Process separate presentations in parallel
Overview
Process several presentations in parallel by giving each worker its own presentation instance. Your application controls scheduling, output paths and resource use. This is useful for a server that handles independent documents at the same time.
Usage example
Create two input decks, then process them with at most two workers. Each worker opens its own presentation instance and writes unique PPTX/PDF filenames. North-output contains North processed and South-output contains South processed. This C# example uses Aspose.Slides for .NET; no input files are required. Initialize any required file-based licensing before starting workers. The worker limit is an application choice, not a shared-instance thread-safety guarantee.
using System;
using Aspose.Slides;
using Aspose.Slides.Export;
using System.IO;
using System.Threading.Tasks;
var names = new[] { "North", "South" };
foreach (var name in names)
{
using var input = new Presentation();
var shape = input.Slides[0].Shapes.AddAutoShape(ShapeType.Rectangle, 60, 60, 540, 100);
shape.TextFrame.Text = name;
input.Save(name + ".pptx", SaveFormat.Pptx);
}
// Apply your file-based license before starting workers when one is required.
var options = new ParallelOptions { MaxDegreeOfParallelism = 2 };
Parallel.ForEach(names, options, name =>
{
using var presentation = new Presentation(name + ".pptx");
var shape = (IAutoShape)presentation.Slides[0].Shapes[0];
shape.TextFrame.Text = name + " processed";
presentation.Save(name + "-output.pptx", SaveFormat.Pptx);
presentation.Save(name + "-output.pdf", SaveFormat.Pdf);
});
Console.WriteLine("Processed North and South with separate presentation instances.");
FAQ
Can several threads load, save or clone the same presentation instance?
Do not share a presentation instance for those operations. Apply licensing outside concurrent work as well, since license setup is not thread-safe.
How should I limit worker count on a memory-constrained server?
Choose a bounded worker count from measured peak memory and temporary-storage use for representative inputs. Two workers in this example is a starting choice, not a universal limit.
When should I initialize licensing relative to worker startup?
Apply the required license during application startup before starting parallel work. Each worker should then own and dispose its own presentation instance.
Guides
Related feature: Process multiple presentations in a loop.