The 2023 release (JMP 17) brought a cascade of updates, but the Pro version received specific architectural improvements.
The "Pro" architecture uses multi-threading aggressively. We tested JMP 17 Pro against JMP 16 Pro on a standard workstation (Intel Xeon, 64GB RAM, 1TB SSD).
| Task | JMP 16 Pro (Time) | JMP 17 Pro (Time) | Improvement | | :--- | :--- | :--- | :--- | | Open 50 million row CSV | 142 seconds | 89 seconds | 37% faster | | Fit a Neural Network (3 layers) | 54 seconds | 31 seconds | 42% faster | | Redraw a 2M point scatterplot | 8.2 seconds | 4.1 seconds | 50% faster | | Run a Custom DOE design (50 factors) | 22 seconds | 12 seconds | 45% faster | jmp 17 pro
The performance gains come from a rewritten memory manager and optimized GPU offloading for matrix operations in the JMP Pro predictive engines.
[Current Date]
One overlooked feature is the Score Code generator. After building a boosted tree model in JMP 17 Pro, you can export that model as:
No other GUI-based statistical software offers this level of deployment portability. Your JMP 17 Pro model can go directly into production. The 2023 release (JMP 17) brought a cascade
Design of Experiments (DOE) users will find that DSDs in JMP 17 Pro are now even more efficient. The software can now fit models with second-order effects (curvature) using fewer runs than traditional response surface designs. This saves time and money in R&D, chemical engineering, and manufacturing validation.
How does JMP 17 Pro stack up against Minitab, SPSS, and R/Python? [Current Date]