Bitcoin Forum
November 04, 2024, 01:19:05 AM *
News: Latest Bitcoin Core release: 28.0 [Torrent]
 
   Home   Help Search Login Register More  
Pages: « 1 ... 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 [60] 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 ... 1240 »
  Print  
Author Topic: CCminer(SP-MOD) Modded GPU kernels.  (Read 2347563 times)
sp_ (OP)
Legendary
*
Offline Offline

Activity: 2954
Merit: 1087

Team Black developer


View Profile
January 20, 2015, 08:08:30 AM
 #1181

I have solved a bug wich causes the cpu to use alot of cpu. Now x11 is hashing with 0-1% cpu usage. Fixed in the unpublished spreadcoin fork for now, but will submitt to github later. Less CPU-> less energy and more stable hashrates.
A noticable difference when using the computer while mining.
Hi sp_.
We need the spreadx11 opt. miner asap, dif. is skyrocketing...

So is the price. 600% in one month. This coin is gold Smiley

I will test on my 970/980 rig on friday/saturday and release the first beta to the donators on sunday.

Team Black Miner (ETHB3 ETH ETC VTC KAWPOW FIROPOW EVRPROGPOW MEOWPOW + dual mining + tripple mining.. https://github.com/sp-hash/TeamBlackMiner
StuffOfInterest
Sr. Member
****
Offline Offline

Activity: 401
Merit: 250


View Profile
January 20, 2015, 11:09:22 AM
 #1182

I have solved a bug wich causes the cpu to use alot of cpu. Now x11 is hashing with 0-1% cpu usage. Fixed in the unpublished spreadcoin fork for now, but will submitt to github later. Less CPU-> less energy and more stable hashrates.
A noticable difference when using the computer while mining.

That will be most appreciated.  I had definitely noticed slugishness issues when my 2x750Ti would switch from Quark to X11.  Fortunately, it has been staying in Quark most of the time lately.

BExR exchange rates on your phone's home screen.
Miner Control to get auto algorithm switching for multiple mining services. (please donate if you like)
Could Proof of Blockchain (PoBC) help secure a coin and avoid runaway ASIC mining?
antonio8
Legendary
*
Offline Offline

Activity: 1400
Merit: 1000


View Profile
January 20, 2015, 03:19:32 PM
 #1183

@ sp_

Is qubit algo broken after release 24?

I get an error message "stratum extranonce id was not correct" and all "booo's" 

If you are going to leave your BTC on an exchange please send it to this address instead 1GH3ub3UUHbU5qDJW5u3E9jZ96ZEmzaXtG, I will at least use the money better than someone who steals it from the exchange. Thanks Wink
sp_ (OP)
Legendary
*
Offline Offline

Activity: 2954
Merit: 1087

Team Black developer


View Profile
January 20, 2015, 08:21:31 PM
 #1184

Submitted a fix on the cpu usage. x11 is now using 0-1% cpu on a slow celeron. Quark is up another 50KHASH 750ti. !



Team Black Miner (ETHB3 ETH ETC VTC KAWPOW FIROPOW EVRPROGPOW MEOWPOW + dual mining + tripple mining.. https://github.com/sp-hash/TeamBlackMiner
sp_ (OP)
Legendary
*
Offline Offline

Activity: 2954
Merit: 1087

Team Black developer


View Profile
January 20, 2015, 08:27:17 PM
Last edit: January 20, 2015, 09:18:36 PM by sp_
 #1185

I tried v31, my fast rig is again stable, my slow rig crashes after 100-200 accepts.  I lowered the intensity to 18 (and lost 1mh/s performance!) and it was the same thing... crash.   I run stock clocks, period.
Both rigs are headless.  I plugged into the slow rig and booted normally, and noticed some error messages about nvidia driver 346.35.  It was released on 17 JAN 15, yesterday.  I think that I will have to purge the drivers and then re-install.  For Linux, v346.35 is a long-term release, no longer a beta, as 346.22 was.  I may burn the whole OS and move to 14.10, which is now supported by 346.35.      --scryptr

Please try a new build from github, the cpu usage is gone, and I have added more cuda cleanupcode. power should be down abit. cpu usage on x11, x13 was 10-20% in release 31 but reduced to 0-1% in the github version

Team Black Miner (ETHB3 ETH ETC VTC KAWPOW FIROPOW EVRPROGPOW MEOWPOW + dual mining + tripple mining.. https://github.com/sp-hash/TeamBlackMiner
scryptr
Legendary
*
Offline Offline

Activity: 1797
Merit: 1028



View Profile WWW
January 20, 2015, 10:53:35 PM
 #1186

I tried v31, my fast rig is again stable, my slow rig crashes after 100-200 accepts.  I lowered the intensity to 18 (and lost 1mh/s performance!) and it was the same thing... crash.   I run stock clocks, period.
Both rigs are headless.  I plugged into the slow rig and booted normally, and noticed some error messages about nvidia driver 346.35.  It was released on 17 JAN 15, yesterday.  I think that I will have to purge the drivers and then re-install.  For Linux, v346.35 is a long-term release, no longer a beta, as 346.22 was.  I may burn the whole OS and move to 14.10, which is now supported by 346.35.      --scryptr

Please try a new build from github, the cpu usage is gone, and I have added more cuda cleanupcode. power should be down abit. cpu usage on x11, x13 was 10-20% in release 31 but reduced to 0-1% in the github version

v31-git BUILD --

I just got home and was about to burn 14.10 when I read this.  I tried to compile the code, but I suspect a TYPO:

{QUOTE/}

mv -f .deps/ccminer-crc32.Tpo .deps/ccminer-crc32.Po
g++ -DHAVE_CONFIG_H -I.  -fopenmp  -pthread -fno-strict-aliasing -I./compat/jans                                                                                                                                                             son -I/usr/local/cuda/include -DUSE_WRAPNVML -DSCRYPT_KECCAK512 -DSCRYPT_CHACHA                                                                                                                                                              -DSCRYPT_CHOOSE_COMPILETIME   -O3 -march=native -D_REENTRANT -falign-functions=1                                                                                                                                                             6 -falign-jumps=16 -falign-labels=16 -MT ccminer-util.o -MD -MP -MF .deps/ccmine                                                                                                                                                             r-util.Tpo -c -o ccminer-util.o `test -f 'util.cpp' || echo './'`util.cpp
ccminer.cpp:217:22: error: conflicting declaration ‘char* device_name [16]’
 char * device_name[16];
                      ^
In file included from ccminer.cpp:44:0:
miner.h:488:14: error: ‘device_name’ has a previous declaration as ‘char* device                                                                                                                                                             _name [8]’
 extern char* device_name[8];
              ^
ccminer.cpp:218:18: error: conflicting declaration ‘int device_map [16]’
 int device_map[16] = { 0, 1, 2, 3, 4, 5, 6, 7,8,9,10,11,12,13,14,15 };
                  ^
In file included from ccminer.cpp:44:0:
miner.h:489:14: error: ‘device_map’ has a previous declaration as ‘short int dev                                                                                                                                                             ice_map [8]’
 extern short device_map[8];
              ^
ccminer.cpp:219:19: error: conflicting declaration ‘long int device_sm [16]’
 long  device_sm[16] = { 0 };
                   ^
In file included from ccminer.cpp:44:0:
miner.h:490:14: error: ‘device_sm’ has a previous declaration as ‘long int devic                                                                                                                                                             e_sm [8]’
 extern long  device_sm[8];
              ^
make[2]: *** [ccminer-ccminer.o] Error 1
make[2]: *** Waiting for unfinished jobs....
mv -f .deps/ccminer-hefty1.Tpo .deps/ccminer-hefty1.Po
mv -f .deps/ccminer-scrypt.Tpo .deps/ccminer-scrypt.Po
mv -f .deps/ccminer-util.Tpo .deps/ccminer-util.Po
make[2]: Leaving directory `/home/borqd/spfresh'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/borqd/spfresh'
make: *** [all] Error 2
{/QUOTE}

Sorry for the fuss.  I hope that you can fix it, I really want to try the executable.       --scryptr

SCRYPTR'S NOTEBOOK: https://bitcointalk.org/index.php?topic=5035515.msg46035530#msg46035530
GITHUB: "github.com/scryptr"  MERIT is appreciated, also.  Thanks!
flipclip
Member
**
Offline Offline

Activity: 111
Merit: 10


View Profile
January 20, 2015, 10:56:27 PM
 #1187

Edit miner.h:

From (start line 488):
extern char* device_name[8];
extern short device_map[8];
extern long device_sm[8];

To:
extern char* device_name[16];
extern int device_map[16];
extern long device_sm[16];

Made the edit and just compiled on my Ubuntu 14.04 system.  Also submitted a pull to git.
tbearhere
Legendary
*
Offline Offline

Activity: 3206
Merit: 1003



View Profile
January 20, 2015, 11:02:49 PM
 #1188


Don't forget to get these usb hold down clips...they work great...no more loose cables.  Wink
http://www.miningcryptocurrency.com.au/store/p4/RiserSafe_By_the_Inventor_Patent_%28Pending%29.html
that looks interesting...
They work great.  Wink
scryptr
Legendary
*
Offline Offline

Activity: 1797
Merit: 1028



View Profile WWW
January 20, 2015, 11:48:14 PM
Last edit: January 21, 2015, 12:23:58 AM by scryptr
 #1189

Edit miner.h:

From (start line 488):
extern char* device_name[8];
extern short device_map[8];
extern long device_sm[8];

To:
extern char* device_name[16];
extern int device_map[16];
extern long device_sm[16];

Made the edit and just compiled on my Ubuntu 14.04 system.  Also submitted a pull to git.

v31b TYPO-

THAT IS THE FIX.  It compiled and has just started running.  Initial impression:  Loss of 500kh/s for the rig from v31a at its fastest.  If this version runs smoothly and remains stable, I don't care.  Stability is more important.  I am right on 75 accepts and 100%  error-free.      --scryptr

EDIT:

It ran for a couple hundred accepts, and then dropped to double-digit kh/s.  A couple cards would resurge to 6mh/s, and then drop immediately.  When is sent "^c", I got a "SIGINT received" response, but the program would not close.  I have seen this pattern before in previous builds.   

I am going to save the config files and important data, and move to 14.10, as the 146.35 drivers now support Utopic Unicorn.  I may have junk in my OS that is a factor.       --scryptr

SCRYPTR'S NOTEBOOK: https://bitcointalk.org/index.php?topic=5035515.msg46035530#msg46035530
GITHUB: "github.com/scryptr"  MERIT is appreciated, also.  Thanks!
flipclip
Member
**
Offline Offline

Activity: 111
Merit: 10


View Profile
January 21, 2015, 04:23:30 AM
Last edit: January 21, 2015, 12:38:56 PM by flipclip
 #1190

From release 31 to the 1/20/2015 commit, for quark I'm seeing a drop of 25 kH/s gain of 50 kH/s per card (750Ti), and a large drop in CPU usage (from ~16% to ~1%).  Great stuff.
sp_ (OP)
Legendary
*
Offline Offline

Activity: 2954
Merit: 1087

Team Black developer


View Profile
January 21, 2015, 10:43:51 AM
 #1191

From release 31 to the 1/20/2015 commit, for quark I'm seeing a drop of about 25 kH/s per card (750Ti), but also a large drop in CPU usage (from ~16% to ~1%).

Yes, The miner starts well at first, but then the hashrate is dropping over time. I reverted the sync code. Now quark is using 3-5% cpu, but the hash is more stable.(hopefully)

Team Black Miner (ETHB3 ETH ETC VTC KAWPOW FIROPOW EVRPROGPOW MEOWPOW + dual mining + tripple mining.. https://github.com/sp-hash/TeamBlackMiner
sp_ (OP)
Legendary
*
Offline Offline

Activity: 2954
Merit: 1087

Team Black developer


View Profile
January 21, 2015, 10:46:55 AM
 #1192

Edit miner.h:
From (start line 488):
extern char* device_name[8];
extern short device_map[8];
extern long device_sm[8];
To:
extern char* device_name[16];
extern int device_map[16];
extern long device_sm[16];
Made the edit and just compiled on my Ubuntu 14.04 system.  Also submitted a pull to git.

Yeah, forgot to checkin one file. CCminer is now supporting 16 cards per rig. Smiley

Team Black Miner (ETHB3 ETH ETC VTC KAWPOW FIROPOW EVRPROGPOW MEOWPOW + dual mining + tripple mining.. https://github.com/sp-hash/TeamBlackMiner
Grim
Sr. Member
****
Offline Offline

Activity: 506
Merit: 252


View Profile
January 21, 2015, 11:20:58 AM
 #1193

Edit miner.h:
From (start line 488):
extern char* device_name[8];
extern short device_map[8];
extern long device_sm[8];
To:
extern char* device_name[16];
extern int device_map[16];
extern long device_sm[16];
Made the edit and just compiled on my Ubuntu 14.04 system.  Also submitted a pull to git.

Yeah, forgot to checkin one file. CCminer is now supporting 16 cards per rig. Smiley

nice...

I can't get more than 6 working in win 8.1 (the same mobo had 7 amd cards running)
sp_ (OP)
Legendary
*
Offline Offline

Activity: 2954
Merit: 1087

Team Black developer


View Profile
January 21, 2015, 11:23:12 AM
 #1194

nice...
I can't get more than 6 working in win 8.1 (the same mobo had 7 amd cards running)

You probobly need a modded driver. I usually have 2-3 amd cards, and fill the rest of the slots with nvidia cards. Good for the powersupply as well.

Team Black Miner (ETHB3 ETH ETC VTC KAWPOW FIROPOW EVRPROGPOW MEOWPOW + dual mining + tripple mining.. https://github.com/sp-hash/TeamBlackMiner
Grim
Sr. Member
****
Offline Offline

Activity: 506
Merit: 252


View Profile
January 21, 2015, 11:27:45 AM
 #1195

nice...
I can't get more than 6 working in win 8.1 (the same mobo had 7 amd cards running)

You probobly need a modded driver. I usually have 2-3 amd cards, and fill the rest of the slots with nvidia cards. Good for the powersupply as well.

ya, only thing that amd is left good at ...

sp_ (OP)
Legendary
*
Offline Offline

Activity: 2954
Merit: 1087

Team Black developer


View Profile
January 21, 2015, 11:57:35 AM
 #1196

nice...
I can't get more than 6 working in win 8.1 (the same mobo had 7 amd cards running)
You probobly need a modded driver. I usually have 2-3 amd cards, and fill the rest of the slots with nvidia cards. Good for the powersupply as well.
ya, only thing that amd is left good at ...

This looks interesting. 16 GPU's on one motherboard using 4 PCI-e slots.

http://amfeltec.com/products/gpu-oriented-cluster/

But Might have to modify the driver...

EDIT: just checked the price.. £140 for pciE splitter 1->4 slots. Bether to buy another motherboard. Not worth it.

Team Black Miner (ETHB3 ETH ETC VTC KAWPOW FIROPOW EVRPROGPOW MEOWPOW + dual mining + tripple mining.. https://github.com/sp-hash/TeamBlackMiner
chrysophylax
Legendary
*
Offline Offline

Activity: 2898
Merit: 1091


--- ChainWorks Industries ---


View Profile WWW
January 21, 2015, 12:34:03 PM
 #1197

nice...
I can't get more than 6 working in win 8.1 (the same mobo had 7 amd cards running)
You probobly need a modded driver. I usually have 2-3 amd cards, and fill the rest of the slots with nvidia cards. Good for the powersupply as well.
ya, only thing that amd is left good at ...

This looks interesting. 16 GPU's on one motherboard using 4 PCI-e slots.

http://amfeltec.com/products/gpu-oriented-cluster/

But Might have to modify the driver...

EDIT: just checked the price.. £140 for pciE splitter 1->4 slots. Bether to buy another motherboard. Not worth it.

sp - i checked these out a while ago - even contacted the people ...

this will work ONLY if the bios of the motherboard allows it to ...

most standard bios only allows ( its own memory limitation ) 8 gpu's per motherboard ...

this limitation is only compensated when you have motherboards with enhanced bios - like some of the supermicro motherboards - that are designed to have more than 8 gpus via a bios augmentation ...

if you are familiar with reprogramming bios - then you will be able to recompile the kernels of the operating systems alongside the a hacked / reprogrammed bios ...

but for now - if you are using standard motherboards that are the 'norm' these days - then the limitation usually always ends up being 8 gpus ...

you can definitely add more ( i have had 12 gpus on the one system in linux ) and the os ( in my case ) could see ALL the gpus - but could only use 8 at a time due to this limitation ...

if there was no limitation - i would be running them right this minute Wink

btw - the supermicro dual xeon boards that can handle 12gpus are all around $600-700AUD ... not worth the effort for mining unless you have the cash for it as a project ...

#crysx

flipclip
Member
**
Offline Offline

Activity: 111
Merit: 10


View Profile
January 21, 2015, 12:41:19 PM
 #1198

Quark:
v28= ~11,160 kh/s
v30= ~11,441 kh/s
v31b= ~11,650 kh/s (and decrease in CPU usage)
x11:
v28= ~5,730 kh/s
v30= ~5,850 kh/s
lyra2:
v28= ~1,350 kh/s
v30= ~1,350 kh/s
(2 750Ti's, no overclock)
bathrobehero
Legendary
*
Offline Offline

Activity: 2002
Merit: 1051


ICO? Not even once.


View Profile
January 21, 2015, 01:13:29 PM
 #1199

I have solved a bug wich causes the cpu to use alot of cpu. Now x11 is hashing with 0-1% cpu usage. Fixed in the unpublished spreadcoin fork for now, but will submitt to github later. Less CPU-> less energy and more stable hashrates.
A noticable difference when using the computer while mining.
Hi sp_.
We need the spreadx11 opt. miner asap, dif. is skyrocketing...

So is the price. 600% in one month. This coin is gold Smiley

I will test on my 970/980 rig on friday/saturday and release the first beta to the donators on sunday.

Can I ask how long do you intend to keep the miner exclusive to donators?

Not your keys, not your coins!
jpouza
Legendary
*
Offline Offline

Activity: 2856
Merit: 1122


View Profile
January 21, 2015, 01:22:59 PM
 #1200

I have solved a bug wich causes the cpu to use alot of cpu. Now x11 is hashing with 0-1% cpu usage. Fixed in the unpublished spreadcoin fork for now, but will submitt to github later. Less CPU-> less energy and more stable hashrates.
A noticable difference when using the computer while mining.
Hi sp_.
We need the spreadx11 opt. miner asap, dif. is skyrocketing...

So is the price. 600% in one month. This coin is gold Smiley

I will test on my 970/980 rig on friday/saturday and release the first beta to the donators on sunday.

Can I ask how long do you intend to keep the miner exclusive to donators?

I think enough time to do not hurt the SPR coin as he stated before.
Pages: « 1 ... 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 [60] 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 ... 1240 »
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!