The decompression of graphics materials in Intel, AMD, and NVIDIA is improved by DirectStorage 1.1, however the blue team performs somewhat better.
On the website’s YouTube channel, the most current benchmark test created by Compusemble demonstrates the technology’s power first hand. For ease of watching, we have given the video below:
The API enables the CPU to use “highly parallel” graphics cards to decompress game content and reduce the number of key cycles required for major operations without the operating system taking control and exploiting the processor at a higher level. When the NVMe storage requests data, Microsoft DirectStorage 1.1 limits the strain exerted on the processor. The time it takes to load assets is largely constrained by the algorithms employed in the compression and decompression of assets using DirectStorage, which allow for the transfer of higher data levels than what the NVMe SSD is typically assumed to be able to manage.
In order to determine which graphics card handles data decompression the best, PC Games Hardware used Compusemble’s benchmark test to compare three of the best graphics cards currently available. The website evaluated the following three GPUs:
- Intel Arc A770
- AMD Radeon RX 7900 XT
- NVIDIA GeForce RTX 4080
The Alder Lake-generation Intel Core i9-12900K processor was paired with each of the three graphics cards. Decompression was done by all three graphics cards over 2.5 times faster than the Intel processor. The “asset decompression” performance of the Intel Arc A770 GPU was superior to that of AMD and NVIDIA. According to the statistics, the AMD RX 7900 XT processed the identical data at 14.6 GB/s whereas the Intel Arc A770 did it at 16.8 GB/s, a thirteen percent difference.
All three cards, however, decompressed assets equally because the loading times were slashed from five seconds to less than a second, indicating that either setup would still benefit from DirectStorage 1.1.
Sources: Tom’s Hardware, Compusemble, PC Games Hardware (PCGH)
