Benchmarks

Benchmarks

This page will provide benchmarks between Sage and Mathematica, please check back periodically as I am still in the process of determining the best way to benchmark each application while ensuring fair comparisons.

System Used

Hardware

• HP Pavillion dv2000
• Intel(R) Core(TM)2 Duo CPU T5250 @ 1.50GHz
• Memory: 2Gb

Software

• Operating System: Linux - OpenSuse 11.0 (x86) Kernel: 2.6.25.16-0.1-pae
• Sage Version: Version 3.1.2, Release Date: 2008-09-19 (Note: If a benchmark uses a newer version of Sage the version will specified with the benchmark)
• Mathematica Version: 6.0.2.1

Benchmarks

Contour Plots

Comparison of Sage and Mathematica's Contour Plot functions.

Function Plotted: ${y}^{2}-{x}^{3}+x$

X Range: -3,3

Y Range: -3,3

Benchmarks are based on the average execution time of 100 calls to the Contour Plot Function.

Sage

Code

times = []
for i in range(0,100):
    t1 = time()
    pl = contour_plot(y^2 - x^3 + x, (-3,3),(-3,3),fill=False,contours=0,plot_points=num_points)
    t2 = time()
    times.append(t2 - t1)

Comparison with different number of points plotted

The Sage contour_plot function takes a plot_points parameter which specifies " the number of points to plot in each direction of the grid." The chart below shows the execution time of contour_plots with plot_points values ranging from 100 to 600 at 100 point increments.

Results

Points	Execution Time in Seconds
100	0.0482304501533508
200	0.0869896674156189
300	0.153420553207397
400	0.242804827690125
500	0.360980730056763
600	0.493348760604858

Mathematica

Code

time  = Timing[
Do[ContourPlot[y^2 - x^3 + x, {x, -3, 3}, {y, -3, 3},Contours -> {0}, ContourShading -> False], {a, 0, 100}]]

Results

Average Execution Time: 0.0516832 seconds

Side by Side

Since Mathematica doesn't have parameter for number points to plot in the grid it makes a completely accurate comparison impossible. Subjectively however with plot_points in Sage set to 100 you get a picture of the same visual quality as with Mathematica's.

Sage with plot_points=100: Avg Execution Time: 0.0482304501533508 seconds

Mathematica: Avg Execution Time: 0.0516832 seconds

Interpretation

While it would be a mistake to assume from these benchmarks that Sage's contour plot is faster than Mathematica's. It does point to the conclusion that the difference in performance between the two functions is minimal.

Factoring Integers

Comparison of Sage's factor function and Mathematica's FactorInteger function.

Note: Using Sage Version 3.2.1

Sage Code

t1 = time()
factor( val )
t2 = time()
print t2 - t1

Mathematica Code

Timing[FactorInteger[val]]

50 Digit Integers

---

First Value: 1934830109234738389299110111002010010027213211173

Times:

Sage: 0.191963911057

Mathematica: 0.368024

---

Second Value: 5223742730473248747591023429843298492835092582481

Times:

Sage: 0.0329298973083

Mathematica: 0.004

---

Third Value: 2348172983123812382173987894732567643276473646343

Times:

Sage: 1.05752491951

Mathematica: 1.85212

75 Digit Integers

---

First Value: 12093434274949817483748923743748732489734873847384738297489237483274823744

Times:

Sage: 87.9907610416

Mathematica: 111.299

---

Second Value: 92394892301893143256458324935465418124832974265687348465723012392392733473

Times:

Sage: 5.50352191925

Mathematica: 1.09607

---

Third Value: 50982472839478916416453643410109192831273987194673611347834632456123894327

Times:

Sage: 38.9131929874

Mathematica: 237.883

Results

For both the sample 50 and 75 digit integers Sage was faster 2 out of the 3 times. However most cases the difference is small especially when considering the simplicity of the benchmark.

There were a couple interesting results:

For the second 75 digit integer Sage was almost 5.5 times slower than Mathematica.

For the third 75 digit integer Mathematica was approx. 6 times slower than Sage.

Note: Sage results were obtained using the PARI Package, PARI uses the number field sieve algorithm for integer factorization. I'm not sure if Mathematica also uses the number field sieve algorithm.