OUT-OF-DATE! Read http://shootout.alioth.debian.org/      

- all benchmarks - binary-trees chameneos-redux fannkuch fasta k-nucleotide mandelbrot meteor-contest n-body nsieve nsieve-bits partial-sums pidigits recursive regex-dna reverse-complement spectral-norm startup sum-file thread-ring - all languages - Ada 2005 GNAT BASIC FreeBASIC C CINT C GNU gcc C Intel C# Mono C++ GNU g++ C++ Intel CAL Clean D Digital Mars Eiffel SmartEiffel Erlang HiPE F# Mono Forth bigForth Forth GNU GForth Fortran G95 Fortran Intel Groovy Haskell GHC Icon Io Java 1.4 -server Java 6 -client Java 6 -server Java 6 -Xint Java 6 -Xms64m Java GNU gcj JavaScript Rhino JavaScript SpiderMonkey Lisaac Lisp SBCL Lua Lua LuaJIT Mercury Mozart/Oz Nice Oberon-2 OO2C OCaml Pascal Free Pascal Perl PHP Pike Prolog SWI Prolog YAP Python CPython Python IronPython Python Psyco Rebol Ruby 1.9 Ruby JRuby Ruby MRI Scala Scheme Chicken Scheme Ikarus Scheme PLT Smalltalk GNU Smalltalk Squeak Smalltalk VisualWorks SML MLton SML SML/NJ Tcl Zonnon Mono

 thread-ring benchmark hypotheses non fingo

↓ Switch from thread to thread passing one token, N=10,000,000.

Which programs used least Code? Which programs use highly optimised assembly code libraries? Which programs make use of all the processor cores?

		sort	sort	sort	sort
×	Program & Logs	CPU secs	Elapsed secs	Memory KB	Code B	~ CPU Load
1.0	Haskell GHC #2	6.70		2,672	260
1.1	Erlang HiPE	7.49		5,368	273
1.3	Mozart/Oz	8.78		3,888	340
3.3	Smalltalk Squeak	21.95		37,016	544
5.0	Smalltalk VisualWorks #2	33.29		13,012	566
6.3	SML MLton	41.87		6,132	361
8.3	C Intel	55.80		4,584	487
8.4	Scala	56.44		22,752	339
8.5	C GNU gcc	56.74		4,504	487
8.5	Lisaac	56.85		8,596	810
8.6	Pascal Free Pascal	57.53		2,588	523
8.8	Scheme Chicken	59.08		2,612	421
9.1	D Digital Mars	61.01		8,700	560
12	Smalltalk GNU	78.04		9,536	551
12	OCaml #3	80.31		2,900	296
13	OCaml #2	83.89		2,896	350
13	Python Psyco #2	84.01		7,820	300
13	Scheme PLT	90.12		21,388	262
15	C++ GNU g++	101.28		2,960	626
15	Python CPython #2	102.31		6,772	287
17	Python CPython	114.28		6,756	397
20	Nice #2	132.48		23,600	382
20	Java 6 -server #3	134.98		23,336	530
20	Java 6 -Xms64m #3	134.99		27,628	530
21	Java 6 -client #3	138.09		22,456	530
22	Ada 2005 GNAT	149.89		7,752	602
25	C# Mono	168.87		11,956	476
27	Java 6 -Xint #3	183.31		20,948	530
49	D Digital Mars #2	5 min		2,792	555
52	Java 6 -server	5 min		25,364	587
54	Groovy	5 min		53,212	489
55	OCaml	6 min		3,616	282
77	CAL	8 min		27,072	927
88	Ruby JRuby #2	9 min		42,028	228
100	Ruby JRuby	11 min		44,256	342
282	Ruby MRI #2	31 min		16,740	228
623	Ruby MRI	1h 09 min		12,088	342
	C++ Intel	Failed			626
	Perl	Timed Out			353
	Perl #2	Failed			288
	Python IronPython	Failed			397
	Python IronPython #2	Failed			287
	Ruby 1.9	Failed			342
	Ruby 1.9 #2	Failed			228
	Tcl	Failed			445
	Zonnon Mono	Failed			489
	Zonnon Mono #2	Failed			396
interesting alternative programs
	Lua LuaJIT #2	Failed			248
	Lua #2	Failed			248
	Python IronPython #3	Failed			272
0.2	Java 6 -server #2	1.21	0.00	11,900	693
1.6	Python CPython #3	10.68	0.00	2,624	272
2.7	Lua	17.95	0.00	1,280	255
3.0	Lua LuaJIT	19.87	0.00	3,732	255
missing programs
	BASIC FreeBASIC	No program
	C CINT	No program
	Clean	No program
	Eiffel SmartEiffel	No program
	F# Mono	No program
	Forth bigForth	No program
	Forth GNU GForth	No program
	Fortran G95	No program
	Fortran Intel	No program
	Icon	No program
	Io	No program
	Java 1.4 -server	No program
	Java GNU gcj	No program
	JavaScript Rhino	No program
	JavaScript SpiderMonkey	No program
	Lisp SBCL	No program
	Lua	No program
	Lua LuaJIT	No program
	Mercury	No program
	Oberon-2 OO2C	No program
	PHP	No program
	Pike	No program
	Prolog SWI	No program
	Prolog YAP	No program
	Rebol	No program
	Scheme Ikarus	No program
	SML SML/NJ	No program

 thread-ring benchmark : Switch from thread to thread passing one token

diff program output N = 1000 with this output file to check your program is correct before contributing.

Each program should create and keep alive 503 threads, explicity or implicitly linked in a ring, and pass a token between one thread and the next thread at least N times.

Each program should

• create 503 linked threads (named 1 to 503)
• thread 503 should be linked to thread 1, forming an unbroken ring
• pass a token to thread 1
• pass the token from thread to thread N times
• print the name of the last thread (1 to 503) to take the token

Similar benchmarks are described in Performance Measurements of Threads in Java and Processes in Erlang, 1998; and A Benchmark Test for BCPL Style Coroutines, 2004. For some language implementations increasing the number of threads quickly results in Death by Concurrency.

Programs may use kernel threads, lightweight threads; but coroutines, cooperative threads and other programs with custom schedulers will be listed as interesting alternative implementations. Briefly say what concurrency technique is used in the program header comment.