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

[bronan]

and still no anwer from the gurus on my problem with cgminer and bfgminer

[Zanatos666]

and still no anwer from the gurus on my problem with cgminer and bfgminer

Take our the "name" lines in your config file.  Make the first two pools look like the third one you listed.

[HellDiverUK]

Just got an ASICMiner USB Block Erupter.  CGMiner 3.3.1 reports the expected 333MH/s, but I've tried it on two different pools over the course of several hours, and both pools are reporting about 100MH/s.

I don't see any rejects or hardware errors.  The 5s average is varying wildly between 600MH/s and 50MH/s, but average is a solid 333MH/s. 

I'm running the ASICMiner in a USB3 port (ASMedia chipset) off an Asu SABERTOOTH 990FX board, Windows 7 Pro x64, WinUSB drivers.

I also noticed that Karin's version of cgminer is showing an error where half the shares go missing.  Could this be the same issue in the 'official' cgminer?  https://bitcointalk.org/index.php?topic=214268.msg2748110#msg2748110

[Jcw188]

Hey guys, I can't seem to get CGMiner to work under scrypt no matter what I try. My cards (2 5970s WC'd) work perfectly fine with Bitcoin, but I try out scrypt mining for Litecoin nothing seems to be working

Also I have 4GB of RAM and am running a AMD Sempron 145 if that helps.

This is what I am faced with:

[snip]

These are the parameters I am using:

Code:

cgminer.exe --scrypt  -o stratum+tcp://stratum.give-me-ltc.com:3333 -u USERNAME -p PASSWORD -I 9

Also I am using SDK 2.1 and Drivers 11.12

SDK 2.1 won't work with scrypt.

I have a problem here that's similar.  WIndows xp 32 bit, AMD Radeon 6670.  When I try to install newest SDK it will only give me 2.4.x.  To do scrypt I need 2.6 SDK don't I?  Is there any workaround?  Thanks

[Aurum]

I'm trying to compile cgminer for 5970s and 5830s. I tried AMD ADL SDK v2.8.1 but it crashes. ReadMe recommends v2.4 or 2.5, but, AMD does not provide it in the archives any more. Does anyone know where I can download AMD ADL SDK v2.5??? TIA

[daemondazz]

If your usb hub is a usb2 hub and not a usb3 hub you will be far better off moving to the latest version of cgminer instead. The only reason to stick with the old version of cgminer is if you have multiple usb1.1 devices on a usb3 hub for there remains an instability there that we've been unable to debug so far.

Is that the cause for this:

Code:

Jul 16 18:37:12 server1 cgminer[31418]: Accepted 03e8a8b7 Diff 65/1 AMU 2 pool 0
Jul 16 18:37:12 server1 cgminer[31418]: AMU1: Comms error (werr=-7 amt=0)
Jul 16 18:37:12 server1 cgminer[31418]: AMU0: Comms error (werr=-7 amt=0)
Jul 16 18:37:13 server1 cgminer[31418]: AMU2: libusb pipe error, trying to clear

Works fine with 1 USB Erupter on the USB3 hub but it doesn't like multiples.

Let me know if I can help with any debugging.

I had some erupters on a hub that wasn't delivering the power it should have.  Erupters on that hub would give me that error.  After I wired the hub up to my PSU I don't see them any more.

Might not be the only cause, but that was it for me.

It is a powered USB3 hub (with external 3A power brick) and I get those errors with only 2 Erupters plugged in, so it would not appear to be an issue with power draw. The USB3 hub is connected to an internal USB2 header on the motherboard.

I should probably mention this is on Ubuntu 12.04 (precise).

[kano]

For erupter/icarus problems, try current git (with a USB2 hub)

I do know that Raspbian fails badly.

What the details of the problem are is that since the Icarus devices never return a finished message, if they don't find a share, they never return anything.
This is of course expected, however, it would seem that certain linux kernels on certain hardware do not support a usb request timeout on USB1.1 devices (yes weird isn't it)
On these kernels, cgminer simply never gets a reply form the call to libusb (or more correctly, the call to libusb never returns)

I've found on my RPi when I switched from Raspbian to Arch the problem went away.
http://www.kano-kun.net/?p=87

Edit: plugging the USB3.0 hub in locked up the RPi - but once it was plugged in and the RPi rebooted, it worked fine also.

[crazyates]

For erupter/icarus problems, try current git (with a USB2 hub)

I do know that Raspbian fails badly.

What the details of the problem are is that since the Icarus devices never return a finished message, if they don't find a share, they never return anything.
This is of course expected, however, it would seem that certain linux kernels on certain hardware do not support a usb request timeout on USB1.1 devices (yes weird isn't it)
On these kernels, cgminer simply never gets a reply form the call to libusb (or more correctly, the call to libusb never returns)

I've found on my RPi when I switched from Raspbian to Arch the problem went away.
http://www.kano-kun.net/?p=87

Have the issues I see in this thread seem to be focused around those Block Erupters. Seems they're more trouble than they're worth?

[kano]

For erupter/icarus problems, try current git (with a USB2 hub)

I do know that Raspbian fails badly.

What the details of the problem are is that since the Icarus devices never return a finished message, if they don't find a share, they never return anything.
This is of course expected, however, it would seem that certain linux kernels on certain hardware do not support a usb request timeout on USB1.1 devices (yes weird isn't it)
On these kernels, cgminer simply never gets a reply form the call to libusb (or more correctly, the call to libusb never returns)

I've found on my RPi when I switched from Raspbian to Arch the problem went away.
http://www.kano-kun.net/?p=87

Have the issues I see in this thread seem to be focused around those Block Erupters. Seems they're more trouble than they're worth?

Worth?
Way too much and return way too little.
They currently mine only 0.00640097 BTC a day ... ouch!

[HellDiverUK]

Yeah, they're expensive, yeah they're slow, but they should still work. 

Why does cgminer report one thing, and pools report another.  There's obviously a bug or some weirdness somewhere...

[notme]

Yeah, they're expensive, yeah they're slow, but they should still work. 

Why does cgminer report one thing, and pools report another.  There's obviously a bug or some weirdness somewhere...

I'm not sure what's wrong with your setup, but git head works fine with my block erupters with p2pool.

[-ck]

Have the issues I see in this thread seem to be focused around those Block Erupters. Seems they're more trouble than they're worth?

I certainly don't think they're worth it, in fact I'd go so far as to call them a rip off, but there are now hundreds of threads discussing this very issue and there are people still buying them. Most have valid explanations, but it just looks like post purchase rationalisation to me. There are no doubt quite a few new buyers who had no clue about the likelihood of meaningful profit amongst them.

[HellDiverUK]

I stuck the Erupter in a USB2 port this morning and it seems to be going better, BTCGuild is reporting 233MH/s in the space of about 20 minutes, which is much better.

Perhaps it's some funkiness with the USB3 on the SABERTOOTH?

I'm going to try it on an Intel machine this evening (B75 chipset) and see if it's any better.

[.m.]

Hi,
I tried cgminer recently, and I am experiencing strange behaviour.
Starting like this:

cgminer/cgminer -v 1 -w 256 -d 0 -I d -o http://gw:9332 -u cgminer7850 -p 234rert2231 --fix-protocol -g 2

Get rid of --fix-protocol as p2pool should be using stratum by now.

Thanks a lot for the tips.
The problem is cgminer - it just does not like p2pool or the fedora core 17 combination.
p2pool errors : https://bitcointalk.org/index.php?topic=18313.msg2752965#msg2752965
Switched back to DiabloMiner.
I was using BFGminer some years ago with no problem (dunno now).

[HellDiverUK]

I certainly don't think they're worth it, in fact I'd go so far as to call them a rip off, but there are now hundreds of threads discussing this very issue and there are people still buying them. Most have valid explanations, but it just looks like post purchase rationalisation to me. There are no doubt quite a few new buyers who had no clue about the likelihood of meaningful profit amongst them.

I'm personally not interested in profit.  I bought mine with BTC I made from GPU mining, which I also don't see a profit in.  I buy these things for fun - like some people who spend money on rigs for SETI@Home, RC5, etc.  It's a silly little hobby.  Expensive silly little hobby, but you should see my collection of wifi routers, hard-drives, and motherboards.  It's extensive.

Hell, I've even got a Jalapeno on order - I don't expect to see a profit when (if) it arrives later this year.  But I'll have fun playing with it.

I suppose, if that's your attitude to the Erupters, I can't see there being much hope of getting a fix, or you even looking to see if there's a bug.  Maybe BFGMiner will be better?

[-ck]

I certainly don't think they're worth it, in fact I'd go so far as to call them a rip off, but there are now hundreds of threads discussing this very issue and there are people still buying them. Most have valid explanations, but it just looks like post purchase rationalisation to me. There are no doubt quite a few new buyers who had no clue about the likelihood of meaningful profit amongst them.

I'm personally not interested in profit.  I bought mine with BTC I made from GPU mining, which I also don't see a profit in.  I buy these things for fun - like some people who spend money on rigs for SETI@Home, RC5, etc.  It's a silly little hobby.  Expensive silly little hobby, but you should see my collection of wifi routers, hard-drives, and motherboards.  It's extensive.

Hell, I've even got a Jalapeno on order - I don't expect to see a profit when (if) it arrives later this year.  But I'll have fun playing with it.

I suppose, if that's your attitude to the Erupters, I can't see there being much hope of getting a fix, or you even looking to see if there's a bug.  Maybe BFGMiner will be better?

I do like to call things as I see them with the profit side, but that's an aside to our development process. We are investigating the problems continuously as kanoi received one as a development unit, and the problems we're encountering are to do with unexpected behaviour through libusb which is our driver model. Note I do not have one myself but even I am trying to help debug this, however it's harder without one, and right now is bad because I'm currently in Europe on vacation for the next 5 weeks still. This is why our current recommendation is 3.1.1 for anyone using them on usb3 hubs. usb3 in general still appears problematic. You are of course free to use whatever software you like, and it's fair to say our (cgminer's) performance on the AMUs with anything usb3 is lacklustre and it probably annoys me more than anyone else that we have yet to solve them. Just ask kano how much I keep hassling him about progress on the AMUs.

[HellDiverUK]

Sorry, I didn't mean to be an ass about it. 

I have got it working OK by removing the WinUSB drivers, and running it on BFGMiner in the meantime.   I know the USB3 on this machine can be a bit flaky - good old AsMedia, I don't know why Asus bother with them.  I guess they're cheap.    I will investigate further with an Intel machine over the weekend.

I'd prefer to see cgminer working properly with the whole kit, though.  I appreciate the efforts.  I just get cranky first thing before my 4 cups of coffee.

[-ck]

Sorry, I didn't mean to be an ass about it.  

I have got it working OK by removing the WinUSB drivers, and running it on BFGMiner in the meantime.   I know the USB3 on this machine can be a bit flaky - good old AsMedia, I don't know why Asus bother with them.  I guess they're cheap.    I will investigate further with an Intel machine over the weekend.

I'd prefer to see cgminer working properly with the whole kit, though.  I appreciate the efforts.  I just get cranky first thing before my 4 cups of coffee.

Don't worry about it. Text discussions have a way of making people do that. Considering I'm recovering from the worst hangover of my entire life, I think I did pretty well myself in response  

[FiatKiller]

Hey ck, thought occurred to me today(amazing, I know!)... is there ever a point where the mining hardware speed would max out (using current processors) so that it would effectively hit a software speed ceiling? Or is hardware always going to be the limiting factor?

[-ck]

Hey ck, thought occurred to me today(amazing, I know!)... is there ever a point where the mining hardware speed would max out (using current processors) so that it would effectively hit a software speed ceiling? Or is hardware always going to be the limiting factor?

We did some benchmarking on the cgminer code and with a modern CPU it maxes out generating work at about 11TH (if it weren't written in C, it would hit a ceiling much earlier). If hardware gets that fast, hopefully the hardware itself will evolve to a point where it can generate work itself too, offloading some of the work. The minirig itself is the fastest hardware at the moment, and it has issues that make it inefficient to talk to, so running one of those alone uses 14% CPU of a 1.6Ghz CPU. Better communication protocols (and better mining hardware designs) should be able to decrease that dramatically. Since we are only on the first generation of ASICs things will change.