Bug 216866 - math tgammal(3) uses a poor algorithm due to clang kludge
Summary: math tgammal(3) uses a poor algorithm due to clang kludge
Status: Closed FIXED
Alias: None
Product: Base System
Classification: Unclassified
Component: standards (show other bugs)
Version: CURRENT
Hardware: Any Any
: --- Affects Many People
Assignee: Mark Murray
URL: https://reviews.freebsd.org/D33444
Keywords:
Depends on:
Blocks:
 
Reported: 2017-02-06 19:56 UTC by sgk
Modified: 2021-12-15 18:42 UTC (History)
3 users (show)

See Also:

Attachments
diff for tgamma() (25.34 KB, patch)
2021-12-13 20:56 UTC, Steve Kargl 		no flags Details | Diff
description of patch1.diff (2.18 KB, text/plain)
2021-12-13 20:57 UTC, Steve Kargl 		no flags Details
implementation of ld80 tgammal(x) (28.15 KB, patch)
2021-12-14 05:20 UTC, Steve Kargl 		no flags Details | Diff
description for patch2.diff (584 bytes, text/plain)
2021-12-14 05:21 UTC, Steve Kargl 		no flags Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description sgk 2017-02-06 19:56:30 UTC 
The algorithm used by FreeBSD libm for tgammal severely reduces both
its precision and domain.  See msun/src/imprecise.c.
Comment 1 Steve Kargl freebsd_committer freebsd_triage 2021-12-13 02:43:37 UTC 
Well, it's taken me a few years, but I finally have fixed
the theraven@freebsd.org's questionalble mapping of tgammal(x)
to tgamma(x) on at least Intel 80-bit long double targets.
Watch this space.  Patches will be arriving shortly.
Comment 2 Mark Murray freebsd_committer freebsd_triage 2021-12-13 09:03:05 UTC 
I have an interest in tgamma(3), so I'll take this for committing. Thanks, Steve!
Comment 3 Steve Kargl freebsd_committer freebsd_triage 2021-12-13 20:56:12 UTC 
Created attachment 230088 [details]
diff for tgamma()

This is the diff to tgamma(x) (the double function) that I did while reverse engineering the algorithms to write tgammal(x) (the long double function).
Comment 4 Steve Kargl freebsd_committer freebsd_triage 2021-12-13 20:57:20 UTC 
Created attachment 230089 [details]
description of patch1.diff

Here's a brief list of changes.
Comment 5 Steve Kargl freebsd_committer freebsd_triage 2021-12-14 05:20:44 UTC 
Created attachment 230101 [details]
implementation of ld80 tgammal(x)

This is the patch that implements ld80 tgammal(x).  It should be applied after patch1.diff.
Comment 6 Steve Kargl freebsd_committer freebsd_triage 2021-12-14 05:21:31 UTC 
Created attachment 230102 [details]
description for patch2.diff

This is brief description of patch2.diff.
Comment 7 Mark Murray freebsd_committer freebsd_triage 2021-12-15 18:42:42 UTC 
Patches accepted, thank you!
https://reviews.freebsd.org/D33444