| Commit | Date | |
|---|---|---|
|
[r9621]
by
dwarren
Fixed memory error when trying to access compl frame to reset level_num when there is no frame. |
2019-08-19 14:01:20 | Tree |
|
[r9620]
by
tswift
Moved num_sched and friends from schedrev_xsb_i.h where they were defined inside a function (emuloop) to emuloop.c where they were defined globally. This means that they can be used when expanding the compstack. |
2019-08-12 23:17:17 | Tree |
| 2019-08-12 23:15:26 | Tree | |
|
[r9618]
by
"kifer"
two new preds: machine:epoch_nanoseconds(Secs,Nanosecs) This one actually works in windows, but windoze resolution is 100ns, while linux mac is 1 ns. This is a windoze limitation for some reason. file_io:file_time_ns(file,secs,nanosecs) This one only works in linux and mac. In windows, nanosecs=0. If somebody has the patience to figure the windoze way for getting nanosecs here, it'd be great. I tried but could not get the full picture and gave up. There is this thing: https://docs.microsoft.com/en-us/windows/win32/sysinfo/retrieving-the-last-write-time but unclear how to get from there to sec/nsec since epoch (1970). For some reason, some funcs in windows use 1601 as epoch and some 1970. |
2019-07-28 02:46:12 | Tree |
|
[r9617]
by
dwarren
Fix bug in removing least element from sched heap; also fix bug in scheduling for incremental table maintence where there was a missing add-to-shced-heap. |
2019-07-25 14:47:18 | Tree |
|
[r9616]
by
dwarren
Changed scheduling in fixpoint to use a priority queue. Also made some minimal changes in tst_insert.c to (maybe) make it easier to add iteration to the skip list impl. |
2019-07-24 20:57:53 | Tree |
|
[r9615]
by
dwarren
Modified answer completion to alternate RL and LR selection strategy on meta-interpretation iterations. Makes more likely to be linear, esp. on artifical examples. |
2019-07-10 15:56:11 | Tree |
|
[r9614]
by
dwarren
Added tech documentation of new adjust_levels algorithm and new fixpoint scheduling algorithm. |
2019-07-10 15:51:02 | Tree |
|
[r9613]
by
dwarren
Modified fixpoint scheduling to use a scheduling queue to avoid linear scan of completion frames. Updated adjust_levels to be more multi-thread friendly, I hope. |
2019-07-10 15:47:49 | Tree |
| 2019-06-24 18:17:19 | Tree |