Test system and test preparation
I use the usual suspects such as CrystalDiskMark and Atto to check the theoretical data from the specifications. Due to previous workstation tests, these are no longer brand-new SSDs, but everyday goods that have also already been worn down quite a bit. Let’s see what remains of the theory in everyday life after the wear and tear. The SSDs to be tested are located in the first PCIe 5.0 NVMe slot of the motherboard and are not used as system disks.
I also use AJA as an everyday test to simulate the encoding of larger Ultra HD video streams and the SPECwpc storage test, which contains a lot of real applications and it will be interesting to see what performance remains with the large workloads. However, I picked out the applications with the biggest differences and loads as examples. I have also summarized the individual components of the test system in a table:
Test System and Equipment |
|
---|---|
Hardware: |
AMD Ryzen 9 7950X |
Cooling: |
Alphacool Eisblock XPX Pro Alphacool Eiswolf (modified) |
Case: |
Raijintek Paean |
Monitor: | BenQ PD3220U |
Thermal Imager: |
1x Optris PI640 2x Xi400 Thermal Imagers Pix Connect Software Type K Class 1 thermal sensors (up to 4 channels) |
OS: | Windows 11 Pro (all updates, current certified drivers) |
Sequential performance of the used SSDs
The synthetics are a good way to really run out the big numbers. We will see later how well this works in reality with the real application benchmarks. That’s why I’m starting with CrystalDiskMark and four different file sizes. The SSDs were no longer brand new at the time of testing (I always do these tests at the end for certain reasons) and I had already had fill levels of around 50% before deleting the data several times.
This certainly also explains the fact that the maximum values are missed by a very small margin when writing, but the figures are still impressive. Hence the wording with the “up to”. If we now compare the T-FORCE Cardea Z540 and the MSI Spatium M570 Pro, we see (almost) no difference! How could we, because both SSDs are identical. However, the T-FORCE Cardea Z540 has a slightly better performing firmware installed, which we will see later.
You can see very clearly that the dynamic pSLC does exactly what it should, mind you with an empty (albeit not virgin) SSD. The nice thing about larger SSDs is that there is always space left and it is therefore better never to fill them more than 2/3 full with data. A higher load does not affect reading, but when it comes to writing, the dynamic SLC will certainly reach its limits at some point. And if you do it over and over again, it will eventually no longer be possible to switch the memory modules between the two methods.
ATTO works in a very similar way, although I only work with two sizes here, which ultimately results in the same thing. The limit of 12000 MB/s is clearly missed, but it is still enough for an impressive result.
But what happens when you stream a video? The industry uses the AJA benchmark for this, which is effectively an interface between synthetic benchmarks and practical applications. The increase in the write rate differs significantly at the beginning, because the MSI M570 Pro 2 TB “pumps” a little less, but later the picture is similar.
Reading is also quite fast, although the theoretically possible speed is not reached here either.
I have attached a detailed AJA protocol for you:
BENCHMARK-TABLE
23 Antworten
Kommentar
Lade neue Kommentare
Urgestein
Mitglied
Mitglied
Mitglied
Mitglied
Veteran
Urgestein
Mitglied
Urgestein
Veteran
Veteran
Urgestein
Urgestein
Mitglied
Urgestein
Veteran
Urgestein
Mitglied
1
Alle Kommentare lesen unter igor´sLAB Community →