/[gentoo-x86]/dev-haskell/hashtables/metadata.xml
Gentoo

Contents of /dev-haskell/hashtables/metadata.xml

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1.5 - (show annotations) (download) (as text)
Sat Dec 13 13:46:50 2014 UTC (13 days, 8 hours ago) by gienah
Branch: MAIN
CVS Tags: HEAD
Changes since 1.4: +4 -3 lines
File MIME type: application/xml
Bump hashtables to 1.2.0.1

(Portage version: 2.2.15/cvs/Linux x86_64, signed Manifest commit with key 618E971F)

1 <?xml version="1.0" encoding="UTF-8"?>
2 <!DOCTYPE pkgmetadata SYSTEM "http://www.gentoo.org/dtd/metadata.dtd">
3 <pkgmetadata>
4 <herd>haskell</herd>
5 <herd>proxy-maintainers</herd>
6 <maintainer>
7 <email>haskell@gentoo.org</email>
8 </maintainer>
9 <longdescription>
10 This package provides a couple of different implementations of mutable hash
11 tables in the ST monad, as well as a typeclass abstracting their common
12 operations, and a set of wrappers to use the hash tables in the IO monad.
13
14 /QUICK START/: documentation for the hash table operations is provided in the
15 &quot;Data.HashTable.Class&quot; module, and the IO wrappers (which most users will
16 probably prefer) are located in the &quot;Data.HashTable.IO&quot; module.
17
18 This package currently contains three hash table implementations:
19
20 1. &quot;Data.HashTable.ST.Basic&quot; contains a basic open-addressing hash table
21 using linear probing as the collision strategy. On a pure speed basis it
22 should currently be the fastest available Haskell hash table
23 implementation for lookups, although it has a higher memory overhead
24 than the other tables and can suffer from long delays when the table is
25 resized because all of the elements in the table need to be rehashed.
26
27 2. &quot;Data.HashTable.ST.Cuckoo&quot; contains an implementation of \&quot;cuckoo
28 hashing\&quot; as introduced by Pagh and Rodler in 2001 (see
29 &lt;http://en.wikipedia.org/wiki/Cuckoo_hashing&gt;). Cuckoo hashing has
30 worst-case /O(1)/ lookups and can reach a high \&quot;load factor\&quot;, in which
31 the table can perform acceptably well even when more than 90% full.
32 Randomized testing shows this implementation of cuckoo hashing to be
33 slightly faster on insert and slightly slower on lookup than
34 &quot;Data.Hashtable.ST.Basic&quot;, while being more space efficient by about a
35 half-word per key-value mapping. Cuckoo hashing, like the basic hash
36 table implementation using linear probing, can suffer from long delays
37 when the table is resized.
38
39 3. &quot;Data.HashTable.ST.Linear&quot; contains a linear hash table (see
40 &lt;http://en.wikipedia.org/wiki/Linear_hashing&gt;), which trades some insert
41 and lookup performance for higher space efficiency and much shorter
42 delays when expanding the table. In most cases, benchmarks show this
43 table to be currently slightly faster than @Data.HashTable@ from the
44 Haskell base library.
45
46 It is recommended to create a concrete type alias in your code when using this
47 package, i.e.:
48
49 &gt; import qualified Data.HashTable.IO as H
50 &gt;
51 &gt; type HashTable k v = H.BasicHashTable k v
52 &gt;
53 &gt; foo :: IO (HashTable Int Int)
54 &gt; foo = do
55 &gt; ht &lt;- H.new
56 &gt; H.insert ht 1 1
57 &gt; return ht
58
59 Firstly, this makes it easy to switch to a different hash table implementation,
60 and secondly, using a concrete type rather than leaving your functions abstract
61 in the HashTable class should allow GHC to optimize away the typeclass
62 dictionaries.
63
64 This package accepts a couple of different cabal flags:
65
66 * @unsafe-tricks@, default /ON/. If this flag is enabled, we use some
67 unsafe GHC-specific tricks to save indirections (namely @unsafeCoerce#@
68 and @reallyUnsafePtrEquality#@. These techniques rely on assumptions
69 about the behaviour of the GHC runtime system and, although they&#39;ve been
70 tested and should be safe under normal conditions, are slightly
71 dangerous. Caveat emptor. In particular, these techniques are
72 incompatible with HPC code coverage reports.
73
74 * @sse41@, default /OFF/. If this flag is enabled, we use some SSE 4.1
75 instructions (see &lt;http://en.wikipedia.org/wiki/SSE4&gt;, first available on
76 Intel Core 2 processors) to speed up cache-line searches for cuckoo
77 hashing.
78
79 * @bounds-checking@, default /OFF/. If this flag is enabled, array accesses
80 are bounds-checked.
81
82 * @debug@, default /OFF/. If turned on, we&#39;ll rudely spew debug output to
83 stdout.
84
85 * @portable@, default /OFF/. If this flag is enabled, we use only pure
86 Haskell code and try not to use unportable GHC extensions. Turning this
87 flag on forces @unsafe-tricks@ and @sse41@ /OFF/.
88
89 This package has been tested with GHC 7.0.3, on:
90
91 * a MacBook Pro running Snow Leopard with an Intel Core i5 processor,
92 running GHC 7.0.3 in 64-bit mode.
93
94 * an Arch Linux desktop with an AMD Phenom II X4 940 quad-core processor.
95
96 * a MacBook Pro running Snow Leopard with an Intel Core 2 Duo processor,
97 running GHC 6.12.3 in 32-bit mode.
98
99 Please send bug reports to
100 &lt;https://github.com/gregorycollins/hashtables/issues&gt;.
101 </longdescription>
102 <use>
103 <flag name="bounds-checking">if on, use bounds-checking array accesses</flag>
104 <flag name="portable">if on, use only pure Haskell code and no GHC extensions.</flag>
105 <flag name="sse4_1">Enable optimization for SSE4_1 capable processors (Intel Core 2 Penryn and later chips)</flag>
106 <flag name="sse4_2">Enables SSE4.2 optimizations: Nehalem-based Intel Core i7 or later.</flag>
107 <flag name="unsafe-tricks">turn on unsafe GHC tricks</flag>
108 </use>
109 </pkgmetadata>

  ViewVC Help
Powered by ViewVC 1.1.20