OFFICIAL CGMINER mining software thread for linux/win/osx/mips/arm/r-pi 4.11.1

[rkinnin]

...
Would you suggest I use the same instructions to upgrade my s1's and s2's as you posted for the s3?

S1: https://github.com/kanoi/cgminer-binaries/tree/master/AntS1

S2: https://github.com/kanoi/cgminer-binaries/tree/master/AntS2

when attempting to use the s2 link on the actual s2 on the upgrade page I get the following error after I hit "flash image":

tar: invalid magic tar: short read cp: can't stat '*': No such file or directory

Thoughts??

[kano]

Download it again and check the checksum on the file?

[rkinnin]

Download it again and check the checksum on the file?

i misunderstood one of your steps....sorry about that. 

RTFM.   

I am good.  Thanks.

[kenshirothefist]

ckolivas, could you please pull this pull request: https://github.com/ckolivas/cgminer/pull/679? This is a small but annoying bug, causing small loss of performance (in terms of accounted shares on pool) when changing difficulty - especially when using vardiff or often diff changing (when switching from low to high diff miner immediately sends high diff while is should/could still send low diff and pool might wrongly account for shares). Thanks!

[-ck]

ckolivas, could you please pull this pull request: https://github.com/ckolivas/cgminer/pull/679? This is a small but annoying bug, causing small loss of performance (in terms of accounted shares on pool) when changing difficulty - especially when using vardiff or often diff changing (when switching from low to high diff miner immediately sends high diff while is should/could still send low diff and pool might wrongly account for shares). Thanks!

I will look at it eventually; the original behaviour was chosen to be robust with dodgy pools' handling of diff so it was actually there intentionally. Saying you lose a lot of shares is a gross exaggeration as most decent pools change diff only once or twice ever and that many shares will not even amount to a single satoshi over a lifetime of mining.

[NiceHashSupport]

ckolivas, could you please pull this pull request: https://github.com/ckolivas/cgminer/pull/679? This is a small but annoying bug, causing small loss of performance (in terms of accounted shares on pool) when changing difficulty - especially when using vardiff or often diff changing (when switching from low to high diff miner immediately sends high diff while is should/could still send low diff and pool might wrongly account for shares). Thanks!

I will look at it eventually; the original behaviour was chosen to be robust with dodgy pools' handling of diff so it was actually there intentionally. Saying you lose a lot of shares is a gross exaggeration as most decent pools change diff only once or twice ever and that many shares will not even amount to a single satoshi over a lifetime of mining.

A robust implementation would be, if the miner was still sending low diff shares when pool requested higher diff. That way you may get some % of rejects, but would work without speed loss on pools that don't implement diff change correctly and on pools that do.

Of course, it is not a big loss, a percent or two on a pool that does vardiff moderately. But I cannot agree with satoshi claim. You forgot about reconnects (whenever miner goes offline and comes back, switches to another pool etc). There is a lot more lost than a single satoshi. Besides, according to how simple the fix is, it is also totally non-understandable to us, why it is not pulled yet (or taken look at it).

And, say this to miners; there is a patch waiting for cgminer that improves performance but "I will look at it eventually".

[-ck]

ckolivas, could you please pull this pull request: https://github.com/ckolivas/cgminer/pull/679? This is a small but annoying bug, causing small loss of performance (in terms of accounted shares on pool) when changing difficulty - especially when using vardiff or often diff changing (when switching from low to high diff miner immediately sends high diff while is should/could still send low diff and pool might wrongly account for shares). Thanks!

I will look at it eventually; the original behaviour was chosen to be robust with dodgy pools' handling of diff so it was actually there intentionally. Saying you lose a lot of shares is a gross exaggeration as most decent pools change diff only once or twice ever and that many shares will not even amount to a single satoshi over a lifetime of mining.

A robust implementation would be, if the miner was still sending low diff shares when pool requested higher diff. That way you may get some % of rejects, but would work without speed loss on pools that don't implement diff change correctly and on pools that do.

Of course, it is not a big loss, a percent or two on a pool that does vardiff moderately. But I cannot agree with satoshi claim. You forgot about reconnects (whenever miner goes offline and comes back, switches to another pool etc). There is a lot more lost than a single satoshi. Besides, according to how simple the fix is, it is also totally non-understandable to us, why it is not pulled yet (or taken look at it).

And, say this to miners; there is a patch waiting for cgminer that improves performance but "I will look at it eventually".

No, you are speaking ONLY of your pool/proxy/service which does lots of reconnects. You are biased towards your pool whereas I'm biased towards normal pools. Nonetheless the patch has been pulled into master git but it may be a very long time before a new cgminer release comes out since there's virtually nothing happening anywhere with regards to hardware till the next generation comes around.

[TheRealSteve]

If the issue is largely with reconnects to pools that start out with a too low diff for the device, and for whatever reason there's no minimum vardiff that can be set on the pool interface, see if the pool used accepts the --suggest-diff parameter?  If it's a proxy type service, the end-user doesn't even have to worry about setting that up.

[luthermarcus]

Hey could you add support for S5 like you did for s3?

[Mikestang]

Hey could you add support for S5 like you did for s3?

If you pay for it I bet they'd be happy to. 

[kano]

Hey could you add support for S5 like you did for s3?

Already working on it.

[toptek]

Hey could you add support for S5 like you did for s3?

Already working on it.

nice

I hope you actually fix it , so far you have once you do it . The new bitmain Firmware worked ok for me till today on one of my S5, then  for no good reason the speed started dropping then one of the blades just stopped and displayed

-------- from  zeros and No X es .


 i rebooted it, no go . so i reset it to defaults, no go, finally i was forced to go back to one of the  Firmware modified by Smit1237 with out #xnsub support sense that seems to do strange things to my S5's on some pools . it is working like it should be now with Smit1237 firmware but not with the new released firmware from bitmain . that did work well for a few days then to day for no reason it did , what i just said .

[luthermarcus]

Hey could you add support for S5 like you did for s3?

Already working on it.

nice

I hope you actually fix it , so far you have once you do it . The new bitmain Firmware worked ok for me till today on one of my S5, then  for no good reason the speed started dropping then one of the blades just stopped and displayed

-------- from  zeros and No X es .


 i rebooted it, no go . so i reset it to defaults, no go, finally i was forced to go back to one of the  Firmware modified by Smit1237 with out #xnsub support sense that seems to do strange things to my S5's on some pools . it is working like it should be now with Smit1237 firmware but not with the new released firmware from bitmain . that did work well for a few days then to day for no reason it did , what i just said .

I agree the s5 does work different on each pool. I just want it fixed for p2pool becuase it hash drops from 1200 to 800-900. It works good with some pools and like you say others it displays x for no good reason and starts to act up p2pool being one of them.

[toptek]

Hey could you add support for S5 like you did for s3?

Already working on it.

nice

I hope you actually fix it , so far you have once you do it . The new bitmain Firmware worked ok for me till today on one of my S5, then  for no good reason the speed started dropping then one of the blades just stopped and displayed

-------- from  zeros and No X es .


 i rebooted it, no go . so i reset it to defaults, no go, finally i was forced to go back to one of the  Firmware modified by Smit1237 with out #xnsub support sense that seems to do strange things to my S5's on some pools . it is working like it should be now with Smit1237 firmware but not with the new released firmware from bitmain . that did work well for a few days then to day for no reason it did , what i just said .

I agree the s5 does work different on each pool. I just want it fixed for p2pool becuase it hash drops from 1200 to 800-900. It works good with some pools and like you say others it displays x for no good reason and starts to act up p2pool being one of them.

I notice that to on  p2pool you log on it, then about 10 mins later it caps at 800 or 900 GS on my S5s on my S3 even before the fix had no issue there . Or i would be using it over antpool any day .

[tkalfaoglu]

On a fresh fedora 22 install, I compiled cgminer with -enable-icarus
and plugged in two antiminer U3's.

It "appears" to work because I can see the hashing status of the two devices, but it also dumps lots of errors like this:

ftdi_sio  ttyUSB0: Unable to write latency timer: -71
ftdi_sio  ttyUSB0: Unable to read latency timer: -71
ftdi_sio  ttyUSB0: Unable to write latency timer: -71
and it keeps repeating..

What can I do to fix this?

Thanks, -t

[Buchi-88]

Anyone know about this errors?

cgminer version 4.3.3 - Started: [2015-07-28 18:49:24]
--------------------------------------------------------------------------------
 (5s):12.87G (1m):12.78G (5m):12.56G (15m):11.84G (avg):12.65Gh/s
 A:7496  R:56  HW:0  WU:179.7/m
 Connected to multiple pools with block change notify
 Block: 9135be29...  Diff:1.24M  Started: [19:32:22]  Best share: 8.88K
--------------------------------------------------------------------------------
 USB management Pool management Settings Display options Quit
 0: BXM 0       :                         | 4.393G / 4.402Gh/s WU:61.9/m
 2: BXM 2       :                         | 4.305G / 4.203Gh/s WU:60.6/m
 3: BXM 3       :                         | 4.138G / 3.983Gh/s WU:57.9/m
--------------------------------------------------------------------------------
[2015-07-28 19:21:12] BXM 1: SPI RX error 0, read 0 of 81
 [2015-07-28 19:21:12] BXM 1 failure, disabling!
 [2015-07-28 19:21:12] BXM 1: SPI TX error -4, sent 0 of 11
 [2015-07-28 19:21:12] BXM 1: SPI TX error -4, sent 0 of 12
 [2015-07-28 19:21:18] BXM 0: SPI RX error -7, read 0 of 8
 [2015-07-28 19:21:19] BXM 0: SPI RX error -7, read 0 of 9

Then the miner restart?

[kano]

On a fresh fedora 22 install, I compiled cgminer with -enable-icarus
and plugged in two antiminer U3's.

It "appears" to work because I can see the hashing status of the two devices, but it also dumps lots of errors like this:

ftdi_sio  ttyUSB0: Unable to write latency timer: -71
ftdi_sio  ttyUSB0: Unable to read latency timer: -71
ftdi_sio  ttyUSB0: Unable to write latency timer: -71
and it keeps repeating..

What can I do to fix this?

Thanks, -t

See the first post for where you get the current master cgminer.
I've no idea what you are running or who you got it from.

[tkalfaoglu]

Thank you -- today I downloaded the master, and compiled it. So it is the most recent version.

[kano]

Thank you -- today I downloaded the master, and compiled it. So it is the most recent version.

./cgminer -h | head will show what you are running.
./cgminer -n will list the USB devices.
./cgminer -D --verbose (followed by your other options) will show debug of it running.
... and lastly ... the README says how to setup access to USB devices if you are not using root.

[hurricandave]

Um...I thought 4.9.2 was the most recent? Looks like he compiled 4.3.3?