Disk Copy Benchmarks

AS SSD Copy Benchmark

We are using AS SSD’s Copy Benchmark tool to test the duration of file copy for ISO, Program, and Game data. The benchmark outputs the duration in minutes and seconds, and lower is better.

Samsung 970 EVO Plus NVMe M.2 SSD 500GB AS SSD Copy Benchmark ISO

Lower is better, and the Samsung 970 EVO Plus 500GB comes in at 0.49 seconds, which is very close to the 970 EVO 1TB at 0.47 seconds, which is technically faster. There are cache size differences between the SSDs, but really this shows not much difference in large file transfer between them. The Teamgroup MP34 falls way behind here.

Samsung 970 EVO Plus NVMe M.2 SSD 500GB AS SSD Copy Benchmark Program

The Samsung 970 EVO Plus is technically faster than the Samsung 970 EVO 1TB SSD, but only very slight, they are pretty much neck and neck. The ADATA XPG SX8100 falls behind here quite a bit in small file sizes.

Samsung 970 EVO Plus NVMe M.2 SSD 500GB AS SSD Copy Benchmark Game

In-game transfers, using a combination of small and large file sizes, there is a technical slight advantage toward the new Samsung 970 EVO Plus over the Samsung 970 EVO, but it is very small. The ADATA XPG SX8100 falls behind here.

DiskBench

With DiskBench we place a 50GB single file onto the test drive. We copy this file from the test drive to a new folder on the test drive, volume to volume. This tests the performance the drive is able to copy a file from itself to itself as if you were copying files on the same drive. Results are in duration Seconds, lower is better.

Samsung 970 EVO Plus NVMe M.2 SSD 500GB DiskBench

This one was a bit of an oddity, copying a very large 50GB was slower on the Samsung 970 EVO Plus 500GB SSD versus the Samsung 970 EVO 1TB. There are reasons for this, one is the physical size of the SSDs, with the lesser 500GB capacity we are filling up more of the drive with a 50GB file and really exhausting those cache buffers, especially the SLC cache and DRAM capacity. There also may be latency differences between the two. The difference is not extreme. One thing is for sure, the Teamgroup MP34 performance is very extreme.

Brent Justice

Brent Justice has been reviewing computer components for 20+ years, educated in the art and method of the computer hardware review he brings experience, knowledge, and hands-on testing with a gamer oriented...

Join the Conversation

3 Comments

  1. Not a bad little drive, and they have come down significantly in price.

    Right now I am struggling when it comes to what drive to use for an SSD cache drive on a server, where constant heavy writes are a factor.

    WAY back I would have had to pay big bucks for a tiny SLC drive for something like this. Can’t find those anymore.

    More recently Samsung’s Pro drives were MLC and had some pretty serious write endurance.

    The latest gen Pro drives (980 Pro) now appear to be TLC, and it has me a little concerned.

    That said, these things are so cheap now, so maybe I’ll just get a couple of 500GB or 1TB Inland Premium drives for $59.99 or $119.99 respectively, and just beat the crap out of them until they are worn out and replace them.

  2. The 860 Pro drives are MLC, and MLC is def the way to go for write endurance, they may be SATA, but if you are after that endurance prices are cheaper now on them. You are right, Samsung has gone a little backwards with the 980 Pro, being TLC, Pro to me always meant MLC on Samsung, not happy with that trend.
  3. The 860 Pro drives are MLC, and MLC is def the way to go for write endurance, they may be SATA, but if you are after that endurance prices are cheaper now on them. You are right, Samsung has gone a little backwards with the 980 Pro, being TLC, Pro to me always meant MLC on Samsung, not happy with that trend.

    Yeah, I am torn.

    The truth is, that with each generation of controller, NAND quality and whatever magic makes the controllers work (write amplification, wear balancing, DRAM cache’s etc.) improves. There is a reason you essentially can’t buy an SLC drive anymore. They just aren’t necessary. MLC got better to the point where it could fill that role.

    The question is, have we gotten to the point where TLC is really ready to supplant MLC in high write applications?

    My old 512GB Samsun 850 Pro SATA drives I have been using as write cache for years are MLC and are rated at 150TBW each. At 69,000 power on hours, and 317110382829 LBA’s (~147.6 TB) written they are both listed at a wear leveling count of 30%, so they are starting to get close. (well, I mean, if 70% wear came in 69000 hours, that means I have ~30,000 hours or 3.5 years left, but I don’t want to push it TOO far)

    The aforementioned Inland Premium drives are Phison E12 TLC drives. The 512GB model (to keep it as close to an "apples to apples" comparison as possible) is rated at over 5x the write endurance, at 780TBW.

    If these numbers are accurate, and measured the same way Samsung did on my old 850 Pro’s, maybe MLC really is no longer needed? I mean those old MLC 850 Pro’s are going to give me a projected final lifespan of 11.25 years in my high write environment. If the Inland Premiums truly get 5.2x longer life, that should give me 58.5 years. I don’t know if I’ll be around in 2080 (probably not unless we see some amazing medical progress!), but I suspect my current server build will be long obsolete…

Leave a comment