speed of VM vs. speed of database
Locked
speed of VM vs. speed of database
 
|
i am currently developing an app just for fun..
during my day job, i am a developer using a web framework that directly connects to a database (mysql).. one of the things we always do is craft our queries to use use the database only for all queries and selections. when you use the database for anything other than trivial queries, the difference is the speed of the query is totally evident.. i am just wondering how the speed of the smalltalk VM compares to the speed of pulling results from a database and shoving them into an array.. thanks! -- ---- peace, sergio photographer, journalist, visionary http://www.CodingForHire.com http://www.coffee-black.com http://www.painlessfrugality.com http://www.twitter.com/sergio_101 http://www.facebook.com/sergio101 _______________________________________________ seaside mailing list [hidden email] http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside |
Re: speed of VM vs. speed of database
|
|
hey sergio, if objects and tables with relations where truly comparable, objet-relational-impedance-mismatch [1] would not exist. In the other hand if you use a BTree of objects indexed in the way you need, your speed will have no match in the RDBMS world. Proper use of Gemstone, omnibase or even a modest index persisted in the image (a facade to a BTree or something like that) will satisfy quite high expectations. Have in mind that, for performance, the only metric that really counts is the user experience. Maximize anything else and you'll be heading to a non human centered artifact, faster. BTW, this could be off topic but in the economy there is an anti-pattern familiar to this when you see companies maximizing profit instead of wealth creation. for more I highly recommend this video: Re-thinking IT by John Seddon o/ On Dec 3, 2011, at 3:56 PM, sergio_101 wrote:
_______________________________________________ seaside mailing list [hidden email] http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside |
Re: speed of VM vs. speed of database
|
|
In reply to this post by sergio_101-2
2011/12/3 sergio_101 <[hidden email]>:
> i am currently developing an app just for fun.. > > during my day job, i am a developer using a web framework that > directly connects to a database (mysql).. > > one of the things we always do is craft our queries to use use the > database only for all queries and selections. when you use the > database for anything other than trivial queries, the difference is > the speed of the query is totally evident.. I have had good to very good experiences with RDBMs given they were properly set up and I wasn't doing "stupid" things. > i am just wondering how the speed of the smalltalk VM compares to the > speed of pulling results from a database and shoving them into an > array.. What do you mean with "speed of the WM"? Just the rendering speed? Cheers Philippe _______________________________________________ seaside mailing list [hidden email] http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside |
Re: speed of VM vs. speed of database
|
|
>
> What do you mean with "speed of the WM"? Just the rendering speed? > by speed, i mean the time from asking for a queried subset of results, and getting the results.. thanks! -- ---- peace, sergio photographer, journalist, visionary http://www.CodingForHire.com http://www.coffee-black.com http://www.painlessfrugality.com http://www.twitter.com/sergio_101 http://www.facebook.com/sergio101 _______________________________________________ seaside mailing list [hidden email] http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside |
Re: speed of VM vs. speed of database
|
|
In reply to this post by sebastianconcept@gmail.co
> Proper use of Gemstone, omnibase or even a modest index persisted in the
> image (a facade to a BTree or something like that) will satisfy quite high > expectations. > hey, sebastian.. just out of curiosity, which smalltalk are you using? cincom looks fun, but i don't want to get involved in something i can't directly get the price on.. thanks! -- ---- peace, sergio photographer, journalist, visionary http://www.CodingForHire.com http://www.coffee-black.com http://www.painlessfrugality.com http://www.twitter.com/sergio_101 http://www.facebook.com/sergio101 _______________________________________________ seaside mailing list [hidden email] http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside |
Re: speed of VM vs. speed of database
|
|
In reply to this post by sergio_101-2
2011/12/4 sergio_101 <[hidden email]>:
>> >> What do you mean with "speed of the WM"? Just the rendering speed? >> > > by speed, i mean the time from asking for a queried subset of results, > and getting the results.. Compared to what? Cheers Philippe _______________________________________________ seaside mailing list [hidden email] http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside |
Re: speed of VM vs. speed of database
|
|
In reply to this post by sergio_101-2
yeah, me too, Cincom has done a fantastic VM but is too enterprisey for me
I'm using Pharo o/ On Dec 3, 2011, at 11:16 PM, sergio_101 wrote:
_______________________________________________ seaside mailing list [hidden email] http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside |
Re: speed of VM vs. speed of database
|
|
In reply to this post by sergio_101-2
Work done straight with the VM will be at least 100X faster compared
to the "same" work done with a database. To "read" a collection in the VM (e.g., #do:), just a few bytecodes are executed. But to do that with objects in a database means it has to convert the objects in the database into Smalltalk objects first. This conversion requires the execution of thousands of additional bytecodes. But databases are still useful for the features they provide that images don't, such has having large models that exceed the size of memory, or letting multiple users access and update a model at the same time, or high-availability and consistency, etc. On Sat, Dec 3, 2011 at 11:56 AM, sergio_101 <[hidden email]> wrote: > i am currently developing an app just for fun.. > > during my day job, i am a developer using a web framework that > directly connects to a database (mysql).. > > one of the things we always do is craft our queries to use use the > database only for all queries and selections. when you use the > database for anything other than trivial queries, the difference is > the speed of the query is totally evident.. > > > i am just wondering how the speed of the smalltalk VM compares to the > speed of pulling results from a database and shoving them into an > array.. > > thanks! > > -- > ---- > peace, > sergio > photographer, journalist, visionary > > http://www.CodingForHire.com > http://www.coffee-black.com > http://www.painlessfrugality.com > http://www.twitter.com/sergio_101 > http://www.facebook.com/sergio101 > _______________________________________________ > seaside mailing list > [hidden email] > http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside seaside mailing list [hidden email] http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside |
Re: speed of VM vs. speed of database
|
|
The whole subject is like asking if a car goes faster with a trailer attached. Of course not, but it is often needed anyway.
-Boris On 2011-12-04, at 12:29, "Chris Muller" <[hidden email]> wrote: > Work done straight with the VM will be at least 100X faster compared > to the "same" work done with a database. To "read" a collection in > the VM (e.g., #do:), just a few bytecodes are executed. But to do > that with objects in a database means it has to convert the objects in > the database into Smalltalk objects first. This conversion requires > the execution of thousands of additional bytecodes. > > But databases are still useful for the features they provide that > images don't, such has having large models that exceed the size of > memory, or letting multiple users access and update a model at the > same time, or high-availability and consistency, etc. > > On Sat, Dec 3, 2011 at 11:56 AM, sergio_101 <[hidden email]> wrote: >> i am currently developing an app just for fun.. >> >> during my day job, i am a developer using a web framework that >> directly connects to a database (mysql).. >> >> one of the things we always do is craft our queries to use use the >> database only for all queries and selections. when you use the >> database for anything other than trivial queries, the difference is >> the speed of the query is totally evident.. >> >> >> i am just wondering how the speed of the smalltalk VM compares to the >> speed of pulling results from a database and shoving them into an >> array.. >> >> thanks! >> >> -- >> ---- >> peace, >> sergio >> photographer, journalist, visionary >> >> http://www.CodingForHire.com >> http://www.coffee-black.com >> http://www.painlessfrugality.com >> http://www.twitter.com/sergio_101 >> http://www.facebook.com/sergio101 >> _______________________________________________ >> seaside mailing list >> [hidden email] >> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside > _______________________________________________ > seaside mailing list > [hidden email] > http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside seaside mailing list [hidden email] http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside |
Re: speed of VM vs. speed of database
|
|
In reply to this post by sebastianconcept@gmail.co
> Have in mind that, for performance, the only metric that really counts is
> the user experience. > > Maximize anything else and you'll be heading to a non human centered > artifact, faster. Utility matters more than raw speed. Applications that provide high functional-density even at the cost of some raw "speed" before returning control to the user provide a better user-experience than hyperspace-speed apps. Take a simple example. Say I want to know why an object is not being GC'd. Compare the "speed" of explore pointers vs. chase pointers. Explore pointers "performs" better in terms of getting back some kind of response, but chase pointers gets me to the REAL answer I'm looking for much faster. The machine does all the work while I was taking a sip of hot coffee, instead of me clicking around while my coffee gets cold. Now, I know some folks use computers as if they've already had two cups of coffee. They click around very fast but with lots of misclicks, and lurching back and forth through a high-speed system but which provides a very low functional-density. The user often ends up talking to themself (perhaps to keep their train of thought, "oops, I didn't want to click that... this is what I want..."), and so train of thought about what they're doing is somewhat broken. When using a system, ask yourself, how many gestures are solely for manipulating the UI (e.g., maximizing, resizing) and how many are for actually accessing/manipulating the domain? For each of the former, speed doesn't count -- in fact it counts negatively because its a distraction from the users domain. So, Interacting with software can be as "slow" as interacting with another human and still be a great experience. e.g., 800-1000 ms response is "fast enough" as long as the user feels the software is _doing_ useful things. - Chris _______________________________________________ seaside mailing list [hidden email] http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside |
Re: speed of VM vs. speed of database
|
|
thanks all for your input.. i think my concerns were unfounded.. i
just didn't want to spend time writing a bunch of test cases to figure out the answer.. i am going to keep on my current trajectory.. thanks again! -- ---- peace, sergio photographer, journalist, visionary http://www.CodingForHire.com http://www.coffee-black.com http://www.painlessfrugality.com http://www.twitter.com/sergio_101 http://www.facebook.com/sergio101 _______________________________________________ seaside mailing list [hidden email] http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside |
«
Return to Seaside General
|
1 view|%1 views
| Free forum by Nabble | Edit this page |