning: 36 enumeration values not handled in switch:
      'ALGO_ANIME', 'ALGO_BITC', 'ALGO_BITCOIN'... [-Wswitch]
                        switch (opt_algo) {
                                ^
ccminer.cpp:2011:12: warning: 36 enumeration values not handled in switch:
      'ALGO_ANIME', 'ALGO_BITC', 'ALGO_BITCOIN'... [-Wswitch]
                        switch (opt_algo)
                                ^
6 warnings and 2 errors generated.
make[2]: *** [ccminer-ccminer.o] Error 1

[Fuzzbawls]

process/thread affinity doesn't really exist in OS X so thats why the error is being tossed.

i've been poking around with it too but haven't had much time to really dig at it.

[michaelbippom]

process/thread affinity doesn't really exist in OS X so thats why the error is being tossed.

i've been poking around with it too but haven't had much time to really dig at it.

Is that a vital part of the code?  Sounds like it may be a lot of trouble to get it working.  From the code it was flagged as "for windows" in an if statement.  So perhaps I need to add flags to the ./confugre to specify darwin enviornment?

[chrysophylax]

It is profitable when few are mining it. The windows exe is included, as well as the sourcecode.

Most coins are more profitable than the multipools/rentalsites if you mine direcly. Being able to switch between algos, will make you make a profit in the long run.

Spreadcoin can only be solomined.

spreadcoin can be pool mined too sp ...

code just needs to be changed at the pool end ...

i was using your spreadminer ( v8 ) to mine on a private pool for months - which is why i gave you the headache trying to compile it under linux the last time if you recall ...

there is also a spreadcoin pool that ocminer created - but only with his miner - NOT with yours ( which is no where near as optimized as your miner ) ... believe me ive tried ...

its one of the arguments i have put forward to the spr community - and well - its sorta just being heard about 'maybe' looking at a pool / multimine scenario ... maybe ...

im all for it if you wanted to know - so i support spreadx11 all the way ...

#crysx

[Fuzzbawls]

process/thread affinity doesn't really exist in OS X so thats why the error is being tossed.

i've been poking around with it too but haven't had much time to really dig at it.

Is that a vital part of the code?  Sounds like it may be a lot of trouble to get it working.  From the code it was flagged as "for windows" in an if statement.  So perhaps I need to add flags to the ./confugre to specify darwin enviornment?

afaik, process/thread affinity didn't exist at all in the early versions, so unless there are optimizations that specifically REQUIRE it, it should theoretically be ok to remove/disable it for OS X

[chrysophylax]

Release 56 has finally made be break 100 mh/s on quark. Virtual beers for everyone!

Aliman,
Your 980ti is beating the Titan-X  

The titans in this picture are to close to eachother. They will heat up and trottle.
Bether to use 70cm usb risers and spread the cards.
The titans are not good for mining. 12GB of ram doesn't help, and the prise is skyhigh.

Wrong about the heat; servers tend to be cooled differently. Right about everything else.

agreed ...

ours cool much more efficiently in a closed environment now than an open one - with the inital prototype closed design we have currently setup ... we already have server cases that do the same thing - but at m,any thousands of dollars each ( without components ) for high grade server cases and redundant psu - its worth the development of a cheap alternative design ...

this frame is taking a LONG time to get right - but will be a very cheap alternative to those servers cases - with a similar cooling system ...

solid aluminium tube frame - and hardened perspex - with high powered fans for tunnelled front-to-back air throughput ... thats our aim ...

anyway - back to work ...

edit - just to make sure if anyone asks ... yes - we have a 5ru case ( and 3ru - though its too tight for the cards which looks like a 4ru case may be the better way to go ) being tested for efficiency and airflow as opposed to 'our' designed frame and working systems ...

so far - a 5ru with usb 3.0 cabled risers and modified railing system may be the best way to go if going with the 'rady made' cases - or the in house designed frame we are getting together ... guess which one we are opting for ...

so we actually can say that cooling works in practice rather than just theory ... still working on the finer details of the frame and adding a few more adjustments ...

the rebuild of the farm WILL comprise of the finished frame design ...

#crysx

[Epsylon3]

@michaelbippom

try to apply that : http://git.io/vY0Jg

[michaelbippom]

@michaelbippom

try to apply that : http://git.io/vY0Jg

Thanks!  Definitely getting on the right path here.  It compiled for a long time and gave a ton of warnings, but then finally halted on this error:

scrypt.cpp:33:10: fatal error: 'omp.h' file not found
#include <omp.h>
         ^
1 error generated.
make[2]: *** [ccminer-scrypt.o] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2

EDIT:

I found the file in /usr/ folder.  I just copied it to the root ccminer directory along with malloc.h and now its compiling.....fingers crossed.

BTW, I applied this code fix to the sp-hash fork.  When I tried to apply it to your fork, it gave a bunch of errors.


UPDATE:

Now it halts on this error:

clang: warning: argument unused during compilation: '-pthread'
ld: warning: directory not found for option '-L/usr/local/cuda/lib64'
ld: library not found for -lcudart
clang: error: linker command failed with exit code 1 (use -v to see invocation)
make[2]: *** [ccminer] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2

UPDATE 2

I used macports instead of homebrew on a fresh osx install and I was able to get gcc5 with openmp support enabled.  I also had to install the CUDA 6.5.27 toolkit for compute 5.2 support.  I did the code fix that Epsylon3 suggested and am now compiling.  Also had to do chmod 775 on install-sh for some reason (it was saying permission denied).  I am also noticing MUCH less warnings while compiling.  Fingers crossed....again!

UPDATE 3

SUCCESS!!!!!!!!!!  I cant beleive it, took me forever and alot of forum searching but I have it successfully compiled and hashing with my GTX 970 on OSX 10.10.  MASSIVE improvement over the ccminer 1.2 from John Chapman.  I was getting 4-4.5 MH X11 on that build and now I am getting 7-7.5 MH X11. 

Epsylon3, you got a little BTC coming your way for your help.  If anyone wants this build of ccminer sp-mod pm me and I'll help you out. 

[rednoW]

Thanks!  Definitely getting on the right path here.  It compiled for a long time and gave a ton of warnings, but then finally halted on this error:

scrypt.cpp:33:10: fatal error: 'omp.h' file not found
#include <omp.h>
         ^
1 error generated.
make[2]: *** [ccminer-scrypt.o] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2

I tried to locate that file but couldn't.  Can I just disable scrypt somehow?  I dont plan on using it.  I really only plan on using X11.

BTW, I applied this code fix to the sp-hash fork.  When I tried to apply it to your fork, it gave a bunch of errors.

omp.h is a header file for OpenMP threading technology. There are multiple ways to organize threading (multiple lightweight processes) in computer system: POSIX threads (pthreads), windows API, OpenMP etc.

You need a compiler with OpenMP support to compile this (or to turn that support on in config)

[michaelbippom]

Thanks!  Definitely getting on the right path here.  It compiled for a long time and gave a ton of warnings, but then finally halted on this error:

scrypt.cpp:33:10: fatal error: 'omp.h' file not found
#include <omp.h>
         ^
1 error generated.
make[2]: *** [ccminer-scrypt.o] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2

I tried to locate that file but couldn't.  Can I just disable scrypt somehow?  I dont plan on using it.  I really only plan on using X11.

BTW, I applied this code fix to the sp-hash fork.  When I tried to apply it to your fork, it gave a bunch of errors.

omp.h is a header file for OpenMP threading technology. There are multiple ways to organize threading (multiple lightweight processes) in computer system: POSIX threads (pthreads), windows API, OpenMP etc.

You need a compiler with OpenMP support to compile this (or to turn that support on in config)

Thanks!  I notice in the log when I run ./configure it says:

checking for gcc option to support OpenMP... unsupported

I installed the latest gcc-5 with homebrew via:  brew reinstall gcc --without-multilib

How do I get it to compile with OpenMP support?