Intel whitepaper demonstrates substantial performance gains from 5th Gen Xeon processors including up to 2.08x improvement compared to 3rd Gen Xeon CPUs using MatLogica AADC.

Comprehensive benchmark comparing JAX PyTorch and TensorFlow vs MatLogica AADC for quantitative applications and ML workloads. Independent results with downloadable source code demonstrating 10x+ performance advantage for quantitative finance workloads.

GPUs are often assumed to be faster for AAD but MatLogica's CPU-optimized AADC can outperform GPU-based solutions for many quant workloads especially when leveraging AVX2/AVX512 vectorization and multithreading.

Comprehensive performance comparison between MatLogica AADC and popular machine learning tools for quantitative finance applications.

Independent open-source benchmark comparing MatLogica AADC with Adept and other AAD libraries for quantitative finance workloads.

Intel-led whitepaper demonstrating up to 1770x performance increase for XVA pricing and 830x for XVA risks on Intel processors when using MatLogica AADC. Open-source benchmark available on GitHub.

In this blog post Maksim explores how MatLogica's AADC supercharges QuantLib dramatically accelerating pricing and risk calculations without requiring a rewrite. By leveraging cutting-edge compiler techniques and automatic differentiation this integration unlocks unprecedented performance gains for quants and developers.

Presentation by Stephan Bosch Quantitative Developer at ING comparing AAD techniques and performance. Highlighted enormous computational acceleration achieved with MatLogica AADC in production environments.

Presentation by Svetlana Borovkova: Algorithmic Adjoint Differentiation For Tail Risk Model Risk & Stress Testing demonstrating MatLogica AADC for expected shortfall sensitivities.

MatLogica's AADC enabled 15-20x speedups reducing overnight portfolio risk computation from 8+ hours to 2 hours and intraday risk calculation from 30+ minutes to a few minutes. Opened new revenue streams reduced infrastructure costs and improved risk management.

Risk calculations need scalable cloud execution but exposing source code or models in the cloud risks intellectual property or regulatory compliance. Traditional approaches also struggle to deliver high performance without rewriting or parallelizing code manually.

Comprehensive guide to implementing valuation adjustments (XVA) using automatic adjoint differentiation for efficient risk calculation.

Legacy quant code is often single-threaded and scalar limiting performance on modern CPUs. MatLogica's AADC tool transforms object-oriented scalar code into vectorized multithreaded and NUMA-aware machine code unlocking parallelism without requiring major code refactoring.

Traditional AAD methods require extensive code changes or templates. MatLogica's Code Generation AAD uses code transformation and operator overloading to generate optimized machine code at runtime enabling fast and flexible gradient calculations without constraining control flow.

Detailed comparison of different automatic differentiation techniques including operator overloading source transformation and tape-based approaches.

AAD is complex and slow with traditional libraries. MatLogica's JIT compiler transforms user code into highly optimized vectorized and multithreaded machine code at runtime maximizing CPU utilization for AAD tasks.

In-depth explanation of the compiler technology and optimization techniques that make MatLogica AADC significantly faster than traditional AAD approaches.

Free workshop providing a practical solution to the quant trade-off: Python prototyping speed vs model performance. Demonstrated an easy-to-use framework combining Python flexibility with AAD and bare-metal performance speeding up models by 1000x.

Demonstrates how we transformed ORE into a streamlined LiveRisk service for FX products using AADC achieving 100x acceleration.

Short video demonstration of MatLogica AADC interface showing a basic example that's available online. Demonstrates the ease of use and immediate performance benefits.

Online sandbox with MultiCurve fitting and AAD risk demo using QuantLib and Levenberg-Marquardt library. Includes video demonstrating 350x performance boost for XVA pricing using Intel AVX2 and 5 threads.

Interactive demonstration of XVA calculations using MatLogica AADC integrated with QuantLib showing dramatic performance improvements.

MatLogica recognized as Category Leader in Automatic Differentiation with 5 total awards from Chartis Research.

MatLogica awarded Asia Risk Technology Newcomer of the Year for breakthrough performance in quantitative risk with AADC.

MatLogica selected as finalist in Accenture's global FinTech Innovation Lab validating AADC's transformative impact on financial analytics.

MatLogica recognized as Category Leader in Automatic Differentiation across multiple years by Chartis Research.

MatLogica presented at the 21st WBS Quantitative Finance Conference in Palermo Italy.

Presentation by Antoine Savine on using modern C++ with AADC for high-performance quantitative finance applications.