Storage Game Loading Test: PCIe 4.0 SSD vs. PCIe 3.0 vs. SATA vs. HDD

QuantumPhysics

Posts: 3,719   +3,614
I typically used 2TB Crucial MX500 SATA SSD drives for my OS, games and cloud service files. I’ve been very happy.

I recently bought a single Samsung 8TB drive to consolidate those 2TB drives. My laptop came with a 256GB Samsung M.2 and a 500GB HDD - which I replaced with a Crucial MX500.

I saw no difference in loading times when playing DCS WORLD or CS GO. Nothing measurable without a benchmarker. I am just happy to be able to put more files in the same drive. I only use an HDD to backup important files and
large video files now.

I focus on capacity rather than theoretical speed. SSD is already so much faster than HDD that I am completely satisfied with the mere switch from HDD to SSD.

As you point out, most games were designed for HDD rather than SSD and get tremendous speed boosts.

A single 2TB or 4TB SSD should be more than enough for a gaming laptop owner or a gaming desktop. $200 - $400 right now, respectively, ain’t that bad.
 
Last edited:
  • Like
Reactions: Reehahs and PEnnn

Achaios

Posts: 61   +203
Dunno abt you, but for me shaving off 5 secs from a loading screen is a big deal.

There's ppl who dump north of 2000 euro on GPU's and they're proud of it too (ahahahaha), why shouldn't anyone pay a couple of hundred bucks more for PCIe 4.0 and a PCIe 4.0 SSD if they can afford it?
 

Heavens

Posts: 9   +5
Dunno abt you, but for me shaving off 5 secs from a loading screen is a big deal.

There's ppl who dump north of 2000 euro on GPU's and they're proud of it too (ahahahaha), why shouldn't anyone pay a couple of hundred bucks more for PCIe 4.0 and a PCIe 4.0 SSD if they can afford it?
I'd say it depends, maybe on something like Skyrim where you have loadscreens galore especially in cities.
If the game only rarely has loadscreens I'd take a bigger SSD over a faster one every time.
Obviously personal choice, I can wait a couple of seconds here and there, drink some coffee while I wait out those <20 seconds on my SATA SSD.
 
  • Like
Reactions: Bamda

Avro Arrow

Posts: 354   +375
This is one of the most interesting videos that I've seen Tim make because it's not something that is often benchmarked so Bravo Tim and good on ya!

I bought a WD Blue SN550 1TB NVMe drive (jeez, these designations are getting long!) as my gaming drive because, for some odd reason, it was exactly the same price as the SATA version. I may not be getting better load performance out of it (yet) but it sure copies faster! LOL
 

Avro Arrow

Posts: 354   +375
Placebo effect chasing a couple seconds here or fractions there for the average desktop user. Upgrade when it makes sense unless you are a benchmark addict, then go for it!
Well, the load time difference between a mechanical drive and an SSD is DEFINITELY noticeable. Between different SSDs, nope.

Interestingly, this same concept is why I was always happy with my FX-8350 CPU. Sure, the Intel chips were definitely faster but I'm primarily a gamer and the difference in load times for programs, while it looked HUGE on graphs, was really only a few seconds and not noticeable. Frame rates on games were just fine (hell, the FX can still do over 60fps for most games TODAY).
 

fps4ever

Posts: 555   +635
Well, the load time difference between a mechanical drive and an SSD is DEFINITELY noticeable. Between different SSDs, nope.

Interestingly, this same concept is why I was always happy with my FX-8350 CPU. Sure, the Intel chips were definitely faster but I'm primarily a gamer and the difference in load times for programs, while it looked HUGE on graphs, was really only a few seconds and not noticeable. Frame rates on games were just fine (hell, the FX can still do over 60fps for most games TODAY).
My post wasn't directed towards mechanical spinny hard drives at all. Not in the slightest...;) ?
 
A well written, informative, well researched article. I honestly didn't even know they still existed tbh.

I would like to see the numbers on the Samsung 970 and 980 too. Also AMD motherboards are talking about some software to use drives as cache (AMD StoreMI Technology), could you talk about it, a pc engineer I know says such things are generally very bad but I'd love to know what you think.
 

dirtyferret

Posts: 643   +753
Excellent review and one of the best techspot reviews in years. I'm sure a lot of fanboys are cursing you out as they read the facts
 

pildorman321

Posts: 14   +9
removed my 1TB Ex950 and SN550 till they are required, no real speed difference loading from a slow *** tinny crucial sata drive
 

brucek

Posts: 574   +685
TechSpot Elite
Love the article, thanks.

I believe every graph and it's enough to convince me not to spend a lot more on a higher class of drive. I'm still going to break same-ballpark ties in favor of performance, because the upside/downside trade-off there is strongly tilted: I won't care too much if I wasted an extra $20, but if there's ever a use case where it helps, I'd be bummed not to have it.

I also remember when SSDs first came out, and in those first couple years there were plenty of credible sounding articles where the author swore up and down that SSDs made no difference (often by using just FPS measurements, which was of course never the point.) Meanwhile, I had two systems back then, one with SSD and one without, and even on every day tasks the two just felt night and day different. It's left me feeling that whenever someone says storage speed doesn't matter, what they really might mean is they haven't figured out how to test or quantify user experience.
 
  • Like
Reactions: fireworksordie

BadThad

Posts: 314   +241
Excellent article! I've been saying there's no noticeable difference between a SATA SSD and M.2/NVME in game load times.....now I have the proof. A friend I game with has a much newer and faster system than me - including an Optane drive. He gets in games about 1 second before me and I am using an old Samsung 750GB SATA SSD with a 4770k.
 

Lew Zealand

Posts: 1,624   +1,684
TechSpot Elite
We're all rocking 1-2TB SATA SSDs in the house for gaming because for years Tom's Hardware used to add in game loading tests to their SSD reviews (haven't seen them recently, though) and the differences were about what Tim sees here. In fact, the differences were actually a bit smaller but that's 2+ year old PCI 3.0 NVMe drives.

Funny thing-- the best SSD in the house, a 1TB HP EX920 NVMe, is in a NUC. [facepalm] At least it's connected to an eGPU...
 

Avro Arrow

Posts: 354   +375
Why there is no Samsung 970 or 980 in the benchmark?
All of the PCI-e 3.0 and 4.0 models showed the same speeds which means that the bottleneck is the CPU itself. The load time would somewhere between 15.7 and 18.1 seconds. It wouldn't have made a difference.
 

defaultluser

Posts: 49   +49
Its going to be several years before we have more than one game that actually utilizes more than SATA6 bandwidth. By that time, my systems will be ready for a PCIe4 update.

But I told all those folks worshiping that PS5 SSD - it's going to be processor-limited in most cases. Even the R9 3900X is not good-enough to more than slightly exceed the SATA6 drives in a couple games.

If current console-pushing games can't really show more than 10% boost over SATA6 SSD, then it's going to be another 5 years before optimized PS5 titles push us over that load bandwidth hump. And even then, it's going ot be satisfied by cheaper PCIe 3 drives (because they will hit that CPU limit way before PCIE4.)
 
  • Like
Reactions: MWR01

BSim500

Posts: 706   +1,500
Good review. In the era of ever-bloating game sizes, I definitely don't regret buying a 2TB MX500 for the same price as a 1TB 970 EVO Plus. I'd also like to see Real World benchmarks (OS boot, game load times, etc), actually become the main focus of future storage reviews more than theoretical throughput marketing p*ssing contents.
 
  • Like
Reactions: PEnnn

Adhmuz

Posts: 2,064   +855
Honestly not overly surprised from these results, I experienced them all first hand, short of the PCIE 4.0 SSD. Had games on a mechanical drive, moved to an SSD, much faster. Moved to 3 SSD RAID 0, no appreciable difference, move to an entry level PCIE SSD comparable in raw read speed as the SSD Raid 0, again no appreciable difference. Tried once more to an Adata XPG Pro 1TB, same as in these test, and once more, no appreciable difference. I can't say that 1 to 4 seconds is noticeable.
 

mattferg

Posts: 82   +53
Somewhat pointless testing the Samsung T5 - open it up, it’s just an 860 Evo mSATA with a USB interface plugged into it.

Think the T5 was just an easy way for Samsung to ditch old mSATA stock when M.2 came out.
 

PEnnn

Posts: 477   +415
I was wondering how slow my recently bought (and dedicated entirely to the Steam folder) Crucial 2TB MX500 SATA drive is, when compared those expensive PCIe 4.0 / 3.0 SSDs.

Turns out, I could live with 2-5 seconds load time difference between the MX500 and the tops dogs. And saving a couple of hundred $ is even as sweet!
 

flipp3r

Posts: 26   +6
To those that want to build there own systems, never touch the contacts with your hands. Your skin secretes oils. Soooo many times I've had to clean the contacts of M.2's & ram modules because of greasy fingers. This could cause detection issues/dropouts/ram errors. Why would you hold an M.2 by its contacts?
 
  • Like
Reactions: Reehahs

DonquixoteIII

Posts: 14   +8
A few comments: First, your person showing off the SSD's for the camera needs a bit more education on handling electronics... No obvious grounding, touching bare connectors? Tsk...

About drives... NO drive does well with small file sizes. Games typically have many small files Notice that, except for the spinning bucket of rust, all solid state drives loaded games at almost the same rate? Perhaps if you had first shown (say) ATTO's tests on your drives all might have become clear. Check out ANY drive on file sizes under 4k. Yes, there will be speed differences, but not as big a gap as you might think. Now look back at your game loading tests, and note the speed differences between an SATA interfaced drive and an NVME interfaced drive. Or better yet, look at your various games, note the file sizes as a percentage of load times. What struck me most was not how poorly the hard drive did, but how close it was in game loading performance to the fastest NVME M.2 drives tested.

Try this: Make a gigabyte's worth of files in the range of 128 to 512 bytes, then record the speed of reading and writing that folder with the same set of drives.

So, from a pure gaming viewpoint, the slowest SATA drive is good enough for most gamers to not notice the difference, except to their pocket. Gaming is just not the best use case to define drive speeds.
 
  • Like
Reactions: defaultluser