1
0
mirror of https://github.com/m00natic/vlfi.git synced 2025-04-14 14:50:19 +01:00

11 Commits

Author SHA1 Message Date
Andrey Kotlarski
11c7af4b04 Change handling of measurement values to support approximations. 2014-09-07 00:02:14 +03:00
Andrey Kotlarski
5651ee3d61 Rename vlf-tune-optimal to vlf-tune-batch. 2014-09-06 23:59:32 +03:00
Andrey Kotlarski
f3212ec9a6 Fix measure approximation and allow tuning to just over half file
batch.
2014-09-05 19:11:48 +03:00
Andrey Kotlarski
48a014f3bc Fix write measuring and endless loop in nearby approximation. 2014-09-05 14:44:24 +03:00
Andrey Kotlarski
facdb9f6bc Fix binary tune base case and add approximation after access to
previously queried measure that is still missing.
2014-09-05 13:52:46 +03:00
Andrey Kotlarski
e18a05b7cb Add linear search for tuning and prefer smaller batches. 2014-09-05 02:49:55 +03:00
Andrey Kotlarski
d85f3d43fc Fix access to uninitialized measurements when tuning. 2014-09-04 15:43:37 +03:00
Andrey Kotlarski
d67825c4cd Move back to using average speed when measuring and tuning. Be more
precise when choosing index for measurement.
2014-09-04 15:25:44 +03:00
Andrey Kotlarski
5379943cd7 Save times instead of speeds, compare on cumulative speed when tuning. 2014-09-04 03:33:05 +03:00
Andrey Kotlarski
fb0503064d Add basic tune strategies. 2014-09-03 02:35:20 +03:00
Andrey Kotlarski
32ff2cb067 Add bookkeeping module. 2014-08-31 02:10:28 +03:00