Thetapad Speed Benchmark: 3x Faster Than Overleaf

The Compilation Speed Problem

Every LaTeX user knows the frustration: you make a small change, hit compile, and wait... and wait.

On Overleaf during peak hours, a simple document might take 15-30 seconds. A complex thesis? Several minutes. And if you hit the compile timeout on the free tier, your document simply doesn't build.

We decided to measure exactly how much time researchers waste waiting for compilation.

Our Benchmark Methodology

We tested three representative document types across multiple platforms:

Test Documents

1. Simple document - 5 pages, basic formatting, 2 equations
2. Research paper - 15 pages, 20 figures, 50 citations, complex tables
3. PhD thesis - 180 pages, 80 figures, 200 citations, multiple chapters

Platforms Tested

• Overleaf (Free tier and paid tier)
• Thetapad (Standard)
• Local TeX Live (for reference baseline)

Test Conditions

• Same source files across all platforms
• 10 compilation runs per test, averaged
• Tests run during both peak hours (2-4 PM EST) and off-peak hours
• Clean compilation (no cached files)

The Results

Simple Document (5 pages)

| Platform | Off-Peak | Peak Hours | |----------|----------|------------| | Local TeX Live | 1.2s | 1.2s | | Thetapad | 2.1s | 2.3s | | Overleaf (Paid) | 4.8s | 7.2s | | Overleaf (Free) | 6.1s | 18.4s |

Even for simple documents, Thetapad compiles 2-8x faster than Overleaf.

Research Paper (15 pages)

| Platform | Off-Peak | Peak Hours | |----------|----------|------------| | Local TeX Live | 4.8s | 4.8s | | Thetapad | 6.2s | 6.8s | | Overleaf (Paid) | 14.2s | 22.1s | | Overleaf (Free) | 19.8s | 45.3s |

For a typical paper, Thetapad saves 8-38 seconds per compilation.

PhD Thesis (180 pages)

| Platform | Off-Peak | Peak Hours | |----------|----------|------------| | Local TeX Live | 18.4s | 18.4s | | Thetapad | 24.1s | 26.3s | | Overleaf (Paid) | 58.2s | 89.4s | | Overleaf (Free) | 78.6s | TIMEOUT |

Large documents show the most dramatic difference. Overleaf's free tier often times out entirely on thesis-sized documents.

Why the Difference?

Overleaf's Architecture

Overleaf uses a shared server model:

Your Document → Queue → Shared Server → Queue → PDF
                 ↑                         ↓
            Other Users              Other Users

During peak usage:

• Compilations queue behind other users
• Server resources are divided among thousands
• Complex documents may hit timeout limits

Thetapad's Architecture

Thetapad uses a local-first approach:

Your Document → Your Resources → PDF

Benefits:

• No queue: Compilation starts immediately
• Dedicated resources: Your machine or dedicated cloud instance
• Intelligent caching: Unchanged sections don't recompile
• Incremental builds: Only changed parts rebuild

Real-World Impact

Let's quantify what these numbers mean for actual work:

During a Writing Session

A typical PhD student might compile 50-100 times during an intense writing day.

| Platform | 50 Compiles | 100 Compiles | |----------|-------------|--------------| | Thetapad | 22 minutes | 44 minutes | | Overleaf (Paid) | 75 minutes | 150 minutes | | Overleaf (Free, peak) | 38+ minutes | TIMEOUTS |

That's 1-2 hours saved per writing day.

During Deadline Week

Conference deadlines see the worst Overleaf performance. We measured during NeurIPS 2024 submission week:

• Average Overleaf queue time: 2.5 minutes
• Timeout rate: 34% on free tier
• Thetapad performance: Unchanged

When deadlines matter most, Thetapad keeps working at full speed.

Beyond Raw Speed

No Timeouts

Overleaf's free tier has a 1-minute compile timeout. Thetapad has no compile time limit—your document compiles completely, regardless of complexity.

Offline Compilation

Thetapad compiles offline. No internet? No problem.

Scenario: Flight to conference
Overleaf: No compilation possible
Thetapad: Full-speed compilation throughout

Consistent Performance

Overleaf performance varies wildly based on:

• Time of day
• Day of week
• Academic calendar
• Server maintenance

Thetapad performance is consistent and predictable.

Technical Deep Dive

For those interested in the technical details:

Caching Strategy

Thetapad caches:

• Compiled auxiliary files (.aux, .toc, .bbl)
• Unchanged figure conversions
• Package installations

Result: Subsequent compilations are 40-60% faster.

Incremental Compilation

When you change one section, Thetapad identifies what actually needs recompilation:

% This chapter changed → recompile
\chapter{Introduction}
...
 
% These chapters unchanged → use cache
\chapter{Literature Review}
\chapter{Methodology}

Overleaf recompiles everything, every time.

Resource Allocation

| Aspect | Overleaf | Thetapad | |--------|----------|----------| | CPU | Shared | Dedicated | | Memory | Limited | Full system | | Storage | Slow shared | Fast local/SSD | | Network | Required | Optional |

Migration is Easy

Worried about switching? Your existing LaTeX documents work identically:

1. Export your Overleaf project (Menu → Download → Source)
2. Import into Thetapad (drag and drop)
3. Compile—usually faster on the first try

Same LaTeX, same packages, faster compilation.

Conclusion

Our benchmarks show Thetapad compiles LaTeX documents 2-8x faster than Overleaf, with the gap widening for larger documents and peak usage times.

For a PhD student or active researcher, this translates to:

• 1-2 hours saved per intense writing day
• No compile timeouts, ever
• Consistent performance regardless of when you work
• Offline capability for travel and poor connectivity

Time spent waiting for compilation is time not spent on research. Make the switch and get those hours back.

Benchmarks conducted December 2024 using standard document templates. Individual results may vary based on document complexity and system specifications.