README.txt
1V8 Benchmark Suite
2==================
3
4This is the V8 benchmark suite: A collection of pure JavaScript
5benchmarks that we have used to tune V8. The licenses for the
6individual benchmarks are included in the JavaScript files.
7
8In addition to the benchmarks, the suite consists of the benchmark
9framework (base.js), which must be loaded before any of the individual
10benchmark files, and two benchmark runners: An HTML version (run.html)
11and a standalone JavaScript version (run.js).
12
13
14Changes From Version 1 To Version 2
15===================================
16
17For version 2 the crypto benchmark was fixed. Previously, the
18decryption stage was given plaintext as input, which resulted in an
19error. Now, the decryption stage is given the output of the
20encryption stage as input. The result is checked against the original
21plaintext. For this to give the correct results the crypto objects
22are reset for each iteration of the benchmark. In addition, the size
23of the plain text has been increased a little and the use of
24Math.random() and new Date() to build an RNG pool has been removed.
25
26Other benchmarks were fixed to do elementary verification of the
27results of their calculations. This is to avoid accidentally
28obtaining scores that are the result of an incorrect JavaScript engine
29optimization.
30
31
32Changes From Version 2 To Version 3
33===================================
34
35Version 3 adds a new benchmark, RegExp. The RegExp benchmark is
36generated by loading 50 of the most popular pages on the web and
37logging all regexp operations performed. Each operation is given a
38weight that is calculated from an estimate of the popularity of the
39pages where it occurs and the number of times it is executed while
40loading each page. Finally the literal letters in the data are
41encoded using ROT13 in a way that does not affect how the regexps
42match their input.
43
44
45Changes from Version 3 to Version 4
46===================================
47
48The Splay benchmark is a newcomer in version 4. It manipulates a
49splay tree by adding and removing data nodes, thus exercising the
50memory management subsystem of the JavaScript engine.
51
52Furthermore, all the unused parts of the Prototype library were
53removed from the RayTrace benchmark. This does not affect the running
54of the benchmark.
55
56
57Changes from Version 4 to Version 5
58===================================
59
60Removed duplicate line in random seed code, and changed the name of
61the Object.prototype.inherits function in the DeltaBlue benchmark to
62inheritsFrom to avoid name clashes when running in Chromium with
63extensions enabled.
64
65
66Changes from Version 5 to Version 6
67===================================
68
69Removed dead code from the RayTrace benchmark and fixed a couple of
70typos in the DeltaBlue implementation. Changed the Splay benchmark to
71avoid converting the same numeric key to a string over and over again
72and to avoid inserting and removing the same element repeatedly thus
73increasing pressure on the memory subsystem. Changed the RegExp
74benchmark to exercise the regular expression engine on different
75input strings.
76
77Furthermore, the benchmark runner was changed to run the benchmarks
78for at least a few times to stabilize the reported numbers on slower
79machines.
80
81
82Changes from Version 6 to Version 7
83===================================
84
85Added the Navier-Stokes benchmark, a 2D differential equation solver
86that stresses arithmetic computations on double arrays.
87
