Con Kolivas
|
d8e2a43712
|
Abstract out the conversion of system time to an lldiv_t in decimicroseconds.
|
12 years ago |
Con Kolivas
|
acc1b7df1b
|
Use our own gettimeofday implementation on windows for it to be consistent across ming builds and higher resolution.
|
12 years ago |
Con Kolivas
|
70bc62c0f8
|
Bump version number to 3.4.0
|
12 years ago |
Con Kolivas
|
fffdea0949
|
Update NEWS.
|
12 years ago |
Con Kolivas
|
c56d44cebd
|
Use stack data for HW error% in avalon stats.
|
12 years ago |
Con Kolivas
|
8a01f5326c
|
Add avalon HW error% to stats and only show BTB variables if avalon is a BTB.
|
12 years ago |
Con Kolivas
|
b9aa5b6528
|
Check for cnx_needed on each loop through wait_lp_current.
|
12 years ago |
Con Kolivas
|
919a836f14
|
Return positive for cnx_needed when no_work is true.
|
12 years ago |
Con Kolivas
|
7dbe4bbeb3
|
Stratum is used more often so test for it first.
|
12 years ago |
Con Kolivas
|
a52ec336bf
|
Reorder support names alphabetically.
|
12 years ago |
ckolivas
|
77f9400ca2
|
Only display the no pool work message once if there are multiple waiters in hash_pop
|
12 years ago |
ckolivas
|
20de9226fd
|
Provide a message and set a bool when no work is available from any pools and when it resumes again.
|
12 years ago |
ckolivas
|
0e43084c09
|
We don't want to continue into the hash_pop function if the getq is frozen.
|
12 years ago |
ckolivas
|
cbea0a5306
|
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.
|
12 years ago |
ckolivas
|
b12badc9f2
|
Thread reportin and out can be static non inline.
|
12 years ago |
kanoi
|
10545d33c7
|
Merge pull request #480 from kanoi/master
usbutils cps sleep_estimate is not an underestimate
|
12 years ago |
Kano
|
612c0b12c8
|
usbutils cps sleep_estimate is not an underestimate
|
12 years ago |
kanoi
|
ff8c832534
|
Merge pull request #479 from kanoi/master
usbutils add cps stats estimates
|
12 years ago |
Kano
|
01855a6051
|
usbutils add cps stats estimates
|
12 years ago |
Con Kolivas
|
903d007cb4
|
Provide cgtimer_sub helper functions.
|
12 years ago |
Con Kolivas
|
9dc63cee53
|
Provide cgtimer_to_ms helper functions.
|
12 years ago |
Con Kolivas
|
7d0c9b5060
|
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
|
12 years ago |
Con Kolivas
|
deb7be72c6
|
Use the reentrant cgsleep functions for usecps in usbutils.
|
12 years ago |
Con Kolivas
|
ee2a5ae8f4
|
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.
|
12 years ago |
Con Kolivas
|
eeac9d81fc
|
Replace all references to the old n*sleep functions with the equivalent cgsleep_*s replacements.
|
12 years ago |
Con Kolivas
|
7d448cd754
|
timeGetTime uses huge resources on windows so revert to using timevals for its implementation of cgtimer_t
|
12 years ago |
Con Kolivas
|
bedd8eeead
|
Revert "Sleep for half a work period only in bflsc_get_results if no dev is suitable."
This reverts commit 9743194547 .
Fine now that sleep works properly with the longer sleep.
|
12 years ago |
Con Kolivas
|
0414bf0960
|
Quotient/remainder error in ms division.
|
12 years ago |
Con Kolivas
|
fb66e8b7b7
|
Only grab a queued work item if we successfully grab the lock to submit work in bflsc_send_work
|
12 years ago |
kanoi
|
39f4cf0557
|
Merge pull request #478 from kanoi/btb
BTB get version from Firmware
|
12 years ago |
Kano
|
96c0c31775
|
BTB get version from Firmware
|
12 years ago |
Con Kolivas
|
a76bc0e043
|
Carve out the unused portions of sha2 implementation.
|
12 years ago |
Con Kolivas
|
78d78b1063
|
Import Aaron D. Gifford's fast sha256 implementation.
|
12 years ago |
Con Kolivas
|
2a78cd36cc
|
Increase the que_low watermarks on BFLSC for they are too low to keep the device busy on scanwork loops.
|
12 years ago |
Con Kolivas
|
dddc2d1e60
|
Revert "Reuse times extracted from the clock in bflsc_get_results."
This reverts commit 1d8235334d .
Monotonic clock does not return the same times as gettimeofday.
|
12 years ago |
Con Kolivas
|
1d8235334d
|
Reuse times extracted from the clock in bflsc_get_results.
|
12 years ago |
Con Kolivas
|
b238139cf9
|
Provide cgtimer_to_timeval helper functions.
|
12 years ago |
Con Kolivas
|
09679e9f8b
|
Provide a timeval_to_cgtime helper function to reuse values.
|
12 years ago |
Con Kolivas
|
9743194547
|
Sleep for half a work period only in bflsc_get_results if no dev is suitable.
|
12 years ago |
Con Kolivas
|
a61e41a070
|
Check for thr->work_restart in restart_wait.
|
12 years ago |
Con Kolivas
|
89f571f527
|
We should be using que_low to decrease scan sleep time in bflsc.
|
12 years ago |
Con Kolivas
|
09dbd616e2
|
Prepare sleep time on bflsc if no dev needs work yet to avoid busy waiting.
|
12 years ago |
Con Kolivas
|
09e9091dd7
|
Simplify cgsleep code for windows by using a typedef for cgtimer_t that resolves to clock resolution, using that internally.
|
12 years ago |
Con Kolivas
|
404cbdea84
|
On windows use the higher accuracy timegettime function to really get 1ms clock and timer accuracy.
|
12 years ago |
Con Kolivas
|
e56c9a5d3b
|
Use the cgsleep reentrant function to sleep for bflsc between read results to account for time taken to perform reads.
|
12 years ago |
Con Kolivas
|
3c707202cf
|
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.
|
12 years ago |
Con Kolivas
|
ffa12a138a
|
Fix missed endtimeperiod in overrun timer on windows.
|
12 years ago |
Con Kolivas
|
ad87db07d9
|
Make cgsleep_us_r take an int64_t for us.
|
12 years ago |
Con Kolivas
|
703c7309c2
|
Make the cgsleep functions build on windows.
|
12 years ago |
Con Kolivas
|
17447ec7e2
|
Use the cgsleep reentrant function in avalon_send_task.
|
12 years ago |