Power Management Features

Real-world client storage workloads leave SSDs idle most of the time, so the active power measurements presented earlier in this review only account for a small part of what determines a drive's suitability for battery-powered use. Especially under light use, the power efficiency of a SSD is determined mostly be how well it can save power when idle.

For many NVMe SSDs, the closely related matter of thermal management can also be important. M.2 SSDs can concentrate a lot of power in a very small space. They may also be used in locations with high ambient temperatures and poor cooling, such as tucked under a GPU on a desktop motherboard, or in a poorly-ventilated notebook.

MyDigitalSSD SBX
NVMe Power and Thermal Management Features
Controller Phison PS5008-E8
Firmware E8FM11.4
NVMe
Version
Feature Status
1.0 Number of operational (active) power states 3
1.1 Number of non-operational (idle) power states 2
Autonomous Power State Transition (APST) Supported
1.2 Warning Temperature 81°C
Critical Temperature 85°C
1.3 Host Controlled Thermal Management Not Supported
 Non-Operational Power State Permissive Mode Not Supported

The E8 controller supports two idle power states, implemented as NVMe Power States 3 and 4 (power states 0, 1 and 2 are operational states with successively lower power limits). The firmware version E8FM11.4 used by the MyDigitalSSD SBX doesn't use power state 4 when the NVMe Autonomous Power State Transition (APST) feature is enabled, which is what most systems in the real world rely on for NVMe power management. The drives can be put into PS4 when APST is disabled and the OS manually manages the drive's power states. The relatively new power and thermal management features from the version 1.3 NVMe spec are not implemented.

MyDigitalSSD SBX
NVMe Power States
Controller Phison PS5008-E8
Firmware E8FM11.4
Power
State
Maximum
Power
Active/Idle Entry
Latency
Exit
Latency
PS 0 3 W Active - -
PS 1 2 W Active - -
PS 2 1 W Active - -
PS 3 0.1 W Idle 1 ms 1 ms
PS 4* 0.005 W Idle 400 ms 90 ms

(PS4 not usable by APST with current firmware)

Note that the above tables reflect only the information provided by the drive to the OS. The power and transition latency numbers are often very conservative estimates, but they are what the OS uses to determine which idle states to use and how long to wait before dropping to a deeper idle state.

Idle Power Measurement

SATA SSDs are tested with SATA link power management disabled to measure their active idle power draw, and with it enabled for the deeper idle power consumption score and the idle wake-up latency test. Our testbed, like any ordinary desktop system, cannot trigger the deepest DevSleep idle state.

Idle power management for NVMe SSDs is far more complicated than for SATA SSDs. NVMe SSDs can support several different idle power states, and through the Autonomous Power State Transition (APST) feature the operating system can set a drive's policy for when to drop down to a lower power state. There is typically a tradeoff in that lower-power states take longer to enter and wake up from, so the choice about what power states to use may differ for desktop and notebooks.

We report two idle power measurements. Active idle is representative of a typical desktop, where none of the advanced PCIe link or NVMe power saving features are enabled and the drive is immediately ready to process new commands. The idle power consumption metric is measured with PCIe Active State Power Management L1.2 state enabled and NVMe APST enabled if supported.

Active Idle Power Consumption (No LPM)Idle Power Consumption

The active idle power draw of the MyDigitalSSD SBX is pretty good for a NVMe drive and on par with many mainstream SATA SSDs. The situation when PCIe ASPM and NVMe APST are enabled isn't great, but at least the Phison controllers are no longer freaking out and paradoxically increasing power consumption. The power savings we observed seem to come mostly from PCIe ASPM, and it makes little difference whether the drive is asked to use APST or not. The drive doesn't stay within its own declared limit of 100mW for PS3.

Also important to the E8 platform's overall idle power consumption is how background processing is handled. When idle, the drive will periodically wake up to perform background processing such as garbage collection. For the first few minutes after the drive is powered on, the interval between those wake-ups is 0.8 seconds, then the drive slows to waking up once every 5 seconds. These wake-ups continue whether or not the drive has background garbage collection or SLC cache flushing to do. Since the drive's power spikes to just over 1W during these active periods and they last for about 200ms each, this increases the overall idle power draw by more than 20%.

Idle Wake-Up Latency

The idle wake-up latency test shows no significant difference in performance for the SBX between having all the power management features enabled or disabled—sometimes the measurements are faster with power management enabled, but still within the margin of error, so we're showing the latency as just zero in those cases. It is clear that no deep power saving measures are being taken within the SSD in this configuration, so there is nothing that would impose a significant wake-up delay.

Mixed Read/Write Performance Conclusion
Comments Locked

46 Comments

View All Comments

  • Samus - Tuesday, May 1, 2018 - link

    Ditto. I think it'd crazy to use a 2.5" over an M2 if you have the M2 slot available, especially since M2 SATA drives are often cheaper than 2.5" drives (because they are less expensive to manufacture, and the OEM market is larger.)

    And as far as SATA M2 drives, if you have an M2 slot that supports NVMe, it's hard to justify not using an NVMe SSD when the cost difference is less than 20%...I picked up the WD Black 512GB NVMe drive last week for $150. A decent 540GB class SATA M2 SSD is at least $120.
  • Death666Angel - Tuesday, May 1, 2018 - link

    If every dollar counts and the performance increase is small or won't be used, it's pretty easy to justify getting a SATA M.2 drive instead of an NVME one.
  • Death666Angel - Tuesday, May 1, 2018 - link

    Especially since small capacities are likely to be very close in speed, when comparing NMVE and SATA M.2.
  • Byte - Tuesday, May 1, 2018 - link

    As someone who does a lot of testing/tweaking, i love the easy formfactor, but hate having to screw and unscrew. We really need a tooless update.
  • MajGenRelativity - Thursday, May 3, 2018 - link

    Honestly, I find it about the same amount of time/difficulty to (un)screw an M.2 drive as it is to work with even a toolless 2.5" drive. Unless the M.2 drive is under the GPU, in which case that really annoys me
  • leexgx - Tuesday, May 1, 2018 - link

    Still £/$30 more for a customer who is not going to benefit from the nvme ssd (and less money for you)

    I hardly notice the difference between the sata and nvme ssd my self, main difference is them above 1GB/s speeds but day to day usage I don't really notice much the difference between them unless I am looking for the difference (as long as it's Not a HDD even a slow ssd is many times faster then a hdd)

    Do Samsung 850 evo have am issue if they have been left on for to long (like 30 days) as my 850 evo just crap it self out smart fail at Bios and can't read it (only done basic not hirions boot CD yet)
  • MajGenRelativity - Thursday, May 3, 2018 - link

    Actually, I'd technically make a little more money if I sold them an NVMe SSD (my labor cost scales with price of parts), but they wouldn't benefit from it, so I generally don't recommend them. 850 Evo's don't normally have that issue.
  • peevee - Tuesday, May 1, 2018 - link

    AT, how about a couple of user-reproducible, real life tests? Compilation of a large software package. Unzipping a large archive. Recoding video. Just to demonstrate the scale of improvement the buyers could actually SEE.
  • SanX - Wednesday, May 2, 2018 - link

    Two reasons come instantly. Because only salespeople left in IT. No one even discuss calling lawyers for such confusing people blatant claims like 1600MB/second read speed this product has. And because Windows for example will load something like in 17.6 seconds instead of 17.9 with this drive vs SSD.

    Funny also is that 2-3 times slower drive which does not deliver at all is just 25-30% cheaper then the leaders.
  • peevee - Friday, May 4, 2018 - link

    This site is often for people for assemble their own PCs and/or choose what to buy for their companies. I'd think a few reproducible, real life tests vs proprietary and compressed tests would show the value of improvements.
    Maybe it is what AT really is afraid of, because tests show the improvements which do not exist in real life?

Log in

Don't have an account? Sign up now