Crucial P5 1 TB M.2 NVMe SSD

[W1zzard]

Crucial's P5 M.2 SSD is finally a design for the high-end NVMe market. The drive uses a completely new controller made by Micron we've never seen before. While synthetic results are just alright, the real-life performance numbers in our Crucial P5 review paint a much better picture.

Show full review

 

[jabbadap]

Throttling while read, erhm what? Maybe some well cooled desktop, but I would think twice putting one these on laptop.

 

[W1zzard]

You’re reading hundreds of gb in a minute on your laptop?

 

[Selaya]

Hm.

I wonder, what's the difference in methodology between these:

Like, unless I'm retarded shouldnt the SLC cache fill after like, uh ~35 seconds? Yet for the thermal test, the drive seems to happily write away at 3,000 MBps indefinitely.

Speaking of the cache, have you performed write test on dirty/full SSD (like, 25%/50%/75% filled)? Because there are models that have scaling SLC cache which gets smaller once the drive fills; that'd actually be a relevant real usage benchmark (I'd wager that nobody will keep their SSD empty forever, seems quite pointless ...) - the best benching SSD would be of little use if its performance plummets once it fills up.

Also, on another random note - unless I'm blind - you don't seem to have ever tested an Optane, right? Maybe it's just me (or rather, because I actually do own one, xD) but I'd love to see it as a reference on the benchmarks, just for scale, basically.

 

[bogami]

I missing momentum cache results with Crucial Storage Executive on the 1 TB model . TT, it has a review with the 2TB model.https://www.tweaktown.com/reviews/9548/crucial-p5-2tb-nvme-2-ssd/index.html .

 

[W1zzard]

I wonder, what's the difference in methodology between these

Nice, finally someone asking this Thermal test is writing to a limited size test area, so that SLC cache is taken out of the equation

you don't seem to have ever tested an Optane, right?

Correct, I have not. Working on PCIe 4 test rig right now, will look into Optane once that is ready

momentum cache

4 GB of RAM wasted to improve synthetic benchmarks and increase risk of data loss. Maybe one day I'll find time to test it, in actual real-life apps

 

[Lionheart]

Was considering of picking one of these up, tossing up between the 970 Evo 1TB or make use of my Gen 4 speeds with a Gigabyte 1TB M.2 drive but it's huge copper heatsink slug puts me off, doubt it would fit with the x570 chipset fan screwed on top of it.

 

[damric]

Has a chip of DDR4-4266?

I doubt that seriously. More likely it is DDR4-2666. Please do verify as I seriously doubt Micron is going to put anything other than JEDEC spec DRAM on an SSD.

 

[Maxx]

Has a chip of DDR4-4266?

I doubt that seriously. More likely it is DDR4-2666. Please do verify as I seriously doubt Micron is going to put anything other than JEDEC spec DRAM on an SSD.

MT53D512M16D1DS-046 IT | Micron Technologies, Inc

www.micron.com

 

[bug]

Throttling while read, erhm what? Maybe some well cooled desktop, but I would think twice putting one these on laptop.

Hear, hear. I always tell people not to worry about throttling while writing, but I expect reading a lot of data at once is a little more common than that.
Even more puzzling, read throttling seems to happen at lower temps then write throttling

 

[AnarchoPrimitiv]

Was considering of picking one of these up, tossing up between the 970 Evo 1TB or make use of my Gen 4 speeds with a Gigabyte 1TB M.2 drive but it's huge copper heatsink slug puts me off, doubt it would fit with the x570 chipset fan screwed on top of it.

I got a 1TB HP EX950 back in November for $120, and I've been completely satisfied and I use it for a lot of content creation workloads... Just check out TPUs review on it and you'll see that it's definitely a better choice than the 970 Evo and far cheaper.

 

[QUANTUMPHYSICS]

TECHPOWERUP

I am a long time user of Crucial's MX300 and now the MX500.

I don't know if you do testing with GAMES or 4K files under 5 hours long, but I'd appreciate it if you could put the conclusions in those terms because I seriously doubt I'd be buying a QLC drive for workloads/workstations and I'd specifically be buying it for games and storage capacity.

 

[R-T-B]

TECHPOWERUP

I am a long time user of Crucial's MX300 and now the MX500.

I don't know if you do testing with GAMES or 4K files under 5 hours long, but I'd appreciate it if you could put the conclusions in those terms because I seriously doubt I'd be buying a QLC drive for workloads/workstations and I'd specifically be buying it for games and storage capacity.

This is a TLC drive.

 

[Selaya]

[ ... ]

Speaking of the cache, have you performed write test on dirty/full SSD (like, 25%/50%/75% filled)? Because there are models that have scaling SLC cache which gets smaller once the drive fills; that'd actually be a relevant real usage benchmark (I'd wager that nobody will keep their SSD empty forever, seems quite pointless ...) - the best benching SSD would be of little use if its performance plummets once it fills up.

[ ... ]

What about this? I'm pretty sure I am in fact, not blind and haven't seen a test like this pop up on TPU's SSD reviews ...

 

[W1zzard]

and haven't seen a test like this pop up on TPU's SSD reviews ...

never tested this, nor have I seen anyone else test it. not sure if feasible, will explore

 

[Selaya]

I could've sworn that I have seen someone perform a test like that in their SSD review, but I don't remember where ... I'll try to delve a bit more in my Chromium history.

 

[Palladium]

Too late, too expensive.

 

[olstyle]

This is a TLC drive.

Which is why I find the performance compared to the P1 even more lacking.
The P1 is reaching nearly the same performance, even better IOPS, using a 4 channel controller and QLC flash.

 

[mahirzukic2]

The Crucial P5 is available in capacities of 250 GB ($55), 500 GB ($80), 1 TB ($150), and 2 TB ($340)

I am not getting something or it makes no sense for prices related to capacity to scale the way they do right now.
Since you have to use all the same components, PCB, Controller, etc. with the only thing memory cells (different number of them) for different capacities, wouldn't that mean that increasing capacities it would lower your costs per GB effectively? Even having for example 15c or 14c/GB for 2TB version would drive some people to buy the 2 TB version instead of 1 TB. Isn't this what you want as a company?
Once again, I am missing something here or am plain stupid?

 

[W1zzard]

it would lower your cost per GB?

Correct, the exception is if you need to use different flash that's more expensive, or a different controller, or a new (low-volume) PCB

 

[bug]

I am not getting something or it makes no sense for prices related to capacity to scale the way they do right now.
Since you have to use all the same components, PCB, Controller, etc. with the only thing memory cells (different number of them) for different capacities, wouldn't that mean that increasing capacities it would lower your costs per GB effectively? Even having for example 15c or 14c/GB for 2TB version would drive some people to buy the 2 TB version instead of 1 TB. Isn't this what you want as a company?
Once again, I am missing something here or am plain stupid?

Yes, the price of a product is not strictly its BOM.
We're being charged here for convenience, too. You could buy two 1TB drives instead of a single 2TB one, but do you have the space for them? Most mobos only have one or two M2 slots. You can add more, but probably not as cheap as the extra $40 you're looking at right here.
Drives in the upper capacity range have always been priced like that. It could also be because they're moving lower quantities.

 

[mahirzukic2]

It could also be because they're moving lower quantities

Ah, that explains everything, even more so than the price of a product not strictly being its BOM.

 

[Selaya]

never tested this, nor have I seen anyone else test it. not sure if feasible, will explore

I could've sworn that I have seen someone perform a test like that in their SSD review, but I don't remember where ... I'll try to delve a bit more in my Chromium history.

Belated update: Found it. Looks like it is Anand after all, my memory wasn't playing tricks on me:

The Best NVMe SSD for Laptops and Notebooks: SK hynix Gold P31 1TB SSD Reviewed

www.anandtech.com

They only test empty and full, but I'd argue that's already 100% more than just one test.

 

[TheDeeGee]

I'm looking to move from a traditional SSD to an m.2 for my new PC.

Already have my board, the Asus ROG Strix B550-E Gaming, since that comes with m.2 heatsinks i guess the heat isn't a concern anymore?

 

[Ricky78]

As part of their M.2 NVMe push, Crucial also announced the Crucial P2 SSD recently—our review is in progress.

Hello,
where is the P2 review?
I can't find it...