Sequential Read Performance

The sequential read test requests 128kB blocks and tests queue depths ranging from 1 to 32. The queue depth is doubled every three minutes, for a total test duration of 18 minutes. The test spans the entire drive, and the drive is filled before the test begins. The primary score we report is an average of performances at queue depths 1, 2 and 4, as client usage typically consists mostly of low queue depth operations.

Iometer - 128KB Sequential Read

The heatsink makes a big difference for sequential read speeds: the M8PeY is essentially tied with the Samsung 960 Pro and EVO (both of which have the advantage of higher capacity). Without the heatsink, the M8PeGN delivers average performance for MLC-based PCIe SSDs.

Iometer - 128KB Sequential Read (Power)

The M8PeY with its large heatsink (and LEDs) draws significantly more power than any of the M.2 PCIe SSDs to deliver top-notch performance. Without the heatsink (but with thermal throttling) the M8PeGN is still one of the most power-hungry drives. Neither configuration offers great efficiency.

The sequential read performance of the M8Pe saturates at QD4 when the heatsink is used, but without the heatsink it is thermally limited for almost all of the test.

Sequential Write Performance

The sequential write test writes 128kB blocks and tests queue depths ranging from 1 to 32. The queue depth is doubled every three minutes, for a total test duration of 18 minutes. The test spans the entire drive, and the drive is filled before the test begins. The primary score we report is an average of performances at queue depths 1, 2 and 4, as client usage typically consists mostly of low queue depth operations.

Iometer - 128KB Sequential Write

As with random writes and sequential reads, the heatsink makes a big difference for sustained sequential writes. Without the heatsink, the M8Pe's high power consumption leads to enough thermal throttling that it is the slowest MLC PCIe SSD in the bunch, but still substantially faster than SATA. With the heatsink, speeds are over 50% faster and the M8Pe is on par with the Intel SSD 750.

Iometer - 128KB Sequential Write (Power)

The power consumption situation is very similar to what we saw on the sequential read test. The thermally-limited M8PeGN draws a similar amount of power to the Samsung 960 Pro, while the M8PeY draws much more, and neither wins a prize for efficiency.

Without a heatsink, thermal throttling starts very early on this test, but the M8PeGN gets in enough of a burst at the beginning to have a noticeably better average at QD 1 than later in the test. With the heatsink, the M8PeY shows mostly flat performance across the entire duration of the test.

Random Performance Mixed Read/Write Performance
Comments Locked

64 Comments

View All Comments

  • The_Assimilator - Wednesday, December 14, 2016 - link

    Conclusion page: the paragraph "The performance differences between the Plextor M8Pe and the Toshiba OCZ RD400..." is repeated. BRO DO YOU EVEN EDITOR?
  • Billy Tallis - Wednesday, December 14, 2016 - link

    I think the problem may have been too much editor. My browser was getting really laggy and unresponsive while I was finishing up the article and rearranging things.
  • Threnx - Friday, December 16, 2016 - link

    Whoa there are still readers here? hah, I was cleaning out my bookmarks and saw anandtech. You realize all the talent left ages ago right? This site is dead. They're just riding on the name now...
  • cbrownx88 - Friday, December 16, 2016 - link

    @Threnx - where do you go now since Anandtech is a shell of its former glory?
  • TemjinGold - Wednesday, December 14, 2016 - link

    So... it basically loses across the board to the EVO but they want to charge more for it?
  • Billy Tallis - Wednesday, December 14, 2016 - link

    Keep in mind that the 960 EVO results here are for the 1TB model, and that kind of capacity advantage usually brings some performance advantage, too. But yeah, once the 960 EVO is actually shipping in volume, a lot of prices will probably have to come down. Unless the 960 EVO price goes up.
  • close - Wednesday, December 14, 2016 - link

    Prices for Samsung SSDs are rising due to the NAND shortage. If Plextor manages to resist this trend they might do well. After all most current workloads don't need anywhere near that kind of level of performance so a slight performance disadvantage is tolerable as long as it's reflected in the price.
  • BrokenCrayons - Wednesday, December 14, 2016 - link

    I wasn't aware of a NAND shortage. Do you happen to know the cause?
  • Samus - Wednesday, December 14, 2016 - link

    It's simply a supply and demand issue. Other causes are manufacturing process adjustments and lower yield issues often associated with transitioning to new processes.

    However, the shortage isn't as dire as 'close' makes it seem. Drive prices aren't necessarily increasing because of the shortage, but they aren't falling as they should be with the density improvements associated with TLC and widespread adoption of 3D VNAND. Samsung drives are increasing in price because vendors have noticed sharper demand for them than other drive, and I'm sure you can guess why based on this article alone...I'm not a Samsung fan, but they are currently the most balanced SSD products on the consumer market.
  • BrokenCrayons - Wednesday, December 14, 2016 - link

    Ah that makes sense. Thanks!

Log in

Don't have an account? Sign up now