How I test for productivity (eBoostr 4.0 Beta 543)

This is how I might test an SSD (if I ever get my hands on one).

System:
VAIO P8400 8GB RAM 500GB Scorpio Blue HD Win 7 x64 Ultimate.
Installed programs and utilities, including Win 7 is about 65GB total.

Downloaded the newest beta version of eBoostr which offers native 64bit support - had to test if a software cache (similar to what Intel's Braidwood will offer in 2010) will have any effect on productivity. I had uninstalled eBoostr for about a week and my gut feeling was that 'productivity' was affected negligibly. This test was to put some numbers to those 'feelings' no matter how smooth the computer ran with eBoostr installed previously.

eBoostr now offers an 'autoconfiguration' option. I tried it and it suggested a 2.5GB RAM cache and also a 4GB cache on my USB based Lexar ExpressCard SSD (16GB size).

I used the computer normally for a couple of days and the RAM cache was finally 100% filled and the cache on the Lexar was 65% filled of 15,268MB (I increased it to its maximum size - didn't like the 4GB default).

Ran some real-world tests that mirror how I actually use my notebook (see below).

I then uninstalled eBoostr and ran the same tests.

Installed eBoostr once again but this time did not use any RAM for the cache. Also, I got a little impatient to post this (also, I'll be too busy in the next few days) and only let the Lexar fill to 48% of 15,268MB.

That's 2.5GB less cache used in the USB based device and also 2.5GB less cache of the first test run with the RAM used as cache also (5GB total, less cache for the 48% column vs. the 65% column).

So these are my tests: time the three different configurations above and see how fast the system: shuts down, shows the desktop, becomes usable (responsive to user input), opens and closes 67 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 Safari 4 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').

Close CS4 and time to start it again (warm start): 4 seconds.

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


............................No eBstr............eBstr+RAMc(65%)......eBstr(48%)

Shut down...............15sec....................13sec..................10sec

Desktop...................55sec....................68sec..................45sec

Usable...................+85sec..................+52sec................+75sec

Total Start Time.......140sec.................120sec.................120sec

Open/Close 67 Apps..18min.....................14min..................12min

Convert 373 images...27min.....................24min..................25min


Okay, what's the verdict?

eBoostr increased productivity by as little as 7% (converting RAW images) and as much as 35% (opening/closing programs). To put this in perspective that 7% is 2 minutes savings and the 35% is a 6 minutes savings.

What eBoostr offers is how responsive the O/S feels - even when it's being pushed hard by a background process. The productivity gains are a bonus, but are not it's main attraction (any specific computer system can only produce so much - no matter how much help you give it).

Another byproduct of having a responsive O/S is that WMP audio stuttered on both eBoostr Lexar + RAM cache and also with no eBoostr installed. When just the Lexar was used as cache for eBoostr, WMP didn't stutter the audio at all. (This is when doing the 'productivity' test, of course).

This is a $40 product and the benefits seemingly far outweigh the cost (hope the shipping final version is this good and I hope Intel's Braidwood does it even better).

The 'traditional' knowledge states that to increase productivity you must upgrade your CPU, your RAM and/or your Video card - as these are the components that most directly affect 'output'.

Lately though, all I read on the web is that SSD's offer the most effective way to upgrade the performance of your computer. But actually using SSD equipped computers have me thinking otherwise. The benefits of SSD's are more in line with what I've found eBoostr to offer, but at a much higher cost (not only in $$, but also in being a 'beta' tester for the SSD industry with your data).

My question to the members here is; do SSD's offer productivity benefits that surpass what a software program does?

I know and understand about power savings, less heat, not prone to vibrating, etc. etc. - but I am curious if SSD's offer productivity gains over mechanical HD's and/or the eBoostr results shown above.

Note that the 4 second CS4 restart ('warm' start) is indicating the maximum my system can load that program from memory (RAM), so eBoostr to load it 'cold' at 5 seconds is, to me, pretty amazing considering that it takes 16 seconds otherwise.

Not even an SSD can beat the 4 second PS CS4 startup time (in my system) - but, will an SSD beat the 2 minute improvement in my productivity test? That's the question.

 

Eboostr's advantages are shown even more significant when used on netbooks with slower processor.

 

Yes, I've tested eBoostr on a nettop computer and it made it much more bearable! (Win XP with 1GB RAM).

But, try to convince the owner of a nettop ($400 total purchase price, or less?) that the $40 program is making his computer finally usable (well, usable to me, he thought it was fine ).

 

Some additional information that may be of interest (before I throw away my notes on this little test).

Shutdown times:
Without eBoostr a 'dirty' shutdown (meaning the computer was used heavily just before turning it off) took about 22 seconds, with eBoostr and RAM + Lexar caches, a similar 'dirty' shutdown took 15 seconds. Without the RAM cache, it took only 16 seconds. Both results are also impressive percentage-wise; over 37 percent longer minimum without eBoostr installed.

CPU usage:
With eBoostr using RAM and the Lexar storage as cache, the CPU usage would fluctuate from 32% to 100% 'wildly' when doing the conversion of the RAW files (and the other programs running).

With no eBoostr installed, doing the same conversion would have the CPU usage pegged from 72% to 100% with it mostly staying around 98%.

With eBoostr using the Lexar, the CPU usage would stay consistently between 88% - 92%.

These results are consistent with the responsiveness of the system under each scenario - the last one was easily the most responsive - even when opening up a program like Premiere Pro (not during the test, of course) and seeing that the eBoostr cache not only helped in launching a heavy duty program like CS4 PP in about 20 sec's, but it also did not measurably impact the conversion of the RAW files, nor did it make WMP stutter on the WMP lossless playback either. Impressive!

Isn't this what an SSD offers? Albeit with slightly more speed and even more additional benefits?

 

Okay, I may have answered my own question (will update and post link to new thread soon).

Edit: Here's my SSD test thread:
http://forum.notebookreview.com/showthread.php?t=436882

 

no. more speed, absolute silence, the speed always, no matter what you access across the disk (which is what i care about much, and why i hate caches), cooler, absolut undestroyable (well, as much as the rest of the laptop)

oh, and, they don't use cpu and ram for their work, which is a batterylifesaving and cpu resource freeing feature, espencially for atoms. even while it might still idle around 0 - 1%, it will force the cpu to not go to idle mode often than an ordinary ssd.

so while it might work fine, it's only a funky patch. the real fix works much better, and is thus, what i like.

 

davepermen,

Thanks for the reply.

I agree with almost everything you say except the 'speed always' part. I have not seen that with the big three I have used.

As a matter of fact, I have seen the opposite. So far, a mechanical HD is much more consistent in my usage scenarios. Glad they work for you.

(Big three; Intel, Indilinx & Samsung).

 

no clue how you're doing it wrong, then. but honestly, i don't care.

 

davepermen::::

You are being RUDE, plz take it easy on the guy, he is just trying to discover sth. new, and he is just SHARING his feedback, not opening a law suite onto you to follow his way.

Let's all be Kind to each other, IF we have nothing to offer.

regards to you, and to him.

Will

 

Willy330Ci,

I thank you for sticking up for me. But me and davepermen are fine (well, I'm fine, anyway).

I just wish he would just show how to get better results than me, or even redefine the test conditions so that I may learn something in one or two posts, instead of over one or two weeks?

Cheers!

 

we actually have a good time together. no clue why, but we get along quite nicely. we want the same, and come from very different experiences. i'm fine, and i like him very much.

 

problem is, i don't know how to help you.. i just see you having much worse results than i have, so your setup failed to deliver. why? i can't say. i'd love to be able to help better but i can't, really. i'm sorry

 

eboostr + SSD = much better responsiveness, i put eboostr on all of my laptops/netbooks.

 

I put eboostr on my Sony with XP and am very happy with the increase in programs loading. I notice the quicker shutdown more than the startup. I do not have the time for actual scientific testing like some of you. I currently have version 5.47 and it runs without a hitch.

 

That kids, is commonly referred to as a "Double-Whammy".

 

Nice to see there are improvements by using eboostr.
I've been using eboostr 4 beta to make use of my ~500MB hidden memory in XP 32 bit
without knowing if there are real improvements or not LOL..