[ANN][BURST] Burst | Efficient HDD Mining | New 1.2.3 Fork block 92000 - page 935.

wmikrut

hero member

Activity: 631

Merit: 501

Quote from: fanepatent on September 08, 2014, 12:12:20 PM

Quote from: wmikrut on September 08, 2014, 11:47:11 AM

Understood... thank you!

I haven't tried to create shortcuts of files directly. It will be very interesting if it will work...

BRB

I know there is a windows tool (NTFS Links maybe) that lets you symlink files like you do in Linux. The only gotcha was if you deleted the link... it removed the file as well.

I haven't used it since XP and would see if there is a W7 version of it.

tex81

full member

Activity: 145

Merit: 100

1) If total number of nonces is 2^64 - is there any (statistical) reason to use all 64 bits in nonce numbers?
2) If I mine on pool for 2 different accounts (colocation) - should I use 2 miners?

fanepatent

full member

Activity: 224

Merit: 100

Quote from: wmikrut on September 08, 2014, 11:47:11 AM

Understood... thank you!

I haven't tried to create shortcuts of files directly. It will be very interesting if it will work...

BRB

edit: Guess what... You can make direct shortcuts of the files in one folder and use one miner. I guess the same could be applied to Windows with symlinks

wmikrut

hero member

Activity: 631

Merit: 501

Understood... thank you!

fanepatent

full member

Activity: 224

Merit: 100

Quote from: wmikrut on September 08, 2014, 11:43:40 AM

So basically I can create a plot file... let's say -- 10G worth.
I test, I am happy... and not I want to soak up another 90G.

I can just create a second plot file (no overlapping nonces) and as long as they are in the same folder, everything is good.
I know under Linux I could symlink the plot files from other drives into the /plots folder -- as long at they are there, they will be used for mining.

Does that sound accurate?
This may actually be answered by the previous post as well.

It was answered. By me.
Short story, symlinks are necessary for Win. In Xubuntu the miner worked with a simple drag n drop shortcut.

wmikrut

hero member

Activity: 631

Merit: 501

So basically I can create a plot file... let's say -- 10G worth.
I test, I am happy... and not I want to soak up another 90G.

I can just create a second plot file (no overlapping nonces) and as long as they are in the same folder, everything is good.
I know under Linux I could symlink the plot files from other drives into the /plots folder -- as long at they are there, they will be used for mining.

Does that sound accurate?
This may actually be answered by the previous post as well.

Patek

member

Activity: 64

Merit: 10

Huh, after couple hours of testing and stuff I finally got miner working. However, still got some questions - here they come:

I'm running miner on 2TB partition - 2 threads cpu and 4gb ram. My plot generator config looks like this (found on previous pages):

Code:

C:\Windows\System32\java.exe -Xmx1000m -cp pocminer.jar;lib/*;lib/akka/*;lib/jetty/* pocminer.POCMiner generate my_account 0 81920 1000 2
C:\Windows\System32\java.exe -Xmx1000m -cp pocminer.jar;lib/*;lib/akka/*;lib/jetty/* pocminer.POCMiner generate my_account 81921 81920 1000 2
C:\Windows\System32\java.exe -Xmx1000m -cp pocminer.jar;lib/*;lib/akka/*;lib/jetty/* pocminer.POCMiner generate my_account 163842 81920 1000 2
C:\Windows\System32\java.exe -Xmx1000m -cp pocminer.jar;lib/*;lib/akka/*;lib/jetty/* pocminer.POCMiner generate my_account 245763 81920 1000 2
pause

It produces plot files of ~20 gb in size. Can I make any changes to make it more efficient? Or is it already running as supposed?
What happens when all 4 files are created and config says "pause"? I assume it will stop generating plots. What then? IS it time for miner to finally solve those plots? How will I know if miner solved everything there was to be solved? What to do once its done? Erase those plot files and start over again?

Amph

legendary

Activity: 3248

Merit: 1070

can someone tell me his kernel memory(paged) for 3tb? my paged is 5gb with 3tb, is this normal?

wmikrut

hero member

Activity: 631

Merit: 501

Quote from: Matt9301 on September 08, 2014, 09:09:27 AM

Quote from: wmikrut on September 08, 2014, 09:04:42 AM

So just a q I have had trouble finding an answer to.
Plot generation... got it. No overlapping nonces.

So let's say I have three HDDs and I am too lazy to make that one big disk.
I could generate plots on each disk (keeping nonces unique).

Do I then solo mine three clients (one for each disk).
So for example I have :

d:\pocminer_v1\plots
e:\pocminer_v1\plots
f:\pocminer_v1\plots

If I were to pool mine, would I also need to run three clients... one on each disk?
Thank you!

You must run 3 instance.. 1 for each hdd

Thank you Matt and Cobra for the responses!

Nefritox

sr. member

Activity: 358

Merit: 251

Quote from: muzkle on September 08, 2014, 11:13:44 AM

Lets call @bittrex to add BURST on the list Cheesy

why we waiting for this so long Huh

paradigmflux

sr. member

Activity: 378

Merit: 254

small fry

Quote from: paradigmflux on September 08, 2014, 10:54:41 AM

Is it possible to plot over SMB?

SMB/CIFS

muzkle

full member

Activity: 266

Merit: 100

The Premier Digital Asset Management Ecosystem

Lets call @bittrex to add BURST on the list Cheesy

uray

hero member

Activity: 1400

Merit: 505

Quote from: ltcnim on September 08, 2014, 11:07:58 AM

looks like the first one is the real read data, and seconds one is calculated nonces (bytes * 1024 * noncesize)

i have one suspect why dcct-miner has bad deadline, i will post after testing it

ltcnim

legendary

Activity: 914

Merit: 1001

looks like the first one is the real read data, and seconds one is calculated nonces (bytes * 1024 * noncesize)

uray

hero member

Activity: 1400

Merit: 505

Quote from: ltcnim on September 08, 2014, 10:54:16 AM

Quote from: uray on September 08, 2014, 10:51:34 AM

Quote from: ltcnim on September 08, 2014, 10:16:56 AM

Quote from: ltcnim on September 08, 2014, 09:49:53 AM

Quote from: dcct on September 08, 2014, 09:45:43 AM

Quote from: ltcnim on September 08, 2014, 09:44:39 AM

hmm..the files are generated just a few hours ago, and my best deadline with the java miner using these files was around 5min an hour ago. is it possible that the miner doesnt like to be a 32bit miner? ^^

Code:

dcct-miner/dcct-miner: ELF 32-bit LSB executable, ARM, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.26, BuildID[sha1]=0x09410d96b865886551d02c360cae703322d5336f, not stripped

Are you using the correct passphrase?

passphrase? where? i used the tutorial here: http://burst-pool.cryptoport.io/howto.html there's nothing written about a passphrase when using your miner. I just checked back with the java miner, and instantly got much better deadlines:

Code:

{"result":"success","deadline":756214}
{"result":"success","deadline":824025}
...

edit: ah, I guess you meant the passphrase when generating the plots? yes, should have been the correct one.

i just checked the code of the miner, and the extremely high deadline that has been submitted can be explained with the fact, that the pool does not have a limit (well...nearly):

Code:

if(best < 5000000000000ULL)

so I still don't understand why the java miner finds much better and more deadlines, and why the C miner does only show ~320GB from ~620GB

edit: i'll try again and let it run a little longer...

actually that limit was from original code of dcct miner, my pool will accept any deadline, you can delete that line if you want,
and for "xxx read/yyy total" , its shows actual byte read, since only 1/4096 data is needed to be read, for example if you have 4 TB plot, dcct-miner will shows highest read is 1 GB

do you mean that 72MB/320GB does not mean my plot files are only detected to be 320GB? also it still does not explain why the java miner finds way more and better deadlines.

no, it shows byte read progress, if it shows 72MB/320 GB, it said miner has read 320GB of your plot.
see this code, both number is from same variable but the second one is 1/4 of the first one ( first on with MB suffix, and second one with GB suffix)

Code:

printf("\r%llu MB read/%llu GB total/no deadline ", (bytesRead / ( 1024 * 1024 )), (bytesRead / (256 * 1024)));

i dont really understand why it write like that

Amph

legendary

Activity: 3248

Merit: 1070

something is wrong, when i close the miner the mem used remain there, and i need to reboot, Huh

i've already removed superfetch and disable it in services

saamxx

hero member

Activity: 546

Merit: 503

I'm wondering about Burst aliases. Seems nobody pays any attention at this powerfull feature http://www.nxttechnologytree.com/nxt-technology/nxt-alias-system#main-wrapper
This fucking Cheesy

BURST-FUYK-3W9E-VUK7-64VRA squatter seats on 98% of all aliases that have been already registered. In other hand, there are thousands of valuable aliases still free.

paradigmflux

sr. member

Activity: 378

Merit: 254

small fry

Is it possible to plot over SMB?

ltcnim

legendary

Activity: 914

Merit: 1001

Quote from: uray on September 08, 2014, 10:51:34 AM

Quote from: ltcnim on September 08, 2014, 10:16:56 AM

Quote from: ltcnim on September 08, 2014, 09:49:53 AM

Quote from: dcct on September 08, 2014, 09:45:43 AM

Quote from: ltcnim on September 08, 2014, 09:44:39 AM

hmm..the files are generated just a few hours ago, and my best deadline with the java miner using these files was around 5min an hour ago. is it possible that the miner doesnt like to be a 32bit miner? ^^

Code:

dcct-miner/dcct-miner: ELF 32-bit LSB executable, ARM, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.26, BuildID[sha1]=0x09410d96b865886551d02c360cae703322d5336f, not stripped

Are you using the correct passphrase?

passphrase? where? i used the tutorial here: http://burst-pool.cryptoport.io/howto.html there's nothing written about a passphrase when using your miner. I just checked back with the java miner, and instantly got much better deadlines:

Code:

{"result":"success","deadline":756214}
{"result":"success","deadline":824025}
...

edit: ah, I guess you meant the passphrase when generating the plots? yes, should have been the correct one.

i just checked the code of the miner, and the extremely high deadline that has been submitted can be explained with the fact, that the pool does not have a limit (well...nearly):

Code:

if(best < 5000000000000ULL)

so I still don't understand why the java miner finds much better and more deadlines, and why the C miner does only show ~320GB from ~620GB

edit: i'll try again and let it run a little longer...

actually that limit was from original code of dcct miner, my pool will accept any deadline, you can delete that line if you want,
and for "xxx read/yyy total" , its shows actual byte read, since only 1/4096 data is needed to be read, for example if you have 4 TB plot, dcct-miner will shows highest read is 1 GB

do you mean that 72MB/320GB does not mean my plot files are only detected to be 320GB? also it still does not explain why the java miner finds way more and better deadlines.

uray

hero member

Activity: 1400

Merit: 505

Quote from: ltcnim on September 08, 2014, 10:16:56 AM

Quote from: ltcnim on September 08, 2014, 09:49:53 AM

Quote from: dcct on September 08, 2014, 09:45:43 AM

Quote from: ltcnim on September 08, 2014, 09:44:39 AM

hmm..the files are generated just a few hours ago, and my best deadline with the java miner using these files was around 5min an hour ago. is it possible that the miner doesnt like to be a 32bit miner? ^^

Code:

dcct-miner/dcct-miner: ELF 32-bit LSB executable, ARM, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.26, BuildID[sha1]=0x09410d96b865886551d02c360cae703322d5336f, not stripped

Are you using the correct passphrase?

passphrase? where? i used the tutorial here: http://burst-pool.cryptoport.io/howto.html there's nothing written about a passphrase when using your miner. I just checked back with the java miner, and instantly got much better deadlines:

Code:

{"result":"success","deadline":756214}
{"result":"success","deadline":824025}
...

edit: ah, I guess you meant the passphrase when generating the plots? yes, should have been the correct one.

i just checked the code of the miner, and the extremely high deadline that has been submitted can be explained with the fact, that the pool does not have a limit (well...nearly):

Code:

if(best < 5000000000000ULL)

so I still don't understand why the java miner finds much better and more deadlines, and why the C miner does only show ~320GB from ~620GB

edit: i'll try again and let it run a little longer...

actually that limit was from original code of dcct miner, my pool will accept any deadline, you can delete that line if you want,
and for "xxx read/yyy total" , its shows actual byte read, since only 1/4096 data is needed to be read, for example if you have 4 TB plot, dcct-miner will shows highest read is 1 GB

Topic: [ANN][BURST] Burst | Efficient HDD Mining | New 1.2.3 Fork block 92000 - page 935. (Read 2171083 times)