Ismar is a Senior Author at How-To Geek. He previously worked as a writer, editor, and general manager at a content agency before joining the team in 2023.
He began his writing career in 2021 after completing a BA in English Language and Literature. Beyond How-To Geek, Ismar has contributed to Red Stag Fulfillment and Authority Hacker. He has also worked on various SEO projects to help clients improve their search visibility.
Ismar has been around Windows PCs since the age of three, so friends and family naturally chose him as the resident tech support. His projects often involve tinkering, such as disassembling mechanical keyboards and gaming mice to mod his beloved gear.
He also enjoys pushing hardware to the max through overclocking while staying on a reasonable budget. An avid gamer, Ismar has logged thousands of hours across various genres, including first-person shooters, RPGs, racing games, and roguelites.
When he is not at his desk, he is probably exploring other people’s phones to discover their quirks and features or experimenting with the latest AI tools.
In his free time, Ismar enjoys spending time with his wife, working out at the gym, playing guitar and bass, cooking gourmet meals, and traveling. Above all else, he’s a proud cat owner.
You’ve just bought a new SSD and want to put it to the test, so you grab a large zip file and copy it between your two drives. Monitoring the transfer speed in real time seems like a straightforward way to see how fast your new SSD is.
The test does tell you something about the SSD’s performance, just not necessarily what you think it does about the drive itself. In reality, the speeds you see can be incredibly misleading due to several factors, including SLC caching.
Copying a file tests the entire transfer, not just the SSD
It’s only as fast as the weakest link
You’d think copying a file is about as real-world a test for an SSD as it gets, but it’s a long way from being a reliable benchmark. Even under ideal conditions, the transfer speed is limited by the weakest link.
If you’re copying the file from one folder to another within the same SSD, it’s not a great way to measure its peak performance because the drive is splitting its internal bandwidth between reading and writing data. And if the file is particularly large, or the SSD is already mostly full, it can forcefully trigger garbage collection, further contributing to reduced performance.
All these factors turn the test into a mixed read/write workload, and while it’s good for testing that specific task, you won’t see the drive reach its peak sequential read or write speeds.
Copying a large file from one SSD to the one you’re testing is a considerably better way to gauge its sequential write performance, but it introduces a different set of potential bottlenecks.
The available free space on both drives, the source drive’s speed, available bandwidth through the motherboard and connection, the USB enclosure and cable, RAM caching, background activity, and how Windows handles the transfer can all affect the results.
The bottom line is that this test measures performance in that particular scenario. The speeds you measure are relevant to that task, which can be useful if, for example, you frequently transfer files from your internal SSD to an external one, but they don’t tell you much about the SSD’s overall performance.
A large enough file transfer can reveal cached and sustained write speeds, but not overall SSD performance
It’s fast until it isn’t
While moving a single large file isn’t the be-all and end-all of SSD testing, it’s not entirely useless. Let’s say you’ve accounted for the main bottlenecks I discussed earlier: an empty SSD to write to, a large file that’s going to fill most of it, a source drive that’s fast enough to keep up, and a motherboard or USB enclosure with enough bandwidth to avoid limiting the transfer.
When you first start copying the file, you might see write speeds relatively close to the SSD’s advertised sequential write speed.
However, as the transfer goes on, you might see a sudden, drastic drop in speed. It’s not uncommon for transfer rates to fall from something as high as 4GB/s to just 80MB/s or even less. You might think something is wrong with your supposedly blazing-fast NVMe SSD, the cable, or your operating system, but this can be completely normal for modern TLC and QLC drives.
This happens because these drives often use something called a pseudo-SLC cache. SLC NAND flash stores one bit per cell, while TLC and QLC store three and four bits per cell, respectively. Storing fewer bits per cell allows for faster writes, so the SSD’s controller can use a portion of its TLC or QLC NAND as if it were SLC by temporarily storing just one bit per cell. This lets the drive write data much faster without the cost of dedicated SLC flash, before folding that data into its normal TLC or QLC format later.
The problem with this so-called pseudo-SLC cache is that it has a limited capacity. Once the available cache fills up faster than the drive can free space, write speeds can drop to the drive’s native TLC or QLC write speed, depending on its design.
These aren’t “fake” speeds. Pseudo-SLC caching allows SSDs to achieve blazing-fast transfer rates while remaining relatively affordable. The drawback is that you’ll eventually encounter the drive’s sustained write speeds if your transfer is large enough to exhaust the available cache.
It’s worth noting that the amount of available SLC cache can vary significantly between SSD models and capacities, as can their post-cache write speeds. Both affect how long a drive can maintain its initial transfer rate and how fast it writes once the cache fills up.
Moving small files exposes performance differences a giant transfer can’t measure
Small files follow different rules
The “real” problem with moving a single large file to test an SSD is that sequential speeds don’t tell you much about its random read/write performance, which can be much slower in terms of raw throughput and is often more relevant to everyday use.
For example, games and programs may need to load many smaller files or chunks of data from different locations rather than read a single file from start to finish. Copying folders full of small files can behave similarly, unlike copying one large file.
Interestingly, random read/write performance doesn’t always scale as much as sequential speeds with newer PCIe generations. That’s why you shouldn’t assume an SSD will perform better just because it advertises higher maximum bandwidth.
You need to look at independent tests of SSDs in your price range to determine which offers the best value, especially if you aren’t buying the SSD for its snappy sequential speeds.
Check the specs and run multiple tests to determine an SSD’s performance
Testing how quickly your SSD handles large file transfers can be a useful indicator of its cached and sustained write speeds, but you need to combine it with synthetic benchmarks and real-world workloads to get the full picture.
But before you even run tests, a lot of an SSD’s performance can be deduced from its spec sheet. TLC costs more, but it’s better in every aspect than QLC, and the presence of DRAM cache can also have a pretty noticeable impact on performance, especially if you keep your SSD mostly full at all times.
Discover more from PMN S.P.O.R.T.S - A PRIME MEDIA NETWORK BRAND
Subscribe to get the latest posts sent to your email.
