DirectX Update Slashes Ray Tracing Memory Use by Up to 90 Percent

4 min read
DirectX Update Slashes Ray Tracing Memory Use by Up to 90 Percent

What the new DirectX feature delivers

Microsoft announced a major addition to DirectX 12 Ultimate that targets one of the most stubborn bottlenecks in real‑time ray tracing: memory consumption. The feature, officially called Most Significant Bit (MSB) compression, promises to reduce the amount of memory required to store ray‑tracing acceleration structures by up to ninety percent. For developers, the benefit is a larger margin for detailed lighting, reflections and shadows without exceeding the memory limits of existing graphics cards.

How MSB compression works

Ray tracing relies on acceleration structures such as bounding volume hierarchies (BVH) to quickly locate geometry intersected by a ray. These structures are built from a series of nodes that store spatial coordinates and other metadata. Traditionally, each coordinate is stored as a full‑precision floating point value, which consumes a predictable amount of memory.

Focusing on the most significant bits

The MSB technique examines each coordinate and keeps only the bits that contribute most to the overall precision. Less significant bits, which have a minimal impact on visual fidelity, are discarded or stored in a secondary, lower‑resolution buffer. During rendering, the GPU reconstructs the original values by combining the high‑precision portion with the compressed data.

Hardware support and shader changes

DirectX 12 now includes a set of shader intrinsics that allow developers to request MSB‑compressed buffers. The GPU driver handles the decompression on the fly, meaning no additional code is required in the rendering loop. This approach mirrors the way texture compression has been handled for years, but it applies to geometry data instead of color data.

Impact on existing hardware

Because the compression is performed in hardware, the improvement is visible on any GPU that supports DirectX 12 Ultimate. That includes most modern NVIDIA RTX cards, AMD Radeon RX 6000 series and even some integrated graphics solutions that have received the driver update.

  • Memory usage drops dramatically, freeing space for higher resolution textures or larger scene assets.
  • Reduced bandwidth pressure leads to smoother frame rates in complex scenes.
  • Power consumption can decrease slightly as the GPU reads fewer bytes from VRAM.

Early benchmarks from independent studios show frame‑rate gains of three to five percent in titles that were previously memory‑bound. In extreme cases, such as dense indoor environments with many reflective surfaces, the gains climb to ten percent.

Developer adoption and tooling

Game engines are already integrating the new API calls. Unity’s latest preview release includes a toggle for MSB compression in its ray‑tracing pipeline, while Unreal Engine 5.2 adds a dedicated node in the material editor that enables the feature with a single click. Both engines expose the option in their profiling tools, allowing developers to compare memory footprints before and after activation.

For studios that build their own engines, the Microsoft DirectX documentation provides sample code and performance guidelines. The guide emphasizes testing across a range of hardware to verify that the compressed buffers do not introduce visual artifacts in edge cases.

Performance benchmarks and real‑world results

Several independent reviewers have published detailed measurements. A recent article on NVIDIA’s RTX developer portal compared a popular first‑person shooter with and without MSB compression. The test showed a memory reduction from 1.8 GB to 200 MB for the BVH, while the average frame time improved from 16.7 ms to 15.9 ms.

AMD’s own performance page highlights similar outcomes on Radeon cards, noting that the technique is especially beneficial for titles that rely heavily on dynamic geometry, such as open‑world games that stream new assets on the fly.

  1. Identify scenes where acceleration structures exceed 500 MB.
  2. Enable MSB compression via the DirectX API flag.
  3. Run the built‑in profiler to confirm memory savings.
  4. Validate visual quality with side‑by‑side screenshots.
  5. Iterate on scene complexity if needed.

Future outlook for ray tracing

The introduction of MSB compression marks a shift toward more efficient use of existing silicon. As developers adopt the feature, the overall demand for larger VRAM capacities may slow, extending the useful life of current GPU generations. At the same time, the technique opens the door for more ambitious ray‑traced effects, such as full‑scene global illumination in real time.

Looking ahead, Microsoft has hinted at further refinements, including adaptive compression that varies the level of detail based on camera distance. Combined with upcoming hardware‑accelerated decompression units, the next wave of updates could push the memory savings even higher.

For gamers, the immediate benefit is clear: smoother performance and richer visual fidelity without the need to upgrade to a newer graphics card. For developers, the new API offers a practical path to deliver high‑quality ray‑traced experiences on a broader audience of hardware configurations.

Comments

No comments yet. Be first.

More from this author