BUYING A SSD - THE MANUFACTURERS BLUFF

BUYING A SSD - LESSON 1 – WHY WOULD I WANT ONE?

BUYING A SSD – LESSON 2 – PRICE, SIZING AND THE MANUFACTURER

BUYING A SSD – LESSON 3 – THE CONTROLLER

BUYING A SSD – LESSON 4

ATTENTION SSD BUYERS!!!!!  If you are considering buying a SSD, this will be the most important thing you will ever learn.  If you have just bought a SSD, you are going to want to sit down and have a coffee before reading this.  If you are a manufacturer... Please don't read this!!! I am about to expose the biggest fallacy of solid state drive sales and get the truth out once and for all. And no I am NOT discussing MTBF (Mean Time Before Failure) and the SSD Life Expectancy as plainly described on every SSD package. That’s for another day.

This little piece of information will be great for one or two manufacturers while damaging the sales of most others.  It is something that should have come out long ago.  And...you have to know that its all tied to the government somehow so if you don’t hear from me after this, you will know why. Send for help!

The Manufacturers Bluff!

Lets get right to the point shall we?  Take a little look at these performance scores and tell me what SSD  you would buy.

Ok a few of you got it but I am disappointed in most so, lets try again.  Look carefully at the left performance results below and then the right.  Which setup would you believe will result in faster visible performance for the typical user? 

 If you guessed the one on the right you are absolutely correct.  I know, I know…some of you are about to write me off as a lunatic and and find another article to help you out with your purchase.  I even cheated a bit putting the performance score of 3x120Gb Vertex SSDs in RAID0 against my own Intel 160Gb X25-m SSD.  I tell you what...  Give me just a few minutes and I will change your mind completely  to the way you look at the performance of a solid state drive.  Deal?


When considering our purchase of a SSD we look at things such as transfer speed to guide us as to which SSD performs best.  Unknowingly, we get roped into the massive large sequential reads and writes that show the highest transfer speed which, in the case of the top left SSD is 245MB/s.  The SSD must be faster than the one beside it which is 183MB/s right?  The answer is both yes and no.  A bit of an understanding of disk access percentages is necessary to be able to intelligently decide specifically what SSD is best for you.  I am very fortunate to have a great support network of people who understand ssd technology very well.  One such person posted this one day which got my attention:


Top 5 Most Frequent Drive Accesses by Type and Percentage
-8K Write (56.35%)
-8K Read (7.60%)
-1K Write (6.10%)
-16 Write (5.79%)
-64K Read (2.49%)
Top 5 account for: 78.33% of total drive access over test period
Largest access size in top 50: 256K Read (0.44% of total)

Using Microsofts Diskmon, he simply monitored his typical computer usage in doing things such as using the internet, running applications, playing music etc.  In short, he did his best to recreate the computer use of a typical user and then used the program to break down the percentage that specific disk access speeds were being utilized.  In the end, it confirms something we always thought but just didn't really understand.  Large sequential read and write access is utilized by the average user less than 1% of the time yet the most used method of access is smaller random write access as shown by the 8k write at over 50%.  Manufacturers showcase the disk access method that is actually used the least (0.44% total) in order to get you to buy their SSDs.


Since the small random write score is accessed at over 50%, lets examine that for a bit.  Take a quick peek at the top two benchmarks followed by the bottom two.  Do you see a pattern at all?  Look closely at the small random 4kb access speed results and how they differ.  The truth is that these are the bread and butter of a good ssd and we have only just started to realize it with the release of the Intel drive.  When you start your computer and/or any software applications, your system relies on several dynamic link libraries (DLLs) to get the OS and applications up and running properly.  DLLs are simply hundreds of smaller programs that are called upon by the main program when needed and do not reside in your RAM.  These DLLs are very small in size and are loaded through 4-8km random access which means that the faster they load, the faster the OS loads as well as its software.  It also means the faster these DLLs can be loaded while a program is in use, the quicker the program.


In other words, the 4-8 kb random write access is the single most crucial access that results in better visible ssd performance.  If you are considering buying an ssd and you want the ssd with the best visible upgrade from your present system, you simply find the ssd with the best transfer results at the 4-8 kb random access level. Having said that, there are situations such as video editing that one will see the benefit of these high sequential benchmarks.  For the typical user, however, they will probably never come across a reason that higher sequential disk access, as advertised, will be utilized by them whatsoever.


This is why the Intels fare so well, both of which are shown in the benchmarks on the right above.  The top most is a Intel X25v 40Gb SSD while the one below it is the benchmark of my Intel X25m 160Gb ssd.  They have hit the ssd arena like a thunderstorm and have become the baseline for just about every ssd comparison one can find.  Only recently has any other SSD even come close as has been seen with the Crucial C300 with the Marvell controller (left) and newer SSDs that have the Sandforce controller such as the OCZ Vertex 2 Pro (right).

Ok so...its out.  The best visible performance improvement one can get for typical ssd use is not the SSD with the highest sequential read benchmark as advertised, but rather, the little known 4-8 kb random access speeds that are heard so little of.  They are seldom advertised on the exterior packaging itself and might only be found in tests found anywhere on the internet.  You can bet the newer drives such as the C300 and others with the newest SandForce controller are going to demand price premium.  Drives like the Intel, on the other hand, have earned their place in home systems and the company has even lowered SSD prices as of late to retain the marketshare and reputation they have gained.


Here is a comparison chart which shows the 4kb random scores of several drives.  Notice that the Intels still have better ratings than even the SandForce drives with only the Crucial with the Marvell controller placing higher. Of particular interest is the placing of the Velociraptor which doesnt even show a performance score as it was such a low comparison.