Feedback

Crucial P1 500GB and 1TB M.2 NVMe SSD Review - More QLC

Subject: Storage
Manufacturer: Crucial

Introduction

Once we saw Intel launch QLC flash installed in their recent 660p M.2 part, I had a feeling that Micron would not be far behind, and that feeling has been confirmed with the launch of the Crucial P1 M.2 SSDs:

View Full Size

Both the 500GB and 1TB models are single sided. The 2TB (not yet released) will likely have packages installed at the rear.

View Full Size

No surprises with the packaging. Does the job just fine.

View Full Size

Specs are also reasonably standard for an NVMe SSD at this point, though we do see a bit more of a falloff at the lower capacities here. This is partially due to the use of QLC flash, even though these specs are likely assuming full use of the available SLC cache. Since QLC allows for higher capacity per die, that translates to fewer dies for a given SSD total capacity, which lowers overall performance even at SLC speeds. This is a common trait/tradeoff for the use of higher capacity dies.

Read on for our review of the Crucial P1!

Now for the performance results. Details on the test suite are located here.

I made it a point to include the Intel 660p near the P1 results as both products are nearly physically identical, with the primary difference being down to firmware implementations between the two products.

View Full Size

The P1 turned in notably low random read performance, with the 500GB dipping just below 860 EVO (SATA) speeds.

View Full Size

Sequential results were good, with the 1TB cousins nearly matching in performance. Intel only sampled us the 1TB 660p, so we could not compare against the 500GB P1.

View Full Size

The 1TB P1 was competitive in the mixed burst test, but the 500GB appeared to dip a bit lower than expected. Let's look at the load time:

View Full Size

Whoa, that total read service time was indeed doing something odd for the 500GB part. Let's zoom in a bit further to look at the data where those figures are derived from:

View Full Size

This might be the first time I needed to present this data for an SSD. Our mixed burst test issues two passes each at three different levels of fill. Things did fine with the drive nearly empty (16GB written), and results were as expected at the 50% fill point as well, but when the 500GB P1 was nearly filled to capacity, we apparently caught it aggressively folding data from SLC to QLC while this workload was being applied, causing a significant performance hit. This is a fairly 'heavy' test, so the takeaway here is that if you suspect you are a power user type, it may be wise to not buy the smallest possible capacity QLC product, nearly fill it to capacity, and then simultaneously perform relatively heavy reads and writes within an hour of filling it :).

Alright, now for the cache test, which was *very* interesting:

View Full Size

View Full Size

Both of the P1 capacities tested here did *extremely* well in the cache test. Regardless of the idle time between applied test writes, the 1TB wrote at 1.6GB/s and the 500GB part consistently exceeded 900MB/s. It appears that Micron/Crucial chose to trade off some maximum performance for improved cache consistency, and the result is outstanding as every single 60-second write pass achieved maximum SLC performance. Compare this to the cache results we saw from the Intel 660p and things will become more clear. Remember, the main difference between those two models is just firmware tuning. What a difference!

Pricing (from Crucial direct):

  • P1 500GB - $90   ($0.18/GB)
  • P1 1TB     - $170 ($0.17/GB)

Warranty period is 5 years.

Looks like the Crucial P1 is a good choice for those looking for a solid budget NVMe SSD. Performance may not be as fire-breathing as some of the competition, but the caching is extremely consistent and the price is right as well. I'd recommend considering the Crucial P1 for consistent performance on a budget, and it can even tread into enthusiast territory so long as you go with a higher capacity model.

View Full Size


December 20, 2018 | 08:05 AM - Posted by Anonymous0 (not verified)

Getting closer to that 10c/GB...

December 20, 2018 | 12:12 PM - Posted by Cyclops

Optane DCPM review when?

December 20, 2018 | 02:47 PM - Posted by Allyn Malventano

The only way to test that is with a specialized server platform. Probably not any time soon unfortunately :)

December 20, 2018 | 04:21 PM - Posted by Cyclops

I'm sure Intel can arrange that if you ask them nicely. Free marketing.

December 20, 2018 | 12:16 PM - Posted by dragosmp (not verified)

The drive looks very good when compared to other QLC or entry level TLC.

About that "last few GB are slow" issue, it was a thing back in 2012. It was a hack back in the day of Jmicron and Sandforce not to partition a few GB to give the controller some extra space. Intel x25 didn't need it. Would that help for the P1? Back then it was never agreed if the choppiness of the full drives were because of controller hardware ineptitude or poor firmware, but it did work.

December 20, 2018 | 02:49 PM - Posted by Allyn Malventano

There's already OP (overprovisioning) at play here, but the issue was more along the lines of us catching the P1 busy trying to reshuffle its dynamic cache after filling its second half with data. Typical use should rarely if ever run into that particular scenario.

December 20, 2018 | 03:23 PM - Posted by Bryan (not verified)

Why would someone choose this over an mx500 which can be had for $118 for 1tb?

December 20, 2018 | 06:37 PM - Posted by Will o (not verified)

As per info on Amazon the p1 has sequential read/write of 2000/1750 MB/s and the mx500 has sequential reads/writes up to 560/510 MB/s

December 21, 2018 | 05:06 AM - Posted by lexx

Also mx500 is sata (m.2 sata or just sata ) m.2 does not mean its a nvme ssd (pci-e)

P1 is a nvme ssd (be interesting over time how reliable QLC is)

January 2, 2019 | 11:48 AM - Posted by Randy (not verified)

If I already use a SATA SSD (Samsung 850 EVO) for my boot drive would I see (from user perspective) any real performance difference here in boot times/app loading or does this fall into paper spec amazingness but just wait till time to replace?

January 13, 2019 | 07:25 PM - Posted by Debra Lance (not verified)

no

January 2, 2019 | 11:49 AM - Posted by Randy (not verified)

If I already use a SATA SSD (Samsung 850 EVO) for my boot drive would I see (from user perspective) any real performance difference here in boot times/app loading or does this fall into paper spec amazingness but just wait till time to replace?

January 10, 2019 | 12:39 PM - Posted by Phil (not verified)

So when will the 2 TB model be available?

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

CAPTCHA
This question is for testing whether you are a human visitor and to prevent automated spam submissions.