[Discuss] SSD drives vs. Mechanical drives

[bogstad at pobox.com (Bill Bogstad)]

On Sat, May 3, 2014 at 11:49 AM, Richard Pieri <richard.pieri at gmail.com> wrote:
> Jerry Feldman wrote:
>> I have seen some competing posts on other forums
>> SSD drives are certainly lighter and faster than mechanical hard drives.
>
> Lighter, yes. Faster, not necessarily. Flash chips are faster for random
> reads, certainly, but they're slower for sustained writes. You need a
> lot of "spindles" to beat rotating media for sustained write speeds.
> That's why SSDs are RAID0 devices inside their shells.

And why does it matter that flash chips are slow?   The question is whether
SATA connected SSDs are slow.  The first 500Gbyte SSD that I looked at
(Samsung 840 EVO MZ-7TE500BW) claims a >500 Mbyte/sec sequential
read/write speed.   Lower capacity SSDs (fewer chips for internal RAID0)
are going to be slower, but you could divide that number
in half and still beat mosts disks.  (Admittedly I seem to have picked
Samsungs "turbowrite" based SSD which claims higher
then normal sequential IO rates.)   On the other hand it seems like all of
the SSDs in that capacity range on Newegg claim 400-500 Mbytes/sec
sequential write speeds are possible.

>> The comment on Windows Boston that has no figures to back up his claim
>> is that the SSD drives actually are less power efficient than mechanical
>> hard drives because of their additional reliance on the CPU. And
>> certainly SSD drives are more expensive.
>
> The guy making that claim is ignorant of how SATA works. It's DMA.
> There's no CPU involved with I/O regardless of the physical medium
> inside the shell. Therefore, CPU load has no bearing on power efficiency.
>
> In real-world operation, SSDs are about the same as HDDs when it comes
> to power consumption. No mechanical parts is helpful, and reads are
> cheap, but writes are expensive when it comes to power consumption. It
> all tends to balance out for typical, day to day operation but more
> specific applications like database journals that are all writes all the
> time may skew power consumption measurements.

Newegg claims same Samsung SSD has the an active power consumption of
0.24 Watts.   Here's some testing of a number of Samsung SSDs for
power consumption:

http://www.tomshardware.com/reviews/samsung-840-evo-review-1tb-ssd,3567-13.html

Power use never gets above 3.5 Watts during any of their testing.
The article also gives peak power for about 20 different devices.
The peak for those is 7.25.

Here is an older (2011) test for notebook hard drives:

http://www.tomshardware.com/reviews/advanced-format-1tb-hard-drive,3046-11.html

It would seem that peak power for SSDs can be higher then notebook
hard drives, but that
idle power is the opposite.   Which is going to give you better power
consumption is going to
depend on your usage patterns.   If this is for a notebook, I suspect
that the storage device will be idle most
of the time and an SSD will result in both lower overall power use as
well as much better IO performance.

Bill Bogstad