SSD's; The Myth Exposed.

Back in this thread:
http://forum.notebookreview.com/showthread.php?t=435052

I asked if an SSD would improve my productivity in my notebook vs. what eBoostr (beta 4 build 543) offered. Nobody cared to answer that particular question, even though quite a few people seemed to read my post.

Having just purchased a Patriot Torqx 64GB SSD, I am now in the position to answer it myself.

System:
VAIO P8400 8GB RAM Patriot Torqx 64GB SSD Win 7 x64. vs.
Identical system hardware except for 500GB Scorpio Blue HD.

First thing I did was install firmware 1819 on this Indilinx SSD.

I then installed Win 7, the hardware drivers, did the Windows updates and installed Security Essentials. Installed maybe 1/4 of my programs (64GB = 59.5GB actual space - less than my Music collection!) and let the system settle for a few hours. I disabled system restore (useless, any time I've ever used it), disabled scheduled defrag and set the pagefile to 512MB. That's all for 'tweaks'.

I copied 7GB of RAW image files and also 350MB (1 album) of Windows Media Lossless uncompressed music too. This left 13GB of free space left, well below the 20% 'suggested' free space for HD's (could have copied 2.1GB more files over, but left some room for temp files).

I ran a subset of the tests I did in the thread linked above (because I didn't have 67 of my programs installed) and the results are what I guessed at last Monday.

You can see the numbers below, but here is my conclusion: eBoostr does offer comparable performance to an SSD. Considering the price difference (10x more for the SSD) eBoostr's Value is phenomenal (for me and my specific system setup).

Now for some disclaimers:
Is this a scientific test that compares apples to apples? Not exactly. But good enough.

Was the Scorpio Blue installation exactly the same as the Torqx SSD's? No.

Was the mechanical HD at a disadvantage because it had more 'garbage' on it? Yes.

Was this as 'real-world' as I could make it considering the time and capacity (of the SSD) limitations? Yes.

The Scorpio Blue installation is over 3 or 4 months old now, includes many more programs installed (including much more startup programs) and also includes VMWare Workstation and XP Mode virtual machines installed. In addition, the hardware screen calibrator program is installed and runs on every boot too - this test definitely was not slanted in any way towards the mechanical HD's favour - no, the SSD had every advantage possible except the advantage of more free space.

The SSD felt fast - very fast, but it was also inconsistent in it's performance (extracting a 180KB RAR file took almost a minute on the Torqx - yes, that's KB, not MB and this aspect of its 'performance' is what makes the Torqx 'garbage').

Remember, that these tests were performed on a Trim enabled O/S with trim enabled firmware.

Okay, the tests and the numbers:

So these are my tests: time the SSD and see how fast the system: shuts down, shows the desktop, becomes usable (responsive to user input), opens and closes 48 of my most used programs, and finally the all important productivity test:

1) Open WMP11 and play an album encoded with Windows Media Lossless compression.
2) Start Outlook
3) Start IE8 and navigate to notebookreview.com
4) Start LR 2.5 and convert 373 12MP images to 1200x1200 pixel jpgs.
5) Wait until LR pops open an explorer window indicating its finished.

The time on the productivity test only includes how long it took to convert all 373 raw files.

Additional tests:
With No eBoostr installed: Adobe PS CS4 takes 16 seconds to start ('cold') (Scorpio Blue HD).

On SSD it takes 7 seconds to start 'cold'.

Close CS4 and time to start it again (warm start): 4 seconds (both SSD and Scorpio drives).

With eBoostr (no RAM cache) CS4 takes 5 seconds to start. Again, to restart its 4 seconds.


..............................SSD..................ScorpioBlue+eBstr(57%)

Shut down...............15sec...........................10sec

Desktop...................42sec...........................45sec

Usable...................+18sec.........................+75sec

Total Start Time....... 70sec..........................120sec

Open/Close 48 Apps.....8min............................9min

Convert 373 images....26min............................25min

Kinda interesting huh?

I also ran the Patriot.exe manual Trim tool - did nothing for the numbers above. I ran PerfectDisk 10 Defrag Free Space. Made the drive slower. I Ran the wiper again - no difference. (This was supposed to be the 'Tony Trim' method; Indilinx, you lie ).

Okay, let me state this straight out; past 50% filled, the Torqx started slowing down. With Win 7 freshly installed (and about 40GB free space) the computer flew. But a 'clean' computer does not do any work; the software does and putting the bare minimum on it (and with only 7.5GB of personal data) makes this size drive unusable. Why?

Well, lets list the reasons:
1) Mech HD don't ever, ever choke on small files (especially with eBoostr).
2) Waiting for a minute to extract a 180KB file is ridiculous for a $380 drive.
3) Scorpio+eBoostr shuts down faster (with a much larger install)! Hmmm....
4) The times 'saved' by using an SSD or eBoostr is only for the first launch of the program, afterwards the speed is the same no matter what HD you are using (thanks to SuperFetch & 8GB of RAM).
5) Converting the RAW images was actually faster using the Scorpio/eBoostr combo! I'm not only surprised at this, I'm truthfully shocked.
6) On a much smaller installation, the SSD is faster by a whole 3 seconds than the Scorpio/eBoostr combo. It is much faster to be usable though (50 seconds faster), but how many times do you reboot in one day?
7) Granted, the SSD opened and closed the same programs one minute faster than the Scorpio, but keep in mind that that one minute is only 5 programs to open/close out of 48 and also remember the much larger install on the Scorpio so although its a technical 'win' I see it as getting about 12% faster for almost 10 times the cost.

I also tried the Torqx with eBoostr and it did make a difference - the biggest being that the audio wouldn't skip during the RAW conversion process.

So, the bottom line is this; if I reboot only once a day I would save 50 seconds with the SSD. Additionally, the first time I open my (48) programs, I would save an additional minute... however, if I convert 373 images during that time (while listening to my music! ), I would loose that minute I just saved. So... $380 for 50 seconds saved? Ya, here's my cash - not!

@hollis_f, I also tried the same conversion test above but to my USB connected Scorpio (rebooted in between, of course) the time to convert was exactly the same; 26min (RAW files imported from and final jpegs outputted to, the USB connected mechanical drive in LR2.5).

@davepermen, so - I've tried one of the latest and greatest and... it's returned already - not even a tear from me... bad, bad drive...

@everyone,
So, for me, Indilinx based SSD's are relegated to the scrap heap, along with Samsung SSD's (they both 'stutter' - even though everyone says they don't).

I still want to try an Intel G2 - but this generation of SSD's looks bleak (two out of three are bad (go Meatloaf! ).

Intel! Where is Braidwood?

 

Do I not see the SSD's attractive side? Yes I do.

No noise, no heat, no vibration.

In my particular case above, I saw no increase in battery life either though.

If I had an underpowered netbook, with barely enough RAM (less than 2GB is silly in 2009) I can see how an SSD can seem to make that type of computer 'better'. It may seem faster, but in the end, you still will not do any real work on it (meaning CPU intensive, not meaning real as in 'valuable' or not).

But if we are trading our hard earned dollars thinking that an SSD can make a productivity increase on our same old hardware - we are sorely misguided.

At least, if we don't explore all our available options ourselves (instead of believing the hype).

 

Which strongly suggests that the jpeg production process is processor/ram limited. The test I did was to render 1:1 previews in LR. That's where I saw a big difference.

 

you just want to be cool (just reading your title i knew several things.. a) it's from you b) you foudn something more to bash on them c) you failed at it)

yes. ssds don't improve your cpu, or your ram.
no, nobody cares about the patriot 64gb.
yes, you should really finally just test out the intel.
yes, you should learn what it's about, when it's good, when it's bad, and why i rocks exactly then.

you can't beat that. at least not for the sound

 

As you know we are still in the early stages of SSD. Things will get better.

 

actually, no, things ARE better (And are since about a year). but he still failed to pair a good system with a good ssd. so far he played with known bad ssds, or known cheap ssds, or on very old systems that just suck by default.

he just doesn't get it. but i said it before, i don't care about it anymore. all the time and money spent trying all those things could have been invested in a single intel 160gb gen2, and all would have been perfect. but no.. he likes to try and play, and bash and cry, and hopes to get famous in the web trough that.

good luck with that.

 

hmmm, I sense some resentment towards the OP.

Things have definitely gotten better in the last year. No one wants to go back to those jmicron controllers but I do think it's still early and in the next year we should see even more improvements.

What's the definition of a good system by the way? Looks like he is testing with a Montevina based Vaio.

 

anything not vaio

well, to be fair: he tested a variety of combinations, but none with a good system + a good ssd together. and always got mediocre results. this is to be expected.

and he, time and time again, tries to push his eboostr, failing to understand how this is just a bandaid fix instead of the real fix. and he directly tries to show me how right he is. while i have to laught about it, sort of, sitting in front of several system proving him wrong each day.

 

Thanks hollis_f,

I too think my 'test' may not show an SSD to its full potential, but let me point out what it did do:

Read 7GB of RAW image data. Spit out 400MB of jpg files. I think that an SSD should have done something, no?

Even if I'm wrong here, I'm not wrong in that this is how I would use my notebook anyway.

Oh yes, reading and outputting the same raw data (but on a USB attached Scorpio) yielded the same time - now, you may say this proves it's slightly CPU limited, but how do you explain the import/export to the USB taking the same time?

My explanation is that the SSD is not fast. At least, it's as fast in this case as a USB connected mechanical HD.

And there is the data point to consider that with that very same Scorpio Blue and eBoostr, the conversion finished a minute faster.

Not doubting you saw your difference, but I think my test is 'fair' too.

 

davepermen,

lets just say that I thirst for real knowledge about these SSD's - what you're suggesting is otherwise and I will not be drawn in.

I agree that I need to test the Intel, but I simply can't get my hands on one yet.

I thought that since Indilinx based SSD's are the next best thing and the opportunity presented itself, I would try them.

I'm glad I did, because I know not to waste anymore time with Samsung or Indilinx based drives based on my specific criteria of 'better', based on my actual experience on them.

As to the video link, yes I've seen that - but opening programs (fast) does not translate to more productivity as I have shown.