60%ish hardware errors seem to be about normal
but...
[2013-11-05 09:03:46] BPM0 | 5s: 2.31 avg: 2.30 u: 0.54 Gh/s | A:66 R:0+0(none) HW:2252/ 89%
[2013-11-05 09:03:46] BPM1 | 5s: 1.99 avg: 1.99 u: 1.89 Gh/s | A:237 R:1+1(.84%) HW:1398/ 59%
[2013-11-05 09:03:46] BPM2 | 5s: 2.29 avg: 2.29 u: 2.16 Gh/s | A:287 R:2+0(.52%) HW:1607/ 59%
[2013-11-05 09:03:46] BPM3 | 5s: 2.21 avg: 2.21 u: 2.05 Gh/s | A:261 R:2+0(.76%) HW:1567/ 60%
seems like I got a couple duds.
@zyk0 & @ssinc
What OS? As I've seen this before but after I recompiled BFG using 3.2.0, I haven't seen the issue since, so, I'm curious what OS you're using these on.
Was originally using a custom cgminer 3.6.6-1 w/ win7 x64 after debugging with ck and getting him to remove the manufacture string check, but it's speeds were horrible and inconsistent, I instead switched to bfgminer under a ubuntu server 13.10 VM in VirtualBox w/ USB passthrough from the win64 host OS. Originally I was using the 3.0.99 repository compiled binary, but have switched to 3.4.0. I have also tried cgminer 3.7.0 and 3.7.2 win32 binaries but they are still not up to snuff.
1/2 the time w/ bfgminer 3.0.99 I'll see random garbage as part of the model name/serial, and I've seen corrupted output on the device name from cgminer's API stats as well.
I'm hoping these are issues with only the firmware on these blue furies and we'll be getting a fix soon.
Sadly... my stick which now is getting more HW errors than the others, used to be my top performing Blue Fury:
[2013-11-05 07:17:33] BF10 BF1 15062915 | 5s:1.996 avg:1.993 u:1.895 Gh/s | A:8702 R:62+0(.70%) HW:0
[2013-11-05 07:17:33] BF11 BF1 15062B15 | 5s:2.302 avg:2.292 u:2.157 Gh/s | A:9907 R:78+2(.80%) HW:0
[2013-11-05 07:17:33] BF12 BF1 15040915 | 5s:2.314 avg:2.305 u:2.200 Gh/s | A:10102 R:66+0(.64%) HW:0
[2013-11-05 07:17:33] BF13 BF1 15062312 | 5s:2.215 avg:2.214 u:2.133 Gh/s | A:9795 R:72+1(.73%) HW:0
But serial no. 15040915 is now my worst one, and no amount of resetting, power cycling, cooling off, leaving unplugged for a while, etc.. is getting it back to where it was.
Edit:
Maybe it needed to "warm up", seems to have finally kicked in to normal after starting with a high HW error rate (unless the difficult change triggered it)
Block: ...2df02926 #268145 Diff:511M ( 3.66Ph/s) Started: [13:39:08]
ST:2 F:0 NB:33 AS:0 BW:[ 62/ 55 B/s] E:33.63 I: 341uBTC/hr BS:85.6k
4 | 8.92/ 8.79/ 8.29Gh/s | A:6405 R:63+0(.97%) HW:37790/ 59%
--------------------------------------------------------------------------------
BPM 0: | 1.99/ 1.99/ 1.90Gh/s | A:1515 R:13+0(.85%) HW: 8501/ 59% <- still 10% slower than most
BPM 1: | 2.28/ 2.29/ 2.17Gh/s | A:1627 R:13+0(.79%) HW: 9817/ 59%
BPM 2: | 2.30/ 2.30/ 2.10Gh/s | A:1587 R:17+0(1.1%) HW:10027/ 61% <- back to working
BPM 3: | 2.22/ 2.21/ 2.13Gh/s | A:1677 R:20+0(1.2%) HW: 9450/ 59%
--------------------------------------------------------------------------------