Interesting observations you've got there. Seeing that you are running on a laptop, I suspect there are other power saving modes which exist within the laptop that you may not have disabled. Have you checked Link Power Management (LPM) features such as HIPM or DIPM? My initial hypothesis is that the drive is entering some sort of power saving mode where the internal SSD controller stops counting clock cycles. Do you simply run the system idle? Perhaps running a continuous workload to ensure the drive stays "awake" would be the best method to test this. (Albeit at the cost of the longevity of your SSD)
or maybe you did find something wrong with the power on counter
Let's see what Intel has got to say about it.
I didn't think of this at first, but even if the drive goes to power saving mode it isn't completely powered off and should be aware of current time. Also, I didn't observe this problem on SSD 510 (though it's got different error reported here), it reports power on hours correctly.
Thank you for posting.
DIPM is indeed related to Power-On Hours Count:
“The raw value reports the cumulative number of power-on hours over the life of the device.
The On/Off status of the Device Initiated Power Management (DIPM) feature will affect the number of hours reported.
- If DIPM is turned on, the recorded value for power-on hours does not include the time that the device is in a "slumber" state.
- If DIPM is turned off, the recorded value for power-on hours should match the clock time, as all three device states are counted: active, idle and slumber.”
Thank you for pointing this out! I was aware of DIPS (every now and then I see on AnandTech people complaining that they don't enable DIPS when testing SSDs on power consumption), but didn't realize it's enabled by default. With DIPS disabled, 'power on hours' keep adding up normally indeed.
Enabled by default on mobile platform, that is (confirmed on Montevina and Calpella platforms). DIPS is disabled by default on desktop machine (at least on Denlow platform). I also noticed when SSD connected via SATA-to-USB bridge, DIPM wouldn't trigger even if it's enabled by operating system (possible feature/bug on my external drive IC).
I don't have my old machine with X25-M available anymore so I've confirmed your info on Denlow platform and Intel 320 model.
How to check DIPS status
Run Intel SSD Toolbox, Drive Summary, Drive Details:
word 79 (Serial ATA Features Enabled), bit 03 (Device Initiated Power Management Enabled): 0 (disabled) or 1 (enabled)
How to control DIPS mode in Windows
It is possible to control DIMP (and HIMP) in power control section with following two entries in registry:
New item would appear in 'hard disk' section which allows to control DIMP:
AHCI Link Power Management with options: Active, HIMP, HIMP + DIMP
Thank you for the feedback and detailed step by step guide.
What is Intel's advice regarding enabling/disabling DIPM on a laptop computer?
I can't speak for Intel, but from Intel's document 325152 State Drive 320 Series, Product Specification, March 2011 the drive's idle power consumption Idle: 100 mW (TYP) is mentioned explicitly with DIMP enabled.
In Windows it seems (see my comment above) that by default mobile platforms has got DIMP enabled as opposed to disabled on desktop platforms and that this is levered by Microsoft, not by Intel (certainly, the drive must be DIMP capable though).