Bitcoin Forum
December 15, 2024, 03:35:26 PM *
News: Latest Bitcoin Core release: 28.0 [Torrent]
 
   Home   Help Search Login Register More  
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 »
  Print  
Author Topic: **US** BitFury Setup Guide  (Read 53236 times)
skrazy
Newbie
*
Offline Offline

Activity: 55
Merit: 0



View Profile
September 24, 2013, 08:49:46 PM
 #361

NOTE: I was one of the late august orders and received a 4 channel 2.3 M-board. Comments in this post only apply to those with this board or later revisions.

I pulled the latest commits from the bfsb/chainminer last night and have improvement in the stability of my starter kit over the last 12 hours. Error rates have dropped from ~7% to ~2%. More importantly my previous MISO errors have nearly disappeared. The comments on the pull suggest that by increasing the spi reset rate stability should increase. The H-Card is stable at around 36GH/s. If you are going to pull the latest commits make sure to back up your current software AND make sure you have the proper hardware to do so.

I'd like to hear about other experiences with the new spi resets.

^how did you do this? the chainminer 2.2 release or something different, and how was it implemented? Ive still got ssh errors (mining just fine) and plan to reformat the SD soon, and would LOVE to get my error rate down to 2% - that sounds like a dream compared to the current 6-12%!

I had to remove the chainminer directory on the MBP supplied sd card as git was not setup on the directory. I then cloned chainminer from https://github.com/bfsb/chainminer and issued a make.

I only have an hour of logs to show here but the standard deviation on my mean chip noncerate has dropped significantly with additional spi resets. The card is consistently pushing above 35 GH/s with a pair of DC fans and a bit of graphite.
https://i.imgur.com/ftjE9ej.jpg
greaterninja
Hero Member
*****
Offline Offline

Activity: 924
Merit: 1000


View Profile
September 24, 2013, 09:21:14 PM
 #362

NOTE: I was one of the late august orders and received a 4 channel 2.3 M-board. Comments in this post only apply to those with this board or later revisions.

I pulled the latest commits from the bfsb/chainminer last night and have improvement in the stability of my starter kit over the last 12 hours. Error rates have dropped from ~7% to ~2%. More importantly my previous MISO errors have nearly disappeared. The comments on the pull suggest that by increasing the spi reset rate stability should increase. The H-Card is stable at around 36GH/s. If you are going to pull the latest commits make sure to back up your current software AND make sure you have the proper hardware to do so.

I'd like to hear about other experiences with the new spi resets.

Very interesting  I have the 2.2 or 2.3 M-boards as well.   So you were at 7% error rate on one card, but now at 2%...pretty good.  What was your previous hash rate?  And what is your resistor reading at then, and what about now?

Sorry for the questions but I am very curious as to what will happen to my bad kit I got.
mo_mo
Full Member
***
Offline Offline

Activity: 194
Merit: 100


View Profile
September 24, 2013, 10:22:33 PM
 #363

I have the V2 motherboard, do I need to update the chainminer to increase the hashrate?
gmannn
Full Member
***
Offline Offline

Activity: 148
Merit: 100


View Profile
September 24, 2013, 11:22:45 PM
 #364

Thank you, I have a disappointing board with a 2 day hashrate average of only ~17 gh/s and a lot of MISO errors.  I'll update to the latest chainminer and see if it helps.  If not, I will return/RMA the board, rather than void my warranty by overclocking it.



NOTE: I was one of the late august orders and received a 4 channel 2.3 M-board. Comments in this post only apply to those with this board or later revisions.

I pulled the latest commits from the bfsb/chainminer last night and have improvement in the stability of my starter kit over the last 12 hours. Error rates have dropped from ~7% to ~2%. More importantly my previous MISO errors have nearly disappeared. The comments on the pull suggest that by increasing the spi reset rate stability should increase. The H-Card is stable at around 36GH/s. If you are going to pull the latest commits make sure to back up your current software AND make sure you have the proper hardware to do so.

I'd like to hear about other experiences with the new spi resets.

Very interesting  I have the 2.2 or 2.3 M-boards as well.   So you were at 7% error rate on one card, but now at 2%...pretty good.  What was your previous hash rate?  And what is your resistor reading at then, and what about now?

Sorry for the questions but I am very curious as to what will happen to my bad kit I got.
skrazy
Newbie
*
Offline Offline

Activity: 55
Merit: 0



View Profile
September 24, 2013, 11:23:48 PM
 #365

The in circuit resistance is ~1.18K at room temperature. I've not measured the resistance once the system is up to operating temperature. I've moving air across the board somewhat aggressively to keep temperatures constant and the regulator happy. My digikey order came in today and I'll try to work up some stats on swapping resistors out this weekend. I'm not interested in using the pencil mod as a long term solution.

I'd take a "if its not broken dont..." stance on software changes unless you are having similar bus issues. I was interested in the spi changes (increased speed and reset rate) and have observed a positive increase in stability/noncerate by upgrading. I'll try to do a sensitivity sweep of the reset rate to find what works best for my setup on the v2 boards.
Swimmer63
Legendary
*
Offline Offline

Activity: 1593
Merit: 1004



View Profile
September 24, 2013, 11:29:07 PM
 #366

My miner has worked great since I got it going 2 weeks ago.  No restarts or anything.  But today I changed the pool configuration to another pool.  Now I can't get it to run again.  I have been restarting it both with the buttons in the browser and physically and it will not go.  Any suggestions?  Below is a photo of the start up page.  I'm a little concerned about the line that says "miner: no process found.

klondike_bar
Legendary
*
Offline Offline

Activity: 2128
Merit: 1005

ASIC Wannabe


View Profile
September 24, 2013, 11:30:47 PM
 #367

The in circuit resistance is ~1.18K at room temperature. I've not measured the resistance once the system is up to operating temperature. I've moving air across the board somewhat aggressively to keep temperatures constant and the regulator happy. My digikey order came in today and I'll try to work up some stats on swapping resistors out this weekend. I'm not interested in using the pencil mod as a long term solution.

I'd take a "if its not broken dont..." stance on software changes unless you are having similar bus issues. I was interested in the spi changes (increased speed and reset rate) and have observed a positive increase in stability/noncerate by upgrading. I'll try to do a sensitivity sweep of the reset rate to find what works best for my setup on the v2 boards.

any sign of chips stopping at random? When I had 1.16K i could rarely get more then 15 chips working, and sometimes it would drop to 14 at times. I now have about 1.178k and 16 chips work well for the first 5-15 minutes, then it seems to drop to 15 chips again. Ive only once in the last 24hrs seen it drop to 14 chips. Im not sure if this is the chips themselves, or if the pencil mod is pushing the limits of the DC 30A converter and forcing a chip to turn off as some result of that.

24" PCI-E cables with 16AWG wires and stripped ends - great for server PSU mods, best prices https://bitcointalk.org/index.php?topic=563461
No longer a wannabe - now an ASIC owner!
-Redacted-
Hero Member
*****
Offline Offline

Activity: 574
Merit: 501


View Profile
September 24, 2013, 11:33:08 PM
 #368

My miner has worked great since I got it going 2 weeks ago.  No restarts or anything.  But today I changed the pool configuration to another pool.  Now I can't get it to run again.  I have been restarting it both with the buttons in the browser and physically and it will not go.  Any suggestions?  Below is a photo of the start up page.  I'm a little concerned about the line that says "miner: no process found.


That may or may not be an issue.  Log in as pi with password raspberry, and do a "sudo screen -ls" to see what processes are running.

You should have one screen process running for each proxy you have enabled, and one for the miner.  I only have one proxy running here...

pi@bitfury ~ $ sudo screen -ls
There are screens on:
        27739..bitfury  (24/09/13 23:30:00)     (Detached)
        3657..bitfury   (21/09/13 02:24:32)     (Detached)
2 Sockets in /var/run/screen/S-root.

EDIT:

Make doubly certain your new pool config is correct.   
You can't use http:// in front of the pool URL.   
You can't be using port 8332, 8333, or 8334.



skrazy
Newbie
*
Offline Offline

Activity: 55
Merit: 0



View Profile
September 24, 2013, 11:35:49 PM
 #369

any sign of chips stopping at random? When I had 1.16K i could rarely get more then 15 chips working, and sometimes it would drop to 14 at times. I now have about 1.178k and 16 chips work well for the first 5-15 minutes, then it seems to drop to 15 chips again. Ive only once in the last 24hrs seen it drop to 14 chips. Im not sure if this is the chips themselves, or if the pencil mod is pushing the limits of the DC 30A converter and forcing a chip to turn off as some result of that.

Have you tried clocking the chips down? Mine seem much happier at 54 or 53 with the higher voltage. I need a bit of patience with mine as well, it often does not achieve its top end for a few hours.
Swimmer63
Legendary
*
Offline Offline

Activity: 1593
Merit: 1004



View Profile
September 24, 2013, 11:40:27 PM
 #370

My miner has worked great since I got it going 2 weeks ago.  No restarts or anything.  But today I changed the pool configuration to another pool.  Now I can't get it to run again.  I have been restarting it both with the buttons in the browser and physically and it will not go.  Any suggestions?  Below is a photo of the start up page.  I'm a little concerned about the line that says "miner: no process found.


That may or may not be an issue.  Log in as pi with password raspberry, and do a "sudo screen -ls" to see what processes are running.

You should have one screen process running for each proxy you have enabled, and one for the miner.  I only have one proxy running here...

pi@bitfury ~ $ sudo screen -ls
There are screens on:
        27739..bitfury  (24/09/13 23:30:00)     (Detached)
        3657..bitfury   (21/09/13 02:24:32)     (Detached)
2 Sockets in /var/run/screen/S-root.



I filled in all three proxy's but they are to the same pool.  That's the way I had it set up when it was working.
My screen is also going off and on occasionally.

https://i.imgur.com/e7Jd9cC.jpg
-Redacted-
Hero Member
*****
Offline Offline

Activity: 574
Merit: 501


View Profile
September 24, 2013, 11:46:30 PM
 #371

That looks correct.  Now do a cat /run/shm/.putstat.log

That will show you if the proxies are doing anything and the port they're connected on.  I only have one running so mine looks like this:

6 1461752 937198 937192 [ 0 ]http://127.0.0.1:8332/
0 1390783 839498 839498 [ 1 ]http://127.0.0.1:8333/
0 1396964 703864 703864 [ 2 ]http://127.0.0.1:8334/


If that looks OK, do a   cat /opt/bitfury/rpoxy_pools.conf and see if it looks mostly like this.  Again, I only have one enabled.

POOL1_START=1
POOL1_ADDRESS=stratum.bitcoin.cz
POOL1_PORT=3333
POOL1_USER=xxxxx.work_01
POOL1_PASS=xxxxx
POOL2_START=0
POOL2_ADDRESS=stratum.btcguild.com
POOL2_PORT=3333
POOL2_USER=rxxxxx_02
POOL2_PASS=xxxxx
POOL3_START=0
POOL3_ADDRESS=us2.eclipsemc.com
POOL3_PORT=3333
POOL3_USER=xxxxx_worker3
POOL3_PASS=xxxxx


Swimmer63
Legendary
*
Offline Offline

Activity: 1593
Merit: 1004



View Profile
September 24, 2013, 11:54:57 PM
 #372

That looks correct.  Now do a cat /run/shm/.putstat.log

That will show you if the proxies are doing anything and the port they're connected on.  I only have one running so mine looks like this:

6 1461752 937198 937192 [ 0 ]http://127.0.0.1:8332/
0 1390783 839498 839498 [ 1 ]http://127.0.0.1:8333/
0 1396964 703864 703864 [ 2 ]http://127.0.0.1:8334/


If that looks OK, do a   cat /opt/bitfury/rpoxy_pools.conf and see if it looks mostly like this:

POOL1_START=1
POOL1_ADDRESS=stratum.bitcoin.cz
POOL1_PORT=3333
POOL1_USER=xxxxx.work_01
POOL1_PASS=xxxxx
POOL2_START=0
POOL2_ADDRESS=stratum.btcguild.com
POOL2_PORT=3333
POOL2_USER=rxxxxx_02
POOL2_PASS=xxxxx
POOL3_START=0
POOL3_ADDRESS=us2.eclipsemc.com
POOL3_PORT=3333
POOL3_USER=xxxxx_worker3
POOL3_PASS=xxxxx




Not really.  I turned off the last two Proxys.  Here is mine:

0 754 0 0 [ 0 ]http://127.0.0.1:8332/
0 773 0 0 [ 1 ]http://127.0.0.1:8333/
0 761 0 0 [ 2 ]http://127.0.0.1:8334/
-Redacted-
Hero Member
*****
Offline Offline

Activity: 574
Merit: 501


View Profile
September 24, 2013, 11:58:45 PM
 #373

Well, if possible post your  proxy_pools.conf  (not rpoxy_pools)  - I don't think they are running.   I'd guess bad config...

As a sanity check, you can always put the original pool configuration back and try restarting the proxies.
Swimmer63
Legendary
*
Offline Offline

Activity: 1593
Merit: 1004



View Profile
September 25, 2013, 12:03:17 AM
 #374

Well, if possible post your  proxy_pools.conf  (not rpoxy_pools)  - I don't think they are running.   I'd guess bad config...

As a sanity check, you can always put the original pool configuration back and try restarting the proxies.

Here you go.  Showing both:

https://i.imgur.com/Lw1zyeg.jpg
Swimmer63
Legendary
*
Offline Offline

Activity: 1593
Merit: 1004



View Profile
September 25, 2013, 12:06:26 AM
 #375

Well, if possible post your  proxy_pools.conf  (not rpoxy_pools)  - I don't think they are running.   I'd guess bad config...

As a sanity check, you can always put the original pool configuration back and try restarting the proxies.

Here you go.  Showing both:

https://i.imgur.com/Lw1zyeg.jpg

Did the Sanity Check and it started right off!  WTF.  I guess I can't use that pool?
-Redacted-
Hero Member
*****
Offline Offline

Activity: 574
Merit: 501


View Profile
September 25, 2013, 12:07:46 AM
 #376

You got me there....   I haven't tried to run bitminter in the bitfury.   Give me a minute to try it....

Seems to work for me, but I use multiple workers, so my username is    redacted_worker3

                                                                 Rej   Acc
I got a crapload of rejects when it started up:   77   43   

Hashrate gauge on the stats page seems screwed....
frankenmint
Legendary
*
Offline Offline

Activity: 1456
Merit: 1018


HoneybadgerOfMoney.com Weed4bitcoin.com


View Profile WWW
September 25, 2013, 12:09:51 AM
 #377

If any of you with 100gh/s or more want an invite to giga's private pool (https://bitcointalk.org/index.php?topic=206181.0) just let me know. My Bitfury starter kit is getting 0.03% reject rate there.


honest question - why would I join a private pool instead of a large one like btcguild or bitminter or eligious?

Swimmer63
Legendary
*
Offline Offline

Activity: 1593
Merit: 1004



View Profile
September 25, 2013, 12:11:27 AM
 #378


You got me.  I haven't tried to run bitminter in the bitfury.   Give me a minute to try it....
Just noticed some earlier posts from today.  Some have it working fine.  Others do not.
klondike_bar
Legendary
*
Offline Offline

Activity: 2128
Merit: 1005

ASIC Wannabe


View Profile
September 25, 2013, 12:17:15 AM
 #379

Well, if possible post your  proxy_pools.conf  (not rpoxy_pools)  - I don't think they are running.   I'd guess bad config...

As a sanity check, you can always put the original pool configuration back and try restarting the proxies.

Here you go.  Showing both:

https://i.imgur.com/Lw1zyeg.jpg

Did the Sanity Check and it started right off!  WTF.  I guess I can't use that pool?

I use this just fine :
mint.bitminter.com             3333           XXX.BF1       XX1
mint.bitminter.com             3333           XXX.BF2       XX2

I think your problem is trying to use port 3332. use 3333 and 2 workers

INTERESTING: I just changed from 3 workers to 2 workers only, and my hashrate has gone from 30.5-33 GHash average per 'bitminter shift' to a much more steady 34.2-35.0 GHash. It seems bitminter does NOT benefit from 3 workers>2 workers. I have yet to see how 1 worker only is handled, but this seems pretty optimal at the moment

24" PCI-E cables with 16AWG wires and stripped ends - great for server PSU mods, best prices https://bitcointalk.org/index.php?topic=563461
No longer a wannabe - now an ASIC owner!
Cablez
Legendary
*
Offline Offline

Activity: 1400
Merit: 1000


I owe my soul to the Bitcoin code...


View Profile
September 25, 2013, 12:47:06 AM
 #380

If any of you with 100gh/s or more want an invite to giga's private pool (https://bitcointalk.org/index.php?topic=206181.0) just let me know. My Bitfury starter kit is getting 0.03% reject rate there.


honest question - why would I join a private pool instead of a large one like btcguild or bitminter or eligious?


First, it has 0% fee which none of the public pools have. It does have a very small withdrawal fee and I do mean small.

Second, it is good to decentralize the pool hierarchy to minimize possible (albeit unlikely) 51% attacks.

Third, I am sure there are probably more but the first 2 suffice.

Tired of substandard power distribution in your ASIC setup???   Chris' Custom Cablez will get you sorted out right!  No job too hard so PM me for a quote
Check my products or ask a question here: https://bitcointalk.org/index.php?topic=74397.0
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 »
  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!