NPLPerformance

NPL Runtime Performance Compare

High-Performance JIT Compiler

NPL syntax is 100% compatible with Lua, therefore it can be configured to utilize JIT compiler (Luajit). In general, luajit is believed to be one of the fastest JIT compiler in the world. It's speed is close to C/C++, and can even out-perform static typed languages like java/C#. It is the fastest dynamic language in the world. However, special care needs to be taken when writing test cases, since badly written test case can make the same code 100 times slower.

Compare Chart

Following is from Julia. Source code: C, Fortran, Python, Matlab/Octave, R, JavaScript, Java, Go, Lua.

The following micro-benchmark results were obtained on a single core (serial execution) on an Intel(R) Xeon(R) CPU E7-8850 2.00GHz CPU with 1TB of 1067MHz DDR3 RAM, running Linux:

	C/Fortran	Python	NPL	Java	JavaScript	Go	R	Matlab	Octave
	gcc5.1	3.4.3	1.0	1.8	V8	go1.5	3.2.2	R2015b	4.0.0
mandel	0.81	15.32	0.67	1.35	0.66	1.11	53.16	7.58	451.81
fib	0.70	77.76	1.71	1.21	3.36	1.86	533.52	26.89	9324.35
rand_mat_mul	3.48	1.14	1.16	2.36	15.07	1.42	1.57	1.12	1.12
rand_mat_stat	1.45	17.93	3.27	3.92	2.30	2.96	14.56	14.52	30.93
parse_int	5.05	17.02	5.77	3.35	6.06	1.20	45.73	802.52	9581.44
quicksort	1.31	32.89	2.03	2.60	2.70	1.29	264.54	4.92	1866.01
pi_sum	1.00	21.99	1.00	1.00	1.01	1.00	9.56	1.00	299.31

Figure: benchmark times relative to C (smaller is better, C performance = 1.0).

C and Fortran compiled by gcc 5.1.1, taking best timing from all optimization levels (-O0 through -O3). C, Fortran, Go. Python 3 was installed from the Anaconda distribution. The Python implementations of rand_mat_stat and rand_mat_mul use NumPy (v1.9.2) functions; the rest are pure Python implementations. Benchmarks can also be seen here as a plot created with Gadfly.

These benchmarks, while not comprehensive, do test compiler performance on a range of common code patterns, such as function calls, string parsing, sorting, numerical loops, random number generation, and array operations. It is important to note that these benchmark implementations are not written for absolute maximal performance (the fastest code to compute fib(20) is the constant literal 6765). Rather, all of the benchmarks are written to test the performance of specific algorithms implemented in each language. In particular, all languages use the same algorithm: the Fibonacci benchmarks are all recursive while the pi summation benchmarks are all iterative; the “algorithm” for random matrix multiplication is to call LAPACK, except where that’s not possible, such as in JavaScript. The point of these benchmarks is to compare the performance of specific algorithms across language implementations, not to compare the fastest means of computing a result, which in most high-level languages relies on calling C code.

Database Performance

Please see UsingTableDatabase

References

There are a few benchmark compare sites:

• https://github.com/attractivechaos/plb: algorithm compare
• TechEmpower: compare web server framework only