- Sandisk reduced SSD preconditioning times from days to just hours
- SPRandom achieves stable performance using a single full disk write
- Open source SPRandom gives hyperscalers faster deployment options
Sandisk has made available a key open source algorithm that can significantly speed up the SSD preconditioning process.
SanDisk Pseudo Random (or SPRandom) reduces the time it takes to prepare a drive for steady-state operation from over 160 hours to approximately 6.5 hours.
Traditional preconditioning methods require writing data at least twice the disk’s total capacity using sequential and random operations. But the new algorithm writes only once to each logical address, thus completing the entire process in less than five percent of the initial time.
SSD preconditioning is important for data centers and AI workloads
Fresh-out-of-the-box SSDs exhibit variable performance, until they go through a process called preconditioning that stabilizes their behavior.
During this process, the controller fills the disk, starts garbage collection, handles wear leveling, and distributes overprovisioned blocks to the storage area.
This background activity must reach a steady state before disk I/O performance becomes predictable and reliable for production use.
Traditional preconditioning of a 128TB drive takes more than 160 hours, or nearly seven days, but Sandisk’s SPRandom accomplishes the same task in just 6.5 hours.
A 256 TB drive requires up to 250 hours, or about 10.5 days, using conventional methods, while SPRandom completes in just 6.5 hours.
These numbers imply that Sandisk’s SPRandom reduced preconditioning time between 95% and 97.4%.
The SPRandom algorithm divides the disk into overlapping sections, with the degree of overlap corresponding to the expected overprovisioning for each section.
As the number of physical addresses increases, the amount of overprovisioning gradually decreases across the entire disk.
The math behind SPRandom calculates how overprovisioning is distributed after preconditioning, ensuring stable performance in a single write pass to physical disk.
Sandisk has released the SPRandom code as an extension to the FIO tool, which stands for Flexible IO Tester, making it freely available to the entire storage industry.
It’s important for AI but not for your gaming PC
Hyperscale data centers and AI infrastructure operators are purchasing SSDs in massive quantities and need them to be production-ready as quickly as possible.
Reducing preconditioning time per disk by more than 150 hours directly translates to faster deployment and reduced operational costs for cloud providers.
AI workloads are particularly sensitive to storage performance variability because training and inference tasks require consistent I/O behavior across thousands of disks operating in parallel.
But, for a typical single-drive gaming PC, spending an extra week preconditioning a new SSD simply isn’t relevant to the end-user experience.
Technology delivers enormous value at scale, but the average consumer will never notice the difference.
Sandisk’s decision to open source the algorithm is truly generous, but the beneficiaries are data center operators with massive storage arrays.
Simply put, the math is nifty, the time savings are real, and the impact on AI infrastructure could be substantial.
However, your gaming PC will continue to perform exactly as it always has, and that’s precisely the point.
Via blocks and files
Follow TechRadar on Google News And add us as your favorite source to get our news, reviews and expert opinions in your feeds.
