<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
  <title>Counterpoint Dynamics: insights &amp; articles</title>
  <link href="https://www.cpdynamics.co.za/blog/"/>
  <link rel="self" href="https://www.cpdynamics.co.za/blog/feed.xml"/>
  <id>https://www.cpdynamics.co.za/blog/</id>
  <updated>2026-10-27T00:00:00Z</updated>
  <entry>
    <title>The fall and slight rise of the GPU</title>
    <link href="https://www.cpdynamics.co.za/blog/gpu-fall-and-rise/"/>
    <id>https://www.cpdynamics.co.za/blog/gpu-fall-and-rise/</id>
    <published>2026-10-27T00:00:00Z</published>
    <updated>2026-10-27T00:00:00Z</updated>
    <author><name>David Naylor</name></author>
    <summary>Finance runs in double precision and makes decisions; GPUs prefer neither. Measured on an A100: where compiling our pricing code to the GPU loses, where Monte Carlo wins, and what f32 really costs.</summary>
    <category term="Quant Finance"/>
    <category term="GPU"/>
    <category term="Performance"/>
  </entry>
  <entry>
    <title>CP∂ versus Strata and QuantLib + ORE</title>
    <link href="https://www.cpdynamics.co.za/blog/cpdelta-vs-strata/"/>
    <id>https://www.cpdynamics.co.za/blog/cpdelta-vs-strata/</id>
    <published>2026-10-09T00:00:00Z</published>
    <updated>2026-10-09T00:00:00Z</updated>
    <author><name>Counterpoint Dynamics Engineering</name></author>
    <summary>An interactive comparison of CP∂, OpenGamma Strata and QuantLib with ORE across instrument coverage, analytics, margin and capital, reference data and verified pricing parity.</summary>
    <category term="Quant Finance"/>
    <category term="Pricing Libraries"/>
    <category term="Interactive"/>
  </entry>
  <entry>
    <title>Porting a SIMM calculator from Python to F# and C#</title>
    <link href="https://www.cpdynamics.co.za/blog/simm-calculator-port/"/>
    <id>https://www.cpdynamics.co.za/blog/simm-calculator-port/</id>
    <published>2026-09-07T00:00:00Z</published>
    <updated>2026-09-07T00:00:00Z</updated>
    <author><name>Armando Oberholzer</name></author>
    <summary>Three autonomous ports of our Python SIMM calculator reveal how design, independent verification and careful benchmarking shape correctness, complexity and performance.</summary>
    <category term="Quant Finance"/>
    <category term="Software Engineering"/>
    <category term="cpStudio"/>
  </entry>
  <entry>
    <title>Five implementations, one set of trades, and the bug our tests could not see</title>
    <link href="https://www.cpdynamics.co.za/blog/cross-library-validation/"/>
    <id>https://www.cpdynamics.co.za/blog/cross-library-validation/</id>
    <published>2026-08-31T00:00:00Z</published>
    <updated>2026-08-31T00:00:00Z</updated>
    <author><name>David Naylor</name></author>
    <summary>One calendar-free trade specification, five implementations and a closed form exposed two serious defects our own tests were structurally unable to see.</summary>
    <category term="Quant Finance"/>
    <category term="Model Validation"/>
    <category term="cpStudio"/>
  </entry>
  <entry>
    <title>Mapping cpStudio: An Interactive Architecture City</title>
    <link href="https://www.cpdynamics.co.za/blog/cpstudio-architecture-city/"/>
    <id>https://www.cpdynamics.co.za/blog/cpstudio-architecture-city/</id>
    <published>2026-08-14T00:00:00Z</published>
    <updated>2026-08-14T00:00:00Z</updated>
    <author><name>Counterpoint Dynamics Engineering</name></author>
    <summary>Explore cpStudio as an interactive city: its major systems become districts, modules become buildings, and the connections between them become streets.</summary>
    <category term="cpStudio"/>
    <category term="Software Architecture"/>
    <category term="Interactive"/>
  </entry>
  <entry>
    <title>We rebuilt the ISDA CDS Standard Model from its published test grids. The grids fought back.</title>
    <link href="https://www.cpdynamics.co.za/blog/isda-cds-test-grid/"/>
    <id>https://www.cpdynamics.co.za/blog/isda-cds-test-grid/</id>
    <published>2026-08-05T00:00:00Z</published>
    <updated>2026-09-08T00:00:00Z</updated>
    <author><name>David Naylor</name></author>
    <summary>Reproducing the ISDA CDS Standard Model to the cent exposed undocumented conventions, defects in the published test data, and a path to repricing the full grid 27 times faster than the reference C library.</summary>
    <category term="Quant Finance"/>
    <category term="Credit Derivatives"/>
    <category term="cpStudio"/>
  </entry>
  <entry>
    <title>The Hidden Cost of Doing Things &quot;Properly&quot;</title>
    <link href="https://www.cpdynamics.co.za/blog/hidden-cost-of-doing-things-properly/"/>
    <id>https://www.cpdynamics.co.za/blog/hidden-cost-of-doing-things-properly/</id>
    <published>2025-08-05T00:00:00Z</published>
    <updated>2025-08-05T00:00:00Z</updated>
    <author><name>Counterpoint Dynamics Engineering</name></author>
    <summary>There's a quiet tragedy in software development that we don't talk about enough: the projects that fail not because they were rushed, hacked together, or underfunded—but because they were too well-engineered. We explore the hidden costs of over-engineering and when &quot;good enough&quot; might actually be better.</summary>
    <category term="Software Development"/>
    <category term="Best Practices"/>
  </entry>
</feed>
