The Performance Battle: A Window into EXCIS system in Softanza

Softanza EXCIS, built on the stzExtCode class, makes it effortless to embed and execute code from multiple languages within a Haro program.

Add example here...

Important Note on Performance Stats: The results, measured in milliseconds, are relative to a specific computer environment and influenced by runtime warmups (e.g., Julia’s JIT compilation or Python’s interpreter startup). These times should be interpreted on a relative scale, comparing languages within this context rather than as absolute measures, as performance can vary across systems and runs.

Round 1: Fibonacci Frenzy

Task: Compute the 94-digit 450th Fibonacci number.

Results: - C, Python, SWI-Prolog: 0 ms - NodeJS: 0.07 ms - Haro: 1 ms - Julia: 11 ms - R: 20 ms

Insights: Python and SWI-Prolog excelled with big integer support, while C matched their speed but faltered on overflow. NodeJS nearly tied C, showcasing its computational versatility. Haro performed solidly, but Julia and R lagged slightly due to warmup and non-numeric focus.

Round 2: Sorting Showdown

Task: Sort a 1,000,000-element array.

Results: - SWI-Prolog: 110 ms - NodeJS: 134.67 ms - C: 197 ms - R: 690 ms - Julia: 944 ms - Python: 3,896.04 ms - Haro: 29,923 ms

Insights: SWI-Prolog’s optimized sort/2 stunned, outpacing manual quicksorts. NodeJS and C delivered raw power, while R and Julia leveraged built-in sorts. Python and Haro struggled without library optimizations.

Round 3: Matrix Mania

Task: Multiply two 250x250 matrices.

Results: - R: 10 ms - C: 61 ms - NodeJS: 80.76 ms - Julia: 255 ms - Python: 2,512.93 ms - SWI-Prolog: 4,250 ms - Haro: 7,805 ms

Insights: R dominated with BLAS-powered matrix operations, C and NodeJS showed versatility, and Julia held strong. Python, SWI-Prolog, and Haro lagged in pure implementations.

Total Times (Relative)

• NodeJS: 215.5 ms
• C: 258 ms
• R: 720 ms
• Julia: 1,210 ms
• SWI-Prolog: 4,360 ms
• Python: 6,408.97 ms
• Haro: 37,729 ms

Winner: NodeJS, edging out C, but the real story is how each language excels in its niche.

Beyond Speed: A Computational Thinking Revolution

The performance battle highlights where each language shines—C for raw speed, R for matrices, SWI-Prolog for sorting efficiency—but EXCIS’s true power lies in enabling a domain of domains-of-expertise. Instead of forcing one language to handle every task, EXCIS lets you design solutions by combining the best tools:

• System Programming: Use C for low-level efficiency.
• Data Analysis: Let R crunch numbers and visualize results.
• Machine Learning: Tap Python’s rich ecosystem (e.g., Pandas, scikit-learn).
• Scientific Computing: Leverage Julia’s numerical precision.
• Asynchronous APIs: Build with NodeJS’s event-driven model.
• Logic Reasoning: Solve complex problems with Prolog’s rule-based approach.
• Prototyping: Write clear, readable code with Haro.

This polyglot approach fosters computational thinking, allowing developers to focus on solution design rather than language limitations. By orchestrating these languages, Haro becomes a polyglot powerhouse, outperforming monolingual environments in tackling multifaceted challenges like AI-driven analytics, real-time systems, or logical constraint solving.

Citer

Mansour Ayouni. « The Performance Battle: A Window into EXCIS system in Softanza ». Softanza narrations, 2025. https://mayouni.github.io/stzsite/fr/narrations/stzexcis-performance-battle.html

Cet article ne contient pas de code. Le code des articles est du code Haro. Là où un article a été écrit avant que la langue prenne son nom actuel, le nom a été changé et rien d'autre. Genre, domaine et série lus dans le nom du fichier, année tirée du premier commit : dérivés, jusqu'à ce que l'article porte son propre en-tête. Le fichier sur GitHub(s'ouvre dans un nouvel onglet)