|
ed40f4ba
|
2013-10-02T03:49:47
|
|
Enable cores back aftre 15 minutes in disabled state
|
|
|
a74236f7
|
2013-10-01T20:58:53
|
|
Remove unused parameter in two inline functions
|
|
|
fa3eb55e
|
2013-09-27T13:58:35
|
|
Disable core if it produces too many HW errors
|
|
|
8b28e009
|
2013-08-28T11:50:49
|
|
Remove stale works from internal queue
(stale = no "work completed" response from FPGA in 5 seconds)
|
|
|
dafaace2
|
2013-08-28T10:19:40
|
|
Add some debug messages
|
|
|
8493b064
|
2013-09-26T12:59:33
|
|
New FPGA SPI response format - 3 words, core id added
|
|
|
addc136e
|
2013-10-01T19:41:55
|
|
More reasonable "stale work" timeout
|
|
|
7ee5e47b
|
2013-09-05T18:08:57
|
|
Flush FPGA at start
|
|
|
67cb850d
|
2013-08-28T12:23:27
|
|
Add some profiling code
|
|
|
ec03d194
|
2013-08-28T10:18:48
|
|
New KnC driver: knc-spi-fpga
|
|
|
d076d70a
|
2013-08-22T16:02:02
|
|
Performance improvement.
Eliminates all performance bottlenecks which can be easily removed.
The next obstacle is SHA256 hash calculation itself, which is much
harder to improve.
|
|
|
bbe8c88c
|
2013-08-09T10:28:17
|
|
'llround' is more suitable here than 'roundl'
|
|
90bf8f5e
|
2013-06-09T16:35:02
|
|
Silence warning if MAX/MIN is already defined
|
|
|
4cbd05f8
|
2013-05-27T16:27:25
|
|
Remove prebuild ccan/opt objects
|
|
70bc62c0
|
2013-08-21T21:16:03
|
|
Bump version number to 3.4.0
|
|
|
fffdea09
|
2013-08-21T21:15:27
|
|
Update NEWS.
|
|
|
c56d44ce
|
2013-08-21T21:00:57
|
|
Use stack data for HW error% in avalon stats.
|
|
|
8a01f532
|
2013-08-21T20:46:07
|
|
Add avalon HW error% to stats and only show BTB variables if avalon is a BTB.
|
|
|
b9aa5b65
|
2013-08-21T20:31:10
|
|
Check for cnx_needed on each loop through wait_lp_current.
|
|
|
919a836f
|
2013-08-21T20:29:46
|
|
Return positive for cnx_needed when no_work is true.
|
|
|
7dbe4bbe
|
2013-08-21T20:12:00
|
|
Stratum is used more often so test for it first.
|
|
|
a52ec336
|
2013-08-21T20:09:17
|
|
Reorder support names alphabetically.
|
|
|
77f9400c
|
2013-08-21T12:15:40
|
|
Only display the no pool work message once if there are multiple waiters in hash_pop
|
|
|
20de9226
|
2013-08-21T11:21:13
|
|
Provide a message and set a bool when no work is available from any pools and when it resumes again.
|
|
|
0e43084c
|
2013-08-21T11:05:21
|
|
We don't want to continue into the hash_pop function if the getq is frozen.
|
|
|
cbea0a53
|
2013-08-21T10:56:31
|
|
Only report threads in and out in queued work devices across a get work since the rest happens asynchronously and the get work is what the device might be waiting on.
|
|
|
b12badc9
|
2013-08-21T10:47:32
|
|
Thread reportin and out can be static non inline.
|
|
10545d33
|
2013-08-20T17:37:06
|
|
Merge pull request #480 from kanoi/master
usbutils cps sleep_estimate is not an underestimate
|
|
612c0b12
|
2013-08-21T10:02:12
|
|
usbutils cps sleep_estimate is not an underestimate
|
|
|
ff8c8325
|
2013-08-20T07:08:06
|
|
Merge pull request #479 from kanoi/master
usbutils add cps stats estimates
|
|
|
01855a60
|
2013-08-21T00:06:05
|
|
usbutils add cps stats estimates
|
|
|
903d007c
|
2013-08-20T22:47:13
|
|
Provide cgtimer_sub helper functions.
|
|
|
9dc63cee
|
2013-08-20T22:35:38
|
|
Provide cgtimer_to_ms helper functions.
|
|
|
7d0c9b50
|
2013-08-20T22:27:14
|
|
Rename cgsleep_prepare_r as cgtimer_time to get time in cgtimer_t format and call cgsleep_prepare_r as a macro for cgtimer_time
|
|
|
deb7be72
|
2013-08-20T21:07:38
|
|
Use the reentrant cgsleep functions for usecps in usbutils.
|
|
|
ee2a5ae8
|
2013-08-20T20:46:09
|
|
TimeBeginPeriod and TimeEndPeriod do not add significant overhead when run the entire time for cgminer so avoid trying to maintain balanced numbers of them for specific time calls to simplify code.
|
|
|
eeac9d81
|
2013-08-20T19:23:49
|
|
Replace all references to the old n*sleep functions with the equivalent cgsleep_*s replacements.
|
|
|
7d448cd7
|
2013-08-19T23:39:49
|
|
timeGetTime uses huge resources on windows so revert to using timevals for its implementation of cgtimer_t
|
|
|
bedd8eee
|
2013-08-19T21:52:07
|
|
Revert "Sleep for half a work period only in bflsc_get_results if no dev is suitable."
This reverts commit 974319454795979dad013c705ca9d53aae67668d.
Fine now that sleep works properly with the longer sleep.
|
|
|
0414bf09
|
2013-08-19T21:36:00
|
|
Quotient/remainder error in ms division.
|
|
|
fb66e8b7
|
2013-08-19T21:13:23
|
|
Only grab a queued work item if we successfully grab the lock to submit work in bflsc_send_work
|
|
|
39f4cf05
|
2013-08-19T03:39:58
|
|
Merge pull request #478 from kanoi/btb
BTB get version from Firmware
|
|
|
96c0c317
|
2013-08-19T20:34:37
|
|
BTB get version from Firmware
|
|
|
a76bc0e0
|
2013-08-19T19:57:50
|
|
Carve out the unused portions of sha2 implementation.
|
|
|
78d78b10
|
2013-08-19T19:10:41
|
|
Import Aaron D. Gifford's fast sha256 implementation.
|
|
|
2a78cd36
|
2013-08-19T18:08:17
|
|
Increase the que_low watermarks on BFLSC for they are too low to keep the device busy on scanwork loops.
|
|
|
dddc2d1e
|
2013-08-19T17:59:59
|
|
Revert "Reuse times extracted from the clock in bflsc_get_results."
This reverts commit 1d8235334d0bac696e91dc246af2aab9f1d35d36.
Monotonic clock does not return the same times as gettimeofday.
|
|
|
1d823533
|
2013-08-19T17:55:32
|
|
Reuse times extracted from the clock in bflsc_get_results.
|
|
|
b238139c
|
2013-08-19T17:44:45
|
|
Provide cgtimer_to_timeval helper functions.
|
|
|
09679e9f
|
2013-08-19T17:38:20
|
|
Provide a timeval_to_cgtime helper function to reuse values.
|
|
|
97431945
|
2013-08-19T16:36:01
|
|
Sleep for half a work period only in bflsc_get_results if no dev is suitable.
|
|
|
a61e41a0
|
2013-08-19T15:58:22
|
|
Check for thr->work_restart in restart_wait.
|
|
|
89f571f5
|
2013-08-19T15:51:14
|
|
We should be using que_low to decrease scan sleep time in bflsc.
|
|
|
09dbd616
|
2013-08-19T13:57:36
|
|
Prepare sleep time on bflsc if no dev needs work yet to avoid busy waiting.
|
|
|
09e9091d
|
2013-08-19T13:54:56
|
|
Simplify cgsleep code for windows by using a typedef for cgtimer_t that resolves to clock resolution, using that internally.
|
|
|
404cbdea
|
2013-08-19T13:29:23
|
|
On windows use the higher accuracy timegettime function to really get 1ms clock and timer accuracy.
|
|
|
e56c9a5d
|
2013-08-18T16:49:15
|
|
Use the cgsleep reentrant function to sleep for bflsc between read results to account for time taken to perform reads.
|
|
|
3c707202
|
2013-08-18T16:46:27
|
|
Use 100ms delay between checking for results on all bflsc devices as the buffering of results mean checking more frequently just wastes CPU and causes more lock contention for only marginally better latencies.
|
|
|
ffa12a13
|
2013-08-18T14:35:01
|
|
Fix missed endtimeperiod in overrun timer on windows.
|
|
|
ad87db07
|
2013-08-18T14:30:54
|
|
Make cgsleep_us_r take an int64_t for us.
|
|
|
703c7309
|
2013-08-18T14:26:41
|
|
Make the cgsleep functions build on windows.
|
|
|
17447ec7
|
2013-08-18T11:59:02
|
|
Use the cgsleep reentrant function in avalon_send_task.
|
|
|
c988587b
|
2013-08-18T11:48:42
|
|
Use the reentrant cgsleep functions within the avalon_send_tasks function.
|
|
|
81144732
|
2013-08-18T11:39:09
|
|
Set high resolution timing on windows within the cgsleep functions.
|
|
|
e784b230
|
2013-08-18T11:02:05
|
|
Use the reentrant cgsleep function to time sleeps on reading from avalon.
|
|
|
621eb003
|
2013-08-18T10:49:52
|
|
Provide reentrant versions of cgsleep functions to allow start time to be set separately from the beginning of the actual sleep, allowing scheduling delays to be counted in the sleep.
|
|
|
0b5edb24
|
2013-08-18T10:36:25
|
|
Make the nmsleep and nusleep functions use the new cgsleep functions internally till functions are migrated to the new cgsleep API.
|
|
|
188210a3
|
2013-08-18T10:28:28
|
|
Add a ms_to_timespec helper function, and create a cgsleep_ms function that uses absolute timers with clock_nanosleep to avoid overruns.
|
|
|
f3b75b0b
|
2013-08-18T01:39:35
|
|
Add rt lib linkage to enable use of clock_nanosleep functions with older glibc.
|
|
|
37ff9f84
|
2013-08-18T01:33:50
|
|
Add necessary time header include to avalon driver.
|
|
|
1864ed26
|
2013-08-18T00:47:05
|
|
Do a sleep of the full duration it would take to do all the work using clock_nanosleep in avalon_send_tasks to avoid sleep overruns before polling to see if it's ready.
|
|
|
fe7c8e82
|
2013-08-18T00:39:13
|
|
Merge branch 'master' of github.com:ckolivas/cgminer
|
|
|
3da121ba
|
2013-08-17T07:39:00
|
|
Merge pull request #477 from someone42/master
Changes to queueing strategy for BitBurner boards
|
|
|
99cbf09d
|
2013-08-18T00:38:29
|
|
Add a timeraddspec helper function.
|
|
|
1bcd9a43
|
2013-08-18T00:34:01
|
|
Provide a us_to_timespec helper function.
|
|
|
33bb8c79
|
2013-08-18T00:28:30
|
|
Use the us_to_timeval helper function in the avalon driver.
|
|
|
2278204d
|
2013-08-18T00:27:33
|
|
Provide a us_to_timeval helper function.
|
|
|
53beae81
|
2013-08-18T00:21:01
|
|
Use timeval_to_spec helper in avalon driver.
|
|
|
a6b1c317
|
2013-08-18T00:19:47
|
|
Add helper functions to convert timespec to timeval and vice versa.
|
|
26de6e64
|
2013-08-17T16:18:12
|
|
simplifying buffer full check
|
|
|
0b9474f3
|
2013-08-17T15:36:47
|
|
forking bitburner write thread function
|
|
|
be0d899b
|
2013-08-16T17:00:56
|
|
making sure original Avalon is unaffected by BitBurner changes
|
|
|
2b932ccb
|
2013-08-16T16:37:43
|
|
changes to queueing strategy for BitBurner boards
|
|
|
1bb6dd40
|
2013-08-16T13:57:54
|
|
Merge branch 'master' of github.com:ckolivas/cgminer
|
|
|
f3f939dd
|
2013-08-16T13:57:07
|
|
Do not poll in avalon_get_results without sleeping if we have finished parsing a full result.
|
|
|
f3837916
|
2013-08-15T17:14:16
|
|
Merge pull request #476 from kanoi/master
avalon for timeout allow d='calculate it' and fix uninitialised
|
|
|
61cb05a9
|
2013-08-16T09:55:45
|
|
Add c to ambient temperature display for avalon driver.
|
|
|
356fcb10
|
2013-08-15T12:00:04
|
|
BTB allow up to 1400mV as per firmware limits
|
|
|
51aaf816
|
2013-08-15T10:57:02
|
|
avalon for timeout allow d='calculate it' and fix uninitialised
|
|
|
3fe89486
|
2013-08-14T23:04:10
|
|
Use cloned work when finding avalon results since another thread can discard the work item while it's in use.
|
|
|
98e338cc
|
2013-08-14T22:59:57
|
|
Provide a variant of find_work_bymidstate that returns a clone of the found work.
|
|
|
853f145e
|
2013-08-14T10:49:40
|
|
Bump version number to 3.3.4
|
|
|
730d0626
|
2013-08-14T10:49:16
|
|
Update NEWS.
|
|
|
c9f58230
|
2013-08-13T17:45:39
|
|
Merge pull request #475 from kanoi/master
API/miner.php add some % fields
|
|
|
6b62c402
|
2013-08-14T10:42:52
|
|
API/miner.php add some % fields
|
|
|
e95ad047
|
2013-08-13T22:44:22
|
|
Nonce2 stratum submission is not working with nonce2 lengths >4, revert the buggy __bin2hex function and use bin2hex.
|
|
|
dd9233f7
|
2013-08-13T17:24:01
|
|
The write thread in avalon is only ever actually woken up by timeout so remove the write semaphore and use a simple sleep poll.
|
|
|
b2a3941e
|
2013-08-13T15:36:52
|
|
Fix warning.
|
|
|
992a97a3
|
2013-08-13T14:52:17
|
|
Interrupting reads on the avalon to start writes loses data so remove the cgsem_post in the read code.
|
|
|
1d614599
|
2013-08-13T14:35:35
|
|
Add room for the null byte at the end of the nonce2 string on stratum share submission and zero the allocated ram.
|