Bitcoin Forum
December 05, 2016, 06:52:41 PM *
News: Latest stable version of Bitcoin Core: 0.13.1  [Torrent].
 
   Home   Help Search Donate Login Register  
Pages: « 1 2 3 4 5 6 7 8 9 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 »
  Print  
Author Topic: Algorithmically placed FPGA miner: 255MH/s/chip, supports all known boards  (Read 109510 times)
mining4fun11
Member
**
Offline Offline

Activity: 110


View Profile
November 23, 2012, 02:14:03 AM
 #861

I'm seeing about 13% rejects on my mmq.  Is this normal for this software. 
1480963961
Hero Member
*
Offline Offline

Posts: 1480963961

View Profile Personal Message (Offline)

Ignore
1480963961
Reply with quote  #2

1480963961
Report to moderator
1480963961
Hero Member
*
Offline Offline

Posts: 1480963961

View Profile Personal Message (Offline)

Ignore
1480963961
Reply with quote  #2

1480963961
Report to moderator
1480963961
Hero Member
*
Offline Offline

Posts: 1480963961

View Profile Personal Message (Offline)

Ignore
1480963961
Reply with quote  #2

1480963961
Report to moderator
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction. Advertise here.
eldentyrell
Donator
Legendary
*
Offline Offline

Activity: 966


felonious vagrancy, personified


View Profile WWW
November 23, 2012, 02:49:19 AM
 #862

I'm seeing about 13% rejects on my mmq.  Is this normal for this software. 

Make sure you're using 1.50a not 1.50.

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
mining4fun11
Member
**
Offline Offline

Activity: 110


View Profile
November 23, 2012, 02:51:04 AM
 #863

That is with 1.50a. 
mining4fun11
Member
**
Offline Offline

Activity: 110


View Profile
November 23, 2012, 12:21:37 PM
 #864

I'm seeing about 13% rejects on my mmq.  Is this normal for this software. 

Make sure you're using 1.50a not 1.50.

It seems to happen frequently when there is a block change.
eldentyrell
Donator
Legendary
*
Offline Offline

Activity: 966


felonious vagrancy, personified


View Profile WWW
November 23, 2012, 11:35:30 PM
 #865

It seems to happen frequently when there is a block change.

Thanks, that is useful information.

What pool are you using?  At the moment I only mine on Eclipse due to the fact that they have both X-Roll-Nonce and non-minimum-difficulty shares -- I need both of these for bandwidth reasons.  But if there is another pool that makes this problem come up more often and/or easier to debug, I can switch temporarily in order to investigate.

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
eldentyrell
Donator
Legendary
*
Offline Offline

Activity: 966


felonious vagrancy, personified


View Profile WWW
November 23, 2012, 11:41:00 PM
 #866

I have posted tml-1.50b which has highly experimental support for the ModMiner Quad's proprietary interface on Linux only:

  http://www.tricone-mining.com/tml/tml-1.50b.jar

There are no plans for Windows support.

Sample usage:


$ java -Dclock_pin=fgg484.K20 -Dclock_pin_freq=100 -jar tml-1.50b.jar modminer:/dev/ttyACM0 <pool-url>


I only have intermittent and remote access to a modminer board, so this is not as well-tested as I would like.  If you find bugs please post as much detail as possible.

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
mining4fun11
Member
**
Offline Offline

Activity: 110


View Profile
November 24, 2012, 12:28:07 AM
 #867

It seems to happen frequently when there is a block change.

Thanks, that is useful information.

What pool are you using?  At the moment I only mine on Eclipse due to the fact that they have both X-Roll-Nonce and non-minimum-difficulty shares -- I need both of these for bandwidth reasons.  But if there is another pool that makes this problem come up more often and/or easier to debug, I can switch temporarily in order to investigate.

I'm using ozcoin. 
eldentyrell
Donator
Legendary
*
Offline Offline

Activity: 966


felonious vagrancy, personified


View Profile WWW
November 24, 2012, 09:22:39 AM
 #868

TML 1.50c has been released, with experimental support for the x6500 proprietary USB interface on Linux only.  Big thanks to mining4fun11 for letting me use his machine to get this working.

  http://www.tricone-mining.com/tml/tml-1.50c.jar

Usage:


  sudo java -Dclock_pin=fgg484.K20 -jar tml-1.50c.jar x6500:AH123123 <pool-url>


Where you replace AH123123 with the serial number of your board's FTDI chip; to get a list of the serial numbers of the chips attached to your machine, type:


$ usb-devices | grep SerialNumber

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
rb2k
Member
**
Offline Offline

Activity: 109


View Profile
November 24, 2012, 11:44:19 AM
 #869

Woke up to this on the "b" version of the jar.
Restarting the process seems to have fixed it.



Quote
        restarting blocked threads for com.triconemining.signcryption.SCClient$Connection@4d4e3b5
[modminer:/dev/ttyACM0:0.0] nonce decryption aborted
               finished restarting blocked threads for com.triconemining.signcryption.SCClient$Connection@4d4e3b5
listener thread for limp.tricone-mining.com/23.23.249.133:7778 exiting due to java.io.IOException: you must perform CRYPTO_INIT before REQUEST_DECRYPT_NONCE
             opening signcryption connection to limp.tricone-mining.com/23.23.249.133:7778
[modminer:/dev/ttyACM0:0.1] decrypting nonce at address 0x000000c2
[modminer:/dev/ttyACM0:0.1]   encrypted nonce = 0x435f5a66
             restarting blocked threads for com.triconemining.signcryption.SCClient$Connection@629999a
[modminer:/dev/ttyACM0:0.1] nonce decryption aborted
               finished restarting blocked threads for com.triconemining.signcryption.SCClient$Connection@629999a
listener thread for limp.tricone-mining.com/23.23.249.133:7778 exiting due to java.io.IOException: you must perform CRYPTO_INIT before REQUEST_DECRYPT_NONCE
             opening signcryption connection to limp.tricone-mining.com/23.23.249.133:7778
[modminer:/dev/ttyACM0:0.2] decrypting nonce at address 0x00000032
[modminer:/dev/ttyACM0:0.2]   encrypted nonce = 0x6dd31044
             restarting blocked threads for com.triconemining.signcryption.SCClient$Connection@4830298
[modminer:/dev/ttyACM0:0.2] nonce decryption aborted
               finished restarting blocked threads for com.triconemining.signcryption.SCClient$Connection@4830298
listener thread for limp.tricone-mining.com/23.23.249.133:7778 exiting due to java.io.IOException: you must perform CRYPTO_INIT before REQUEST_DECRYPT_NONCE
             opening signcryption connection to limp.tricone-mining.com/108.171.201.209:7778
[modminer:/dev/ttyACM0:0.1] decrypting nonce at address 0x00000061
[modminer:/dev/ttyACM0:0.1]   encrypted nonce = 0x3dc0b671
[modminer:/dev/ttyACM0:0.1] decrypting nonce at address 0x00000088
[modminer:/dev/ttyACM0:0.1]   encrypted nonce = 0x6be90b8b
             restarting blocked threads for com.triconemining.signcryption.SCClient$Connection@682042d8
[modminer:/dev/ttyACM0:0.1] nonce decryption aborted
[modminer:/dev/ttyACM0:0.1] nonce decryption aborted
               finished restarting blocked threads for com.triconemining.signcryption.SCClient$Connection@682042d8
listener thread for limp.tricone-mining.com/108.171.201.209:7778 exiting due to java.io.IOException: you must perform CRYPTO_INIT before REQUEST_DECRYPT_NONCE
             opening signcryption connection to limp.tricone-mining.com/109.75.176.36:7778
[modminer:/dev/ttyACM0:0.1] loaded job  832e58082030f7270e7d157b00aa0fe5728718cf599c9491b17e2b0a2a0e8bbe:04dfa1df50b0b2a21a04faeb(age=29s)
[modminer:/dev/ttyACM0:0.2] loaded job  4dd738ec5403316c4a6889772d45d5a0019a6cb7b5584940d849f3a4681e9bc2:97dae23b50b0b29c1a04faeb(age=29s)
[modminer:/dev/ttyACM0:0.2] decrypting nonce at address 0x00000019
[modminer:/dev/ttyACM0:0.2]   encrypted nonce = 0xc90f6bd8
             restarting blocked threads for com.triconemining.signcryption.SCClient$Connection@386b4a7e
[modminer:/dev/ttyACM0:0.2] nonce decryption aborted
               finished restarting blocked threads for com.triconemining.signcryption.SCClient$Connection@386b4a7e
listener thread for limp.tricone-mining.com/109.75.176.36:7778 exiting due to java.io.IOException: you must perform CRYPTO_INIT before REQUEST_DECRYPT_NONCE
             opening signcryption connection to limp.tricone-mining.com/184.82.49.28:7778
[modminer:/dev/ttyACM0:0.0] loaded job  62f055c94015b224ba053a38cbd28b22be410f8c9c2ef7cefbb7fc85d0345acb:59559d9f50b0b29c1a04faeb(age=39s)
[modminer:/dev/ttyACM0:0.1] decrypting nonce at address 0x00000044                              
[modminer:/dev/ttyACM0:0.1]   encrypted nonce = 0x252c3089                                      
             restarting blocked threads for com.triconemining.signcryption.SCClient$Connection@4e661307
[modminer:/dev/ttyACM0:0.1] nonce decryption aborted                                            
               finished restarting blocked threads for com.triconemining.signcryption.SCClient$Connection@4e661307
listener thread for limp.tricone-mining.com/184.82.49.28:7778 exiting due to java.io.IOException: you must perform CRYPTO_INIT before REQUEST_DECRYPT_NONCE
             opening signcryption connection to limp.tricone-mini
eldentyrell
Donator
Legendary
*
Offline Offline

Activity: 966


felonious vagrancy, personified


View Profile WWW
November 24, 2012, 10:53:53 PM
 #870

Woke up to this on the "b" version of the jar.
Restarting the process seems to have fixed it.

Quote
[modminer:/dev/ttyACM0:0.1] nonce decryption aborted                                            
               finished restarting blocked threads for com.triconemining.signcryption.SCClient$Connection@4e661307
listener thread for limp.tricone-mining.com/184.82.49.28:7778 exiting due to java.io.IOException: you must perform CRYPTO_INIT before REQUEST_DECRYPT_NONCE
             opening signcryption connection to limp.tricone-mini

Ah, thank you, I know exactly what this is and how to fix it.  I will have 1.50d posted in the next 12 hours.

Thank you very much for posting a log file!  (you have no idea how many people expect me to be able to fix/troubleshoot stuff without a logfile, which I can't…)

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
lukasbradley
Donator
Member
*
Offline Offline

Activity: 90


View Profile
November 25, 2012, 01:53:11 AM
 #871

Thank you very much for posting a log file!  (you have no idea how many people expect me to be able to fix/troubleshoot stuff without a logfile, which I can't…)

Tyrell, I'd love to get you more information, but not sure how verbose it should be, or if you have some secret logging setting to get what you're after.  With the standard out/err, it seems TML 1.0b (and others before) literally just stops hashing.  I get output of the averages, but it slowly drops to 0.

Would turning on the read/write log help?
lukasbradley
Donator
Member
*
Offline Offline

Activity: 90


View Profile
November 25, 2012, 03:33:19 AM
 #872

This is what I see:

Quote
[modminer:/dev/ttyACM0:2.1] loaded job  67fbb2a394aaeb8c606ff9e966323b306f9e3234166a506b71b4a4fce54fd45c:f7b377c950b190661a04faeb(age=44s)
[modminer:/dev/ttyACM0:3.1] loaded job  0e6d167481fbfda5fa7579996c2136c752a106df11808b3f34ef835775e3659f:6eecee5650b190661a04faeb(age=44s)
[modminer:/dev/ttyACM0:3.2] loaded job  767ff82074dc31c246cf931bea303bbd22252bdcabf65a00ae516bace59766aa:2395452750b190671a04faeb(age=42s)
[modminer:/dev/ttyACM0:2.1] decrypting nonce at address 0x00000017
[modminer:/dev/ttyACM0:2.1]   encrypted nonce = 0xdbfa6d51
[modminer:/dev/ttyACM0:2.1]   found a share: 0x7b3931c2
[modminer:/dev/ttyACM0:2.1]   pool accepted share
[modminer:/dev/ttyACM0:0.0] decrypting nonce at address 0x000000b3
[modminer:/dev/ttyACM0:0.0]   encrypted nonce = 0xcbf41e92
[modminer:/dev/ttyACM0:0.0]   found a share: 0xe3493c71
[modminer:/dev/ttyACM0:0.0]   pool accepted share
[modminer:/dev/ttyACM0:0.0] prediction: optimalFreq=229mhz, errorrate=16%, hashrate=95MH/s, alpha=0.119 beta=1.442 gamma=-830.683
[modminer:/dev/ttyACM0:0.0] setting clock to 165 Mhz, mult=33 div=20
[modminer:/dev/ttyACM0:0.1] prediction: optimalFreq=229mhz, errorrate=23%, hashrate=87MH/s, alpha=0.176 beta=1.442 gamma=-1226.957
[modminer:/dev/ttyACM0:0.1] setting clock to 165 Mhz, mult=33 div=20
[modminer:/dev/ttyACM0:0.2] prediction: optimalFreq=229mhz, errorrate=23%, hashrate=87MH/s, alpha=0.176 beta=1.442 gamma=-1226.957
[modminer:/dev/ttyACM0:0.2] setting clock to 165 Mhz, mult=33 div=20
H:501 X:960 E:0 T:5m   |  H:188 E:0 A:45 R:0 T:17m5s

That 501 hash will fall until it hits zero.  Everything just sits there.
kakobrekla
Hero Member
*****
Offline Offline

Activity: 714


Psi laju, karavani prolaze.


View Profile
November 25, 2012, 04:42:39 AM
 #873

For starters, perhaps try

Code:
-Dread_write_debug=true

From your log seems that a reclock is causing issues? Have you trid playing with 'recalibrate clock' settings?

http://www.tricone-mining.com/tml.html

lukasbradley
Donator
Member
*
Offline Offline

Activity: 90


View Profile
November 25, 2012, 12:54:56 PM
 #874

For starters, perhaps try

Code:
-Dread_write_debug=true

Doesn't really offer clues that I can see:

Quote
H:429 X:922 E:3 T:5m   |  H:171 E:3 A:30 R:0 T:12m30s read(0x417) = 0x2e
read(0x457) = 0x17
read(0x497) = 0x5c
read(0x417) = 0x17
read(0x457) = 0x5c
read(0x497) = 0x5c
read(0x417) = 0x5c
read(0x457) = 0x2e
read(0x497) = 0xe1
read(0x417) = 0x5c
read(0x457) = 0x1
read(0x497) = 0xb8
read(0x417) = 0x2e
read(0x457) = 0x17
read(0x497) = 0x5c
read(0x417) = 0x17
read(0x457) = 0x5c
read(0x497) = 0x5c
read(0x417) = 0x5c
read(0x457) = 0x2e
read(0x497) = 0xe1
read(0x417) = 0x5c
read(0x457) = 0x1
read(0x497) = 0xb8
H:443 X:922 E:3 T:5m   |  H:176 E:3 A:31 R:0 T:12m35s read(0x417) = 0x2e
read(0x457) = 0x17
read(0x497) = 0x5c
read(0x417) = 0x17
read(0x457) = 0x2e
write(0x45c, 0x0)
write(0x45c, 0x1)
read(0x456) = 0x830ecd23
read(0x456) = 0x830ecd23
[modminer:/dev/ttyACM0:1.1] decrypting nonce at address 0x0000005c
[modminer:/dev/ttyACM0:1.1]   encrypted nonce = 0x830ecd23
H:443 X:922 E:3 T:5m   |  H:176 E:3 A:31 R:0 T:12m35s write(0x455, 0x2e)
read(0x497) = 0x5c
read(0x417) = 0x5c
read(0x457) = 0x2e
read(0x497) = 0xe1
read(0x417) = 0x5c
[modminer:/dev/ttyACM0:1.1]   found a share: 0xaecff5dd
[modminer:/dev/ttyACM0:1.1]   pool accepted share
             got long poll from 2.50, new prevblk=47333a8171f923d9ebf574f16e9bcb0c67633a3f9071d9da000002a100000000
H:0 X:922 E:0 T:5m   |  H:81 E:3 A:32 R:0 T:28m

Quote
From your log seems that a reclock is causing issues? Have you trid playing with 'recalibrate clock' settings?

http://www.tricone-mining.com/tml.html

But isn't that where the "magic" happens?
kakobrekla
Hero Member
*****
Offline Offline

Activity: 714


Psi laju, karavani prolaze.


View Profile
November 25, 2012, 01:00:21 PM
 #875

But isn't that where the "magic" happens?

Well, I found its best (at least in my case) to set fix clocks on a 'production run' (with enough distance to max so it doesn't produce even single invalid).

mining4fun11
Member
**
Offline Offline

Activity: 110


View Profile
November 26, 2012, 01:46:43 AM
 #876

But isn't that where the "magic" happens?

Well, I found its best (at least in my case) to set fix clocks on a 'production run' (with enough distance to max so it doesn't produce even single invalid).

What clocks are you using and whats are you entering at the command line. 
eldentyrell
Donator
Legendary
*
Offline Offline

Activity: 966


felonious vagrancy, personified


View Profile WWW
November 26, 2012, 02:33:36 AM
 #877

Ah, thank you, I know exactly what this is and how to fix it.  I will have 1.50d posted in the next 12 hours.

Okay, that took a little longer than I expected, but I got a chance to clean up a lot of rough edges.  1.50d is posted:

  http://www.tricone-mining.com/tml/tml-1.50d.jar

I will post more details shortly, but this has a big pile of minor improvements to the networking code.

It also fixes a bug in the accounting for non-difficulty-1 shares.  If your pool was sending you shares with diffculty>1 the previous TML would mine them just fine and calculate the "valid" share rate correctly (the rate at which correct nonces emerge from the device) but would sometimes under-report the rate of accepts (shares the pool accepts) and/or rejects (shares the pool rejects, including stales).

The new code carefully weights every accepted/rejected share according to its difficulty, so if the pool rejects a difficulty-2.125 share that counts as 2.125 rejections; likewise for accepts.  Shares which are below the pool's target go towards the valid count but do not influence the accept/reject counts.

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
eldentyrell
Donator
Legendary
*
Offline Offline

Activity: 966


felonious vagrancy, personified


View Profile WWW
November 26, 2012, 02:34:18 AM
 #878

With the standard out/err, it seems TML 1.0b (and others before) literally just stops hashing.  I get output of the averages, but it slowly drops to 0.

Wow, okay, that means something is seriously wrong.  Could you please post a Java thread dump instead?  On Linux you get this by pressing control-backslash or sending the process a SIGQUIT.  On Windows I think you use a program called jstack.exe and pass it the pid of the java process as the argument.  It will dump out a ton of stuff (a few thousand lines); probably best to put it on pastebin or you can just email it to me.  It's way too big to post here.  This will list out all the threads that are running and show me what line of code each of them is on.  From this I can figure out what's stuck.

(btw I have just posted 1.50d, so please try the above using that rather than an earlier version, if possible)

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
kakobrekla
Hero Member
*****
Offline Offline

Activity: 714


Psi laju, karavani prolaze.


View Profile
November 26, 2012, 02:36:50 AM
 #879

But isn't that where the "magic" happens?

Well, I found its best (at least in my case) to set fix clocks on a 'production run' (with enough distance to max so it doesn't produce even single invalid).

What clocks are you using and whats are you entering at the command line.  

Code:
java -Xbootclasspath/a:ZtexBTCMiner-120221.jar -Drecalibrate_clock=false -jar tml-1.50.jar ztex:0.0@175 ztex:0.1@172 ztex:0.2@175 http://user:pass@host:port

eldentyrell
Donator
Legendary
*
Offline Offline

Activity: 966


felonious vagrancy, personified


View Profile WWW
November 26, 2012, 02:38:39 AM
 #880

Folks,

At least for me, software and bitstream development requires a certain degree of isolation.  I had to swear off IRC many years ago because I realized that if I let myself hang out on IRC I'd never get any work done (most of you probably have more discipline in this regard than I do).

This works great for improving the TML, but doesn't always work so well for making sure users are well cared-for.  Therefore, to make sure that there is always somebody to respond and escalate serious issues, I will be sharing the signcryption revenues with kakobrekla, who will be staffing the #tml-support channel on irc.tricone-mining.com and will also handle first-line support here in this thread.  I'll still read this thread, hopefully at least a few times a week, but I'll be relying on kako to help people with the easy stuff and sort out (and document!) the real bugs from the getting-started issues.

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
Pages: « 1 2 3 4 5 6 7 8 9 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 »
  Print  
 
Jump to:  

Sponsored by , a Bitcoin-accepting VPN.
Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!