For performance testing on both the X6 and the X8 I went through our normal test suite which touches on a few key synthetic benchmarks then I do a few real-world tests to give you a better idea of what you can expect for transfer speeds depending on different file types. Before doing that I did still run CrystalDiskInfo to try to get a peek at what was inside of the drives. It didn’t give me to much information, but it does let me document the firmware revision of both before testing. Besides, I already know what is inside of both drives. With our last X8, I learned that it had a P1 M.2 inside just by going off the firmware revision of the 1TB sample which was P3CR013 and was the same as the P1s firmware update at that time. The P3CR120 of this drive follows the same numbering but is a much later number. This doesn’t line up with the P2 which the firmware started with P2 and P4 so it seems like even with the higher capacity this is still a P1. Which isn’t a bad thing, the P1 is still faster than this USB interface can handle. As for the X6, I was told it is based on the X500 SATA drives, but the size of the drive as well as the firmware (M6CR030) tells us they had to do some work to make things fit in the small size.
My first tests were in CrystalDiskMark to get a good look at sequential performance mainly. While both do connect through the same interface, you can see a significant performance difference when you are comparing a lower-end NVMe drive against a SATA based drive. Of course, the connection is still the main limitation, our actual P1 pulled twice this in both the read and write speeds. But the X8 is solid right on the upper limits of the normal USB 3.2 Gen 2 which is 1250 MB/s before any overhead. The difference is especially noticeable in the random 4k 32 queue depth and 16 thread test where the X6 is ¼ of the speed on the read but died off on the writes. The controller must get overloaded at those levels. That said, with the pure sequential numbers the X6 still does really well and most importantly is worlds apart from a hard drive based drive (check this one out for example).
In ATTO Disk benchmark I wanted to get another look at the read and write speeds just to confirm things but I also wanted to check out the IOPS performance. The X8 did exactly what we saw in CrystalDiskMark again, as did the X6 including its write performance being on the lower side compared to the reads. As for the IOPS, the X6 did surprisingly well on the read IOPS but struggled on the writes which the high queue depth CDM test also clued us in on that. The X8 on the other hand didn’t stumble at all with its writes doing very well.
Now Anvils’s Storage Utilities is just a second look at the same thing we saw in CrystalDiskMark as well as the IOPS from ATTO and nothing changed here with the exception of the X6’s IOPS not being as high as ATTO indicated and the X8 doing the opposite and doing even better than ATTO. The X8 rear and write speeds in the sequential tests were a touch lower, down in the mid to low 900’s but still good.
The rest of my testing was focused on actual real-world performance. For this, I split the file types into three types. Large single files, specifically movies. For the middle of the road files I went with pictures, then the last file type was documents that I went with word documents. I then just transferred groups of those files to both the X6 and the X8 to see how they would perform, using real windows transfer speeds as an indication then transferred the same fold of files back on to a fast SSD on the PC doing the same again. All three file types are files that I would expect to see people using their portable drives to backup, travel with, or to transfer files to other devices.
The movie file transfers are the best-case scenario for both of the drives. This is what the sequential speeds in the other tests can relate to as it is one large file, not random files or file sizes. Copying our movies to the drives are write speeds and you can see that with the X6 which transferred at 300MB/s nearly the entire time where the X8 was slower than the synthetic tests at 689 MB/s. They both were faster when copying those same files back to the PC with the X6 sticking right at 496 MB/s the entire time and the X8 was flat as well at 869 MB/s. Both are more than fast enough to move these files around, although the X8 is going to save you a little extra time.
X6 Movies to drive
X8 Movies to drive
X6 Movies to PC
X8 Movies to PC
For pictures, these were just copies of pictures used in my reviews which include RAW, edited, and resized pictures for a variety of file sizes and you can see that a lot in the transfer speeds. Especially on the X6 which was about half the speed of its movie write performance at 143 MB/s at its peak speeds. The X8 did better here at 530 with a lot fewer waves. For the read speeds which are moving back to the PC, the X6 didn’t improve much with its average coming closer to 168 with a few peaks higher. The waves are again visible on it as well. The X8 transferred the files at 644 MB/s which was much faster and again there weren’t as many visible waves.
X6 Pictures to drive
X8 Pictures to drive
X6 Pictures to PC
X8 Pictures to PC
The last test with the folders filled with Word documents is a lot more demanding on every drive. Even just deleting the generated files after this test is a headache even though their file size is so small. These small files and high quantities of files are what kill transfer speeds, even with the fastest drives. Copying these to the X6 saw a whopping 246 KB/s which is under even a MB/s. The X8 did better but not well at 1.4 MB/s which could be a lot better as well. The last hard drive portable I tested performed similar to the X8 on this type of transfer and I’ve seen other NVMe based drives see around 3-8MB/s. So both drives fall off here, but the X6 is especially slow with this type of load.
X6 Documents to drive
X8 Documents to drive
X6 Documents to PC
X8 Documents to PC
While doing my testing I did also jump into AIDA64 and used their linear read speed test which I can run for longer periods to do thermal images of both drives after 20+ minute long transfers. Neither drive got too hot. The X8 with its metal casing did a good job of also puling the heat out as well.