How DirectStorage Changes Storage Impact in PC Benchmarking

This article explores how the adoption of Microsoft's DirectStorage API in modern game engines fundamentally alters computer hardware benchmarking. Historically, storage drives only influenced game loading times and had virtually no impact on actual frame rates. By allowing storage devices to stream data directly to the graphics card for decompression, DirectStorage turns solid-state drives (SSDs) into active participants in real-time rendering. Consequently, hardware reviewers and benchmarkers must now account for storage speeds when evaluating CPU performance, GPU workloads, and frame time stability.

The End of the "Storage Doesn't Matter for FPS" Era

In traditional game engines, assets are pulled from the storage drive, placed into system RAM, sent to the CPU to be decompressed, and finally forwarded to the GPU to be rendered. Because the CPU was a massive bottleneck in this decompression pipeline, the maximum speed of the storage drive rarely mattered once a game was loaded. Upgrading from a standard SATA SSD to a high-end NVMe SSD yielded almost zero improvement in frames per second (FPS). Therefore, hardware benchmarkers largely ignored storage drives when testing graphics cards or processors, as any functional SSD produced the same in-game performance results.

Shifting the Workload from CPU to GPU

Game engines utilizing DirectStorage bypass the traditional data pipeline. High-speed NVMe drives now transfer compressed game assets directly to the GPU's memory via the PCIe bus, and the GPU handles the decompression using its vast parallel processing power.

This shift changes hardware benchmarking in two distinct ways:

  • CPU Benchmarks: DirectStorage significantly reduces CPU overhead. When benchmarking CPU-bound games that use this technology, weaker or older processors will show improved performance and higher frame rates because they are freed from heavy data decompression tasks.
  • GPU Benchmarks: Because the graphics card now uses a fraction of its compute power to decompress storage data, it has slightly fewer resources available for pure rendering. Benchmarkers evaluating GPUs must now recognize that the card is simultaneously acting as a rendering engine and a data decompression processor.

Measuring Frame Pacing and 1% Lows

The most critical impact of DirectStorage on gaming benchmarks is not necessarily higher average FPS, but improved frame pacing and higher 1% and 0.1% low frame rates. Modern game engines constantly stream massive textures and geometry in real-time as a player moves through a seamless world.

With DirectStorage, the bandwidth of the NVMe drive directly dictates how quickly these assets are fed to the GPU. When benchmarking slower PCIe 3.0 drives against faster PCIe 4.0 or PCIe 5.0 drives in DirectStorage-enabled engines, testers can finally measure tangible differences in traversal stutter, texture pop-in, and frame time consistency. A slow drive can now actively cause frame drops during fast-paced gameplay.

Redefining Synthetic Storage Metrics

For years, synthetic storage benchmarks focused heavily on maximum sequential read and write speeds—metrics that looked great on a box but rarely translated to real-world gaming performance. DirectStorage forces hardware testers to prioritize different metrics:

  • High Queue Depths: Traditional games process storage requests one or a few at a time. DirectStorage allows game engines to submit thousands of input/output (I/O) requests simultaneously.
  • Sustained Performance: Because assets are streamed continuously, a drive's ability to handle parallel data streams without thermal throttling becomes a primary metric for gaming viability.
  • Controller Efficiency: The raw speed of the NAND flash matters less than how efficiently the SSD's onboard controller can manage massive queues of randomized gaming data.

Stricter Standardization for Test Benches

The integration of DirectStorage requires a stricter standardization of hardware test benches. Reviewers can no longer swap out different SSDs between test runs under the assumption that the results will remain isolated to the GPU or CPU being tested. The storage drive is now a performance-altering variable on par with system memory speeds and latency. To produce accurate data, benchmarkers must now lock in a specific, high-bandwidth PCIe Gen 4 or Gen 5 NVMe drive for CPU and GPU testing, while also creating a new category of benchmarks where the SSD itself is tested using live gameplay framerates rather than simple stopwatch load-time tests.