it feels like random read/write numbers have been stuck for 5 years. 80/266mb/s when my m9pe did 66/220mb/s on z97. it's faster,but it's baby steps [SPOILER] https://static.tweaktown.com/content/9/9/9918_24_plextor-m10p-2tb-ssd-review-new-performance-leader.png [/SPOILER]

Thanks for the review Hilbert, interesting to see new controllers on the market! Even if the main takeaway from SSD reviews seem to be that real-world performance differences between years-old 500GB SATA drives and bleeding edge 2TB PCIe 4.0 drives is negligible at best. Perhaps DirectStorage and similar technologies will change that down the line, let's hope so, but until then I'd go with one of the cheaper 3D TLC options and put the saved money towards components were it makes a tangible difference today. The price delta to newer PCIe 4.0 drives have been going down of late, locally at least, but it's still nowhere near the actual performance difference.

And how long before the ' Plextor M10P 2TB NVMe PCIe' starts using inferior, cheaper controllers...without telling anyone.

What we need is applications. Our favourite type, Games to be able to use these drives effectively. If the keep optimising apps for spinners then whats this for. Maybees the new consoles will lead the way for games to better use an SSD as they do not feature spinners..

PsyaNyde:

And how long before the ' Plextor M10P 2TB NVMe PCIe' starts using inferior, cheaper controllers...without telling anyone.

never my first M.2 was a Plextor, they were the first to put in heatsinks and they were the first used by Enterprise they sell much closer to MSRP than other brands and the premium comes through in the products.

cucaulay malkin:

it feels like random read/write numbers have been stuck for 5 years. 80/266mb/s when my m9pe did 66/220mb/s on z97. it's faster,but it's baby steps [SPOILER] https://static.tweaktown.com/content/9/9/9918_24_plextor-m10p-2tb-ssd-review-new-performance-leader.png [/SPOILER]

yeah but what do you have to say about SATA:p seriously tho' none of that if the fault of the drive maker. that's squarely on standardization, (as seen) controller hardware, and the cost and complexity of memory holding back innovation

Exodite:

Thanks for the review Hilbert, interesting to see new controllers on the market! Even if the main takeaway from SSD reviews seem to be that real-world performance differences between years-old 500GB SATA drives and bleeding edge 2TB PCIe 4.0 drives is negligible at best.

I really can't agree with that, sorry....;) Prior to moving to an SSD couple of years back, my Windows 10 boot time was around 60-70 seconds from a 7200 RPM 2 TB Sata3 hard drive, to ~15-20 secs with my first SSD. Now it's 10-12 seconds for a cold boot with my NVMe 500GB Samsung 980 Pro--and it's on the slow side for the 980 Pro family. Doesn't get more real-world than booting, imo. I notice a huge difference in real-world use. The real fact seems to be if there was no real-world performance difference between SSD's and HDD's, then who would be buying SSDs of any type as they would run no faster, while being far more expensive per TB? But SSDs are being bought up rapidly. Actually, there is quite a difference in real-world performance between an NVMe SSD and a S3 HDD. I've got some very disk-intensive games that run noticeably faster from my SSD than they do from my fastest SATA 3 HDD. Even my PCIe3 SSD is significantly faster in the real world than my fastest HDD. I have no idea where you'd get an idea like that about HDDs...? I still use HDDs--I have four of them installed right now alongside 2 NVMe SSDs, and I can honestly say the real-world performance difference is substantial--as one should immediately think it would be. Maybe you're talking about the fact the 7GB/ps speeds possible in benchmarks for some PCIe4 SSD's doesn't equate to real world average performance but is only a theoretical performance number--and you'd be correct. But, you won't even get that theoretically out of a Sata3 HDD. Anyway...I have to disagree for reasons given.

waltc3:

I really can't agree with that, sorry....;) Prior to moving to an SSD couple of years back, my Windows 10 boot time was around 60-70 seconds from a 7200 RPM 2 TB Sata3 hard drive, to ~15-20 secs with my first SSD. Now it's 10-12 seconds for a cold boot with my NVMe 500GB Samsung 980 Pro--and it's on the slow side for the 980 Pro family. Doesn't get more real-world than booting, imo. I notice a huge difference in real-world use. The real fact seems to be if there was no real-world performance difference between SSD's and HDD's, then who would be buying SSDs of any type as they would run no faster, while being far more expensive per TB? But SSDs are being bought up rapidly. Actually, there is quite a difference in real-world performance between an NVMe SSD and a S3 HDD. I've got some very disk-intensive games that run noticeably faster from my SSD than they do from my fastest SATA 3 HDD. Even my PCIe3 SSD is significantly faster in the real world than my fastest HDD. I have no idea where you'd get an idea like that about HDDs...? I still use HDDs--I have four of them installed right now alongside 2 NVMe SSDs, and I can honestly say the real-world performance difference is substantial--as one should immediately think it would be. Maybe you're talking about the fact the 7GB/ps speeds possible in benchmarks for some PCIe4 SSD's doesn't equate to real world average performance but is only a theoretical performance number--and you'd be correct. But, you won't even get that theoretically out of a Sata3 HDD. Anyway...I have to disagree for reasons given.

99.99999999% sure they're referring to SATA SSDs versus their significantly "faster" NVMe counterparts. They're not wrong either: in a chunk of real world application they're not significantly faster despite having a huge read/write advantage. This largely comes down to NAND being the limiting factor as all one has to do is look at what Optane does. The P5800X completely destroys everything on the market when it comes to random, and that's because of 3D Xpoint being significantly better than NAND.

waltc3:

I really can't agree with that, sorry....;) Prior to moving to an SSD couple of years back, my Windows 10 boot time was around 60-70 seconds from a 7200 RPM 2 TB Sata3 hard drive..

As thestryker noted I'm talking about SSDs, not HDDs. Looking at Hilbert's review the examples of real-world performance are all but identical between all the different stripes of SSDs, only the synthetics show any actual difference. I don't claim there's no performance advantage over HDDs, though personally I were definitely surprised how little difference I noticed between my 5400RPM HDD and a SATA SSD. Though I only cold boot a couple of times a year, maybe that's the reason. 🙂

thestryker:

99.99999999% sure they're referring to SATA SSDs versus their significantly "faster" NVMe counterparts. They're not wrong either: in a chunk of real world application they're not significantly faster despite having a huge read/write advantage. This largely comes down to NAND being the limiting factor as all one has to do is look at what Optane does. The P5800X completely destroys everything on the market when it comes to random, and that's because of 3D Xpoint being significantly better than NAND.

I have to agree with this. I´ve just replaced my Samsung 970 Evo 500g SSD for an Western Digital SN 850 NVME 1tb and the diffferences in normal usage are basically zero, as far as i can tell. So far the only differences i´ve noticed are in gaming, load times are significantly better now but other than that, i struggle to see any difference.

H83:

So far the only differences i´ve noticed are in gaming, load times are significantly better now but other than that, i struggle to see any difference.

Out of curiosity, which games? As that's generally one of the metrics that sees little to no difference between SATA and NVME drives so I'm interested to see if it's something I can test myself. Also worth keeping in mind there might be unrelated reasons as well, a mostly full disk can't do proper garbage collection and tends to perform worse for example.

Exodite:

Out of curiosity, which games? As that's generally one of the metrics that sees little to no difference between SATA and NVME drives so I'm interested to see if it's something I can test myself. Also worth keeping in mind there might be unrelated reasons as well, a mostly full disk can't do proper garbage collection and tends to perform worse for example.

I´ve only tried a few games because i did a fresh install. Speaking of those games, Path of Exile loads are almost instant, with the Samsung 970 evo they would take 15 to 30 seconds. Doom takes between 15 to 25 seconds to load, before it would take between 30 to 60 seconds. Cyberpunk loads in 10/20 seconds, something i consider amazing considering we are talking about an huge open world. I still have to try more games but the improvements are significant so far in this aspect. But to be interily fair, i have to say i did a fresh Windows install and i move from W7 to W10 but i doubt Windows "responsability" in the improvement of loading times because of all the crap running in the background. In W7 i would have around 75 active processes running, in W10 i have around 145 processes... I would love to know what all those processes are doing...