You spent your weekend researching parts, waited for the delivery truck to arrive, and finally installed a shiny new PCIe Gen4 NVMe SSD into your rig. The box proudly boasts sequential speeds of up to 5,000 MB/s, and your first five-second benchmark score seems to back up that claim.
Then, you decide to copy over a massive 150 GB Steam library, or perhaps a folder of raw 4K footage.
For the first twenty seconds, the progress bar flies. Suddenly, the transfer speed plummets from a blistering 4,500 MB/s to a crawling 150 MB/s, which is slower than a ten-year-old mechanical hard drive.
No, your drive is not broken, and your motherboard is not overheating. You have simply hit the brick wall of modern solid-state storage: SLC cache exhaustion.
Understanding how manufacturers hide the true, sluggish performance of raw flash memory is the key to avoiding a terrible purchase, especially when you are upgrading compact gaming handhelds or setting up high-end workstations.
The multi-bit memory trade-off
To understand why your drive slowed down, we have to look at how modern flash storage is manufactured.
Years ago, solid-state drives used Single-Level Cell (SLC) NAND flash, which stores exactly one bit of data per physical cell. This design is incredibly fast, highly durable, and extremely expensive.
To bring prices down, manufacturers developed Multi-Level Cell (MLC), Triple-Level Cell (TLC), and Quad-Level Cell (QLC) memory, which pack two, three, and four bits of data into every single microscopic cell.
This technological leap successfully lowered the price of high-capacity storage, making 2 TB drives incredibly affordable.
However, there is a massive catch. Writing multiple bits of data to a single cell requires different voltage levels, which takes significantly more time, precision, and processing overhead.
Raw TLC and QLC flash memory are inherently slow during data writes. If you were forced to write directly to raw TLC or QLC blocks at all times, your expensive new NVMe drive would feel absolutely sluggish.
Setting up the high-speed landing zone
To mask this structural speed bottleneck, drive designers use a clever workaround called an SLC cache.
The SSD controller is programmed to treat a portion of its slow TLC or QLC blocks as if they were fast, single-bit SLC blocks. When you start writing a file to your drive, the controller routes all incoming data directly into this high-speed landing zone.
This temporary buffer is why short, burst-heavy benchmark utilities show such impressive speeds. The software is only writing a tiny 1 GB or 4 GB test file, which fits entirely inside the fast cache.
But what happens when you transfer a massive video editing project, a raw game folder, or an uncompressed system backup?
The moment your file transfer exceeds the size of that temporary buffer, the cache is completely exhausted.
This buffer exhaustion forces the SSD controller into absolute panic. It must simultaneously accept new incoming data from your computer, read the existing data out of the SLC cache, erase the flash blocks, and write that data back into slow TLC or QLC states.
This double-handling process is known as “folding”, and it is the exact moment your transfer speeds collapse.
Benchmarking your drive’s true raw speed
If you want to expose the true limits of your storage hardware, you have to bypass the short-burst benchmarks that marketing departments rely on.
We highly recommend avoiding quick benchmarks like CrystalDiskMark for this diagnostic test. These tools only measure the speed of the cache, giving you a highly inflated, unrealistic view of daily file-transfer performance.
Instead, use a utility that performs a continuous linear write across the entire span of the drive, such as HD Tune Pro or the Linear Write Test in AIDA64.
When you run a sustained write test, you will see a clear, graphical representation of your drive’s performance profile. You will observe a high, flat line of peak performance for the first 10% to 20% of the drive’s capacity, followed by a sharp drop-off to a lower, jagged line.
This lower level represents the true, unbuffered speed of the raw flash memory.
Sustained speeds are particularly vital if you are running an expansive, high-end desk layout. Sprawling out your workflow across a massive panoramic display panel, similar to the ultrawide desktop environment we broke down in our Gigabyte M34WQ monitor review, is incredible for multitasking. However, your premium screen real estate will still feel completely bogged down if your storage subsystem chokes during a heavy background file transfer.
Measuring your storage write deficit
Once you have recorded your peak speed and your minimum sustained write speed, you can easily calculate the actual performance deficit of your drive.
To standardise solid-state drive performance metrics and prevent deceptive specifications, organisations such as the Storage Networking Industry Association publish strict, real-world rules on evaluating storage under continuous, steady-state workloads.
To find out exactly how much marketing fluff is built into your drive’s retail packaging, you can calculate the statistical discrepancy using a percent error calculator.
For this calculation, treat the advertised sequential write speed printed on the retail box as the theoretical value. Then, take the raw, sustained speed you recorded after cache exhaustion and enter it as the experimental value.
For example, let us look at a typical budget Gen4 drive like the Kingston NV2:
- Advertised Box Speed (Theoretical Value): 3,500 MB/s
- Sustained Raw Speed (Experimental Value): 150 MB/s
Plugging these numbers into the formula reveals a staggering write performance deficit of 95.7%.
This massive percentage means your drive is operating at less than 5% of its advertised capability during heavy, sustained workloads. Even high-end consumer drives often exhibit a write deficit of 40% to 60% once their cache folding begins.
How to choose your next storage upgrade
Does this mean budget drives with massive performance deficits are useless?
Not at all. If you are upgrading a secondary storage drive solely to store games, or if you are swap-upgrading a portable system, a budget drive is often perfectly fine.
For instance, when we analysed the storage changes in our ROG Ally X vs ROG Ally review and comparison, we noted how critical fast, reliable, and power-efficient storage is for handheld gaming. For game-loading workloads, write speeds do not actually matter because the system is almost exclusively reading data.
But if you spend your days editing heavy video projects or shifting massive local files around, grabbing a drive based purely on the sequential speed stamped on the packaging is a fast track to absolute misery.
When your workflow depends on constant, heavy writes, you need to hunt for TLC-based silicon that features dedicated DRAM, a high-grade controller, and a generous static cache. Skipping out on these three essentials means you are bound to hit that folding wall sooner rather than later.
Always look up third-party hardware reviews that feature full-drive sustained write graphs. If a reviewer does not show you what happens to the write speed after 100 GB of continuous transfers, they are only showing you the mirage.
