MatLogica AADC delivers 444× faster performance than Basic Python and 4.6× faster Greeks than hand-optimised C++ — while requiring minimal code changes.
For computationally-heavy workloads (which is most of quant finance), AADC delivers superior performance with dramatically less complexity. No low-level expertise needed — AADC handles optimisations and computation of derivatives (AAD) automatically.
Traditionally, you choose productivity OR performance. AADC gives you both—and more.
Asian Option Monte Carlo - 1000 trades × 500K scenarios - Price + Greeks
1,000 trades × 500K scenarios × 252 timesteps — Price + Greeks
Simple but cannot scale
Fast but complex
Fast AND simple
AAD vs traditional bump-and-revalue — Risk calculations require Greeks (Delta, Rho, Vega)
Requires 4-7 full pricings per trade for Delta, Rho, Vega sensitivities.
Requires just 1 forward + 1 adjoint pass — all Greeks in one sweep.
AADC code is nearly identical to Basic Python (~37 lines added), yet runs 444× faster and delivers 4.6× faster Greeks than hand-optimised C++.
# AADC Python - Key additions only
import aadc
# 1. Create active variables
S0_v = aadc.idouble(S0)
S0_arg = S0_v.mark_as_input()
# 2. Record computation graph
funcs = aadc.Functions()
funcs.start_recording()
# ... your existing Python code ...
payoff_res = discounted_payoff.mark_as_output()
funcs.stop_recording()
# 3. Evaluate with automatic Greeks
request = {payoff_res: [S0_arg, r_arg, sigma_arg]}
results = aadc.evaluate(funcs, request, inputs, workers)
# Greeks from single adjoint pass - NO extra pricings!
delta = results[1][payoff_res][S0_arg]
rho = results[1][payoff_res][r_arg]
vega = results[1][payoff_res][sigma_arg]Greeks via AAD: 1 forward + 1 adjoint pass — +31% overhead vs +593% for bump-and-revalue.
From specification to production benchmark in hours
Use your model or a model developed by AI with your specification. If using AI, be sure to define all parameters, constraints, do's and don'ts.
The AADC-Agent converts your model (Python or C++) to AADC-enabled versions. Works like any Claude task—ask it to fix issues. Choose your language and hardware.
Get your benchmark in hours, not weeks. See exactly how AADC performs. If AI can't help, MatLogica is always there for expert support.
We don't advise developing quant models using AI, but it's a good way to get started prototyping with AADC.
We don't suggest using AI for production integrations. MatLogica's integration/debugging toolkit should be used for production integrations.
Benchmarks executed on enterprise-grade server hardware
| CPU | 2× Intel Xeon Platinum 8280L @ 2.70GHz |
| Cores | 56 physical (28 per socket), 112 threads |
| Architecture | x86_64, Cascade Lake |
| L3 Cache | 77 MiB (38.5 MiB per socket) |
| RAM | 283 GB |
| OS | Linux kernel 6.1.0-13-amd64 (Debian 64-bit) |
CPU Features: AVX-512, AVX2, FMA, AES-NI
| Trades | 1,000 |
| Scenarios | 500,000 |
| Timesteps | 252 |
| Threads | 16 |
| Total Paths | 500M |
Asian Option Monte Carlo with GBM dynamics
| GCC | 12.2.0 (Debian) |
| Clang | 14.0.6 (Debian) |
| Python | 3.11.2 + AADC |
Dual-socket server with AVX-512 vectorization
All benchmarks are transparent, reproducible, and verifiable. Run the code yourself and confirm the results.