Can anyone recommend a USB 3.0 hub? Both I tried so far (Aitech and Orico) suck  Do you know Anker USpeed USB 3.0 Hub (9 Port + 1 Charge Port)? No. Please let me know if it works reliably (and with a power switch..) |
|
|
|
Can anyone recommend a USB 3.0 hub? Both I tried so far (Aitech and Orico) suck |
|
|
|
Any chance of posting the source code for the firmware images on the devices?
The current firmware is based off an old copy of cgminer which has poor performance for large blocks or large coinbase transactions. As a result I'm seeing 100% cpu usage and a stale rate several times higher than my avalons. This could easily be fixed by porting the driver to bfgminer or a current cgminer (or backporting some of the work generation improvements from newer cgminer), and I'm willing to give it a shot at doing it myself— but I need the source.
Bitmain promised me a sample unit about 3 months ago. Haven't heard from them since then, but hopefully they'll catch up soon so I can work with you on the port. Luke |
|
|
|
any special drivers i should be using? The default, standard USB HID driver. |
|
|
|
This feels like stack corruption. How often does it happen? Are you running the latest firmware?
Every time I run it. Can you send me your current firmware? |
|
|
|
3.7.0 bug: [2013-11-28 22:21:58] Started bfgminer 3.7.0Segmentation fault: 11
Cant mine...
Need a backtrace or bisect. 3.6.0 Is not detecting my bitfury redfury ASIC anymore, how do I fix this?
Need output from They don't seem to like to play together. 3.7 starts and recognizes the Jalapeno, Blue Fury and BiFury but if I try to add the erupters it crashes quickly. After dinner I'll try running the BEs on a separate hub and see if that makes any difference.
Need a backtrace or bisect. OK I have a problem
OP Windows 7
BFGMiner 3.7.0
BiFURY running by itself on a usb 3 hub with a fan
I can get it running and the temp is at about 48C
But after a few miner BFGMiner crashes
Can anyone help
0x4898c8 : C:\Users\Douglas Seager\Desktop\bfgminer-3.7.0-win32\bfgminer.exe : /
home/luke-jr/Projects/Education/Tonal/BitCoin/bfgminer/make-release-tmp/bfgminer
-3.7.0-tmp/driver-bifury.c (467) : in function (bifury_poll)
0x42c9ac : C:\Users\Douglas Seager\Desktop\bfgminer-3.7.0-win32\bfgminer.exe : /
home/luke-jr/Projects/Education/Tonal/BitCoin/bfgminer/make-release-tmp/bfgminer
-3.7.0-tmp/deviceapi.c (641) : in function (minerloop_queue)
0x42ccb3 : C:\Users\Douglas Seager\Desktop\bfgminer-3.7.0-win32\bfgminer.exe : /
home/luke-jr/Projects/Education/Tonal/BitCoin/bfgminer/make-release-tmp/bfgminer
-3.7.0-tmp/deviceapi.c (688) : in function (miner_thread)
Failed to init bfd from (C:\Users\Douglas Seager\Desktop\bfgminer-3.7.0-win32\pt
hreadGC2.dll)
0x6248671b : C:\Users\Douglas Seager\Desktop\bfgminer-3.7.0-win32\pthreadGC2.dll
: pthread_create
Failed to init bfd from (C:\Windows\syswow64\msvcrt.dll)
0x77221287 : C:\Windows\syswow64\msvcrt.dll : itow_s
Failed to init bfd from (C:\Windows\syswow64\msvcrt.dll)
0x77221328 : C:\Windows\syswow64\msvcrt.dll : endthreadex
Failed to init bfd from (C:\Windows\syswow64\kernel32.dll)
0x76ce336a : C:\Windows\syswow64\kernel32.dll : BaseThreadInitThunk
Failed to init bfd from (C:\Windows\SysWOW64\ntdll.dll)
0x779a9f72 : C:\Windows\SysWOW64\ntdll.dll : RtlInitializeExceptionChain
Failed to init bfd from (C:\Windows\SysWOW64\ntdll.dll)
0x779a9f45 : C:\Windows\SysWOW64\ntdll.dll : RtlInitializeExceptionChain
C:\Users\Douglas Seager\Desktop\bfgminer-3.7.0-win32>
This feels like stack corruption. How often does it happen? Are you running the latest firmware? |
|
|
|
OK I have a problem
OP Windows 7
BFGMiner 3.7.0
BiFURY running by itself on a usb 3 hub with a fan
I can get it running and the temp is at about 48C
But after a few miner BFGMiner crashes
Can anyone help Can you post a backtrace? |
|
|
|
@Luke-Jr - Any releases with drillbit bitfury drivers coming?
I know the drillbit boards needed their own drivers aside from the bitfury drivers in CGMiner, but I would personally rather use BFGMiner as it just seems to be more stable on the RaspberryPi's, CGMiner seems to be having issues with libusb regarding timing lately, but I've never had an issue with BFGMiner.
If you need the drives being used in CGMiner let me know, happy to provide them.
I've still not had any interaction with Drillbit in a while... Most cgminer drivers are not really usable outside of cgminer, unfortunately. I could try to reverse engineer the code, but until I receive my sample unit to test with, it's difficult to write bug-free code. Plus, there's a possibility when Drillbit does finally contact me, they'll have some protocol documentation which would save a lot of time so I don't need to reverse engineer it. So overall, it makes more sense to just wait for Drillbit to follow through IMO. |
|
|
|
How are GPUs detected now? How do I pass specific settings per device via command flags? (thread-concurrency, gpu-engine, etc)
-S opencl:auto |
|
|
|
On another note, electrum has a payto command (and a paytomany) that allows you to specify the from address. I am assuming this is essentially doing what I am talking about here, right? This is a bug. There is no from address. electrum help payto Create and broadcast a transaction. Syntax: payto <recipient> <amount> [label] <recipient> can be a bitcoin address or a label options: --fee, -f: set transaction fee --fromaddr, -F: send from address - --changeaddr, -c: send change to address Are you saying that the "feature" where I specify the fromaddr above is misnamed, or something else? I assumed that by specifying the from addr, it simply would only use unspent transaction outputs that were going to that specified address, as the inputs for the new outgoing transaction. Is this incorrect? It's not a feature, it's a bug. There is no use case for selecting UTXO entries based on the transaction they were created by. |
|
|
|
On another note, electrum has a payto command (and a paytomany) that allows you to specify the from address. I am assuming this is essentially doing what I am talking about here, right? This is a bug. There is no from address. |
|
|
|
Anybody else have their average stat line disappear on ver 3.4.0? The stats for each miner are still showing but the average for both has disappeared. This happened once before so I tried to zero the stats to see if it would come back and well, that didnt work. Had to restart to get them to come back.
I've seen this, but the system with the problem didn't have a debug build I think hitting DMMM will fix it at least temporarily? Running 3.7.0 compiled using Homebrew.
The running stats log on the second half of the terminal window does not have any line breaks. It updates as one continuous string that just fills the bottom of the screen: Bug when compiled without unicode support. Already fixed in git. |
|
|
|
NEW VERSION 3.6.0, NOVEMBER 12 2013
 oops |
|
|
|
NEW VERSION 3.7.0, NOVEMBER 27 2013Human readable changelog:- Support for new Bi*fury and Twinfury mining devices.
- Device probing has been rewritten to scan all devices in parallel (much faster). As a result, the order of devices found is no longer deterministic (consistent) at all.
- Parallel device probing makes --device's former syntax unusable, so it has been modified to use the same pattern matching as --set-device. Note that things like --scan-serial ztex:all are now valid, but probably much more dangerous than with VCOM drivers; a bare driverless "all" only affects VCOM drivers for this reason.
- Pattern matching for --set-device (and now --device) supports identifying specific devices by serial number or path. To use this, insert either of these between @ symbols where you would have previously put an index. So for example, --device 'BFL@FTWOTOP4' (bitforce device with serial FTWOTOP4) or --set-device '@/dev/ttyUSB0@a:osc6_bits=52' (first processor of device on /dev/ttyUSB0).
- BFGMiner-written config files now include the scan-serial and set-device settings.
- cpu and opencl drivers have been disabled by default in all cases. Use -S cpu:auto or -S opencl:auto if you need them.
- Automatically workaround cgminer driver breakage on Linux.
Full changelog:- FTDI scan: Use intptr_t to avoid warning about different size type cast
- Bugfix: proxy: Add missing unistd.h include
- Bugfix: bifury: Precision specifications should be int, not size_t
- Bugfix: klondike: Include config.h first
- Bugfix: binloader: Include config.h first
- README.ASIC: Include support URI for bi*fury driver and firmware
- Added MultiMiner as an example of a GUI application that wraps bfgminer using the RPC interface
- Silence non-issue warnings
- bifury: Add support for hwerror message
- inc_hw_errors2: Support reporting hw errors without a known nonce
- bifury: Initialise cutofftemp to 75 C
- bifury: Turn chips off and on when disabled and enabled
- bifury: Split out bifury_send_clock
- Bugfix: bifury: Avoid double-free and double-close on version parse errors during probe
- bifury: Add support for clock command
- bifury: Update for "job" protocol change
- bifury: Update to current git
- bifury: Workaround bugs in older firmware by flooding device with work until it proves it has needwork
- bifury: Attempt to submit shares even if maxroll limit is ignored
- Clone cgminer's submit_noffset_nonce interface
- Bugfix: bifury: Check that there is a line before comparing it to "version "
- bifury: Prune jobs older than 16 queued, to eventually clean up discarded work (from flushes)
- Replace ucs2tochar* with ucs2_to_utf8* to handle Unicode
- README.ASIC: Document Bi*Fury firmware upgrade
- bifury: Free work when we know it's complete
- bifury: Implement hashrate reporting using "job" messages
- bifury: Tolerate other data before version reply
- bifury: Abstract line parsing into bifury_readln function
- Bugfix: bifury: Use a char array, rather than pointer to char array, for bifury_init_cmds
- bifury: Unify SEND devprotodump logging code
- fpgautils: Skip SetCommConfig on Windows if baud is not to be set
- Check for U+2022 and replace it with "*" is unprintable
- Use replacement character for non-printable Unicode
- Use UTF-8 for Unicode
- Implement a utf8_decode function to produce wchar_t needed by curses
- bifury: New driver
- Bugfix: twinfury: Handle flash request for both chips, and avoid using proc uninitialised
- Bugfix: twinfury: Fix various technicalities
- Bugfix: twinfury: Remove incorrect tcflush
- Twinfury communication protocol changed to be more robust. A preamble is sent before the actual command.
- LED flashing changed: - setting a flag and flash the LED during the poll cycle - mutex removed
- Renamed bf2 driver -> twinfury driver bin2hex utility used to convert the serial number
- BF2 driver comment edited
- BF2 - Twin Bitfury USB miner driver added
- Deprecate --ndevs|-n (redundant with -d? and GPU-only)
- Never abort startup just because we have no devices (there may be cases this is pointless, but at least it's consistent)
- cgpu_match: Accept ranges for letter processor specification
- cgpu_match: Accept ranges for numeric elements
- test_cgpu_match: Test with more than one device/processor
- configure: Condense configuration options summary a bit
- configure: Put configuration options summary back after everything completes
- Workaround bug in MinGW-w64: Check for strtok_r to be defined at compile-time, and don't override it unless we really want to override
- Bugfix: configure: Use new need_lowl_vcom instead of need_fpgautils
- configure: Explicitly check for sleep
- Bugfix: Set LIBUSB_LIBS for non-pkgconf non-header-subdir libusb installations
- Move unused winsock2.h includes to the end of config.h
- Use lround instead of llround, since it has better compatibility with older MingW and Cygwin versions
- lowl-hid: Save and use hid_enumerate results from library test
- Silence Windows build warnings
- mcp2210: Remove cruft leftover from HID code (moved out)
- lowl-vcom: sysfs scan: Simplify maybe_strdup into _sysfs_do_read
- lowl-vcom: sysfs scan: Don't bother getting product string until we confirm it is in fact a tty device
- input_pool: Accept a null password
- Bugfix: Move const data into the scope it needs to be in
- Bugfix: Restore case insensitivity to --scan driver names
- Accept "@" separator for serial/path in --scan to match --device and --set-device better
- README: Uppercase the sequence letter when talking about multiple pools
- README: Mention "make install" and clarify saving settings
- Bugfix: avalon: Don't "probe" with -S <path/serial>
- bitforce: Add lowl_* to bitforce_queue so it can be specified by name as a driver
- Bugfix: devpath_to_devid: Be strict about COM port specifications on Windows
- Bugfix: avalon: Don't "probe" with -S all
- Bugfix: Only count devices in -d?
- RPC: {dev,proc}details: Add 'Processors', 'Manufacturer', 'Product', 'Serial', 'Target Temperature', 'Cutoff Temperature'
- debian: Need to explicitly add opencl support to build.
- Correctly document klondike as being enabled by default
- Simplify logic for README.ASIC and README.FPGA install
- Split fpgautils into lowl-vcom and binloader, and fix internal build dependencies
- Add README.CPU to "make install" distribution when built with CPU mining support
- line 2913 added urlencode
- Bugfix: RPC: Release apisock on error in tidyup
- klondike - dont try to flush if not initialised
- Rename --scan-serial to --scan
- Bugfix: Handle invalid driver names cleanly
- Workaround libcurl bug to properly support HTTP proxies for stratum
- Bugfix: Write klondike options for klondike-options (not icarus options)
- README.OpenWrt: Change example repository to "latest", and document alternatives
- mcp2210: Add support for --device-protocol-dump
- Bugfix: Initialise dynamic osc6_bits variables
- README: Fixed a few typos here and there and improved wording (Thanks miningpenguin!)
- README: Moved CPU sections to README.CPU and elaborated more on it
- README*: Grammatical cleanups
- README.ASIC: More KnCMiner
- README.ASIC: Add section about KnCMiner
- Bugfix: opencl: Dynamic mode should disable threads for this device, not just N after the first
- Bugfix: Make curses_input return NULL on a blank line, as most callers expect
- Duplicate (hidden) --userpass option before --pool-priority so it can be used in config files
- Only probe VCOM devices on bare "all" scan-serial request
- Provide backports of uthash LL_{PREPEND,FOREACH{,_SAFE}}2 (only added in 1.9.7, not yet in major distros)
- Upgrade required uthash version to 1.9.4 for LL_CONCAT
- README: Update documentation for scan-serial
- README: Bitstreams no longer included with BFGMiner source
- README: Remove incomplete and unnecessary list of ASIC/FPGA devices
- write_config: Add scan-serial and set-device
- write_config: Update to make JSON Array of named parameters
- Remove support for comma-separated values in --device, and update documentation
- lowlevel: Collapse multiple lowlevel drivers on the same devid to a secondary linked list, and only spawn one thread to probe all lowlevels on a given devid
- Bugfix: vcom_devinfo_scan: Clean up hashtable when done
- README: Update cgminer driver FAQ with Windows-specific answer, now that we have a workaround for *nix
- cpu/opencl: Always disable by default; removes deprecated -G option entirely
- vcom_lowl_probe_wrapper: Attempt to reattach detached USB devices
- Bugfix: erupter: Prioritise Emerald after Sapphire
- Bugfix: Allow "drv:auto" to override general "noauto"
- Bugfix: _probe_device_match: Fix inverted logic
- Bugfix: Don't try to call lowl_probe for drivers that don't support that interface
- lowl-hid: Use usb devids for libusb-based hidapi devices
- Replace bfg_claim_hid with lowlevel_claim
- lowlevel: Generic claim interface for lowlevel info
- cgpu_match: Compare dev_t under paths
- Rework -d? output to display more useful information now that device ordering is non-deterministic
- cgpu_match: Add support for matching by serial number or path
- Unit tests for cgpu_match
- Rework --device option to accept a pattern and JSON Array in config; removes support for processor ranges
- Remove --remove-disabled option, as it conflicts with hotplug, per-processor stuff, etc already
- Bugfix: Reimplement "noauto" flag
- Bugfix: Restore "-S foo" without driver name functionality
- Bugfix: scan-serial: Compare dev_t as well
- littlefury: Convert to lowl_probe
- hashbuster: Convert to lowl_probe
- avalon: Convert to lowl_probe
- cairnsmore/erupter/icarus: Convert to lowl_probe
- bigpic: Convert to lowl_probe
- ztex: Convert to lowl_probe
- x6500: Convert to lowl_probe
- nanofury: Convert to lowl_probe
- modminer: Convert to lowl_probe
- klondike: Convert to lowl_probe
- bitforce: Convert to lowl_probe
- fpgautils: Wrapper to easily adapt an old detectone to new lowl_probe
- lowlevel: Add match variants of detect functions
- DevAPI: add_cgpu: Use mutex to ensure thread safety
- Make detectone_meta_info thread-local
- New threaded device probe framework
|
|
|
|
Is the hardware broken? BFL 0: 70.0C | 55.13/56.72/33.15Gh/s | A:1691 R:35+1238( 40%) HW:141/.60% I tried multiple pools. Sounds like a bad internet connection. |
|
|
|
|
The point is to let you micromanage what coins (UTXO entries) you consume in a given transaction.
There is no such thing as a "from address".
|
|
|
|
Well, I can't wait for the next version to come out. I have one of the bi-furies and I can't for the life of me get it recognized and running. I've tried compiling BFGMiner from source and CGMiner from source on Ubuntu. I've tried using precompiled binaries of CGMiner on OS X and I've gotten no love. Mostly it's been unrecognized but when it has been recognized as a bitfury device it hasn't initialized. It can just sit on it's hot little ass until I have more time to screw with it or 3.7 comes out. There's no doubt though, this thing needs a fan. It gets HOT just being plugged in.  Not sure when(if?) they'll be publishing the Windows INF file, but here's one that should work. Edit: But do note that earlier Bi*furys shipped with buggy firmware and might not work very well until you upgrade the firmware (see BFGMiner README.ASIC for instructions. |
|
|
|
|
Show Posts
