| Author |
Message |
< Erlyweb mailing list ~ Performance of ErlyWeb relative other frameworks |
| Guest |
 Posted: Mon Dec 10, 2007 5:30 am |
 |
|
|
Guest
|
Ok, I have long suspected that ErlyWeb is faster than Ruby on Rails,
even on a single server. But I wanted proof. Now I have proof; well,
at least clear indications: http://blog.davber.com/2007/12/10/web-server-performance-shoot-out-simple-pages/
Interesting to note is that ErlyWeb (just like the Haskell cousin,
HAppS) is FASTER with dynamic content than with static content (from
file...) 
From a non-performance aspect, it is interesting to note the
difference in meta programming between ErlyWeb - with its reflective
Smerl - and that of HAppS/Haskell with compilation-based code
generation; ok, HAppS is not using Haskell proper (i.e., 9 but
instead Template Haskell, which gives ALMOST as much flexibility as
reflection/introspection.
Ok, see you soon!
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "erlyweb" group.
To post to this group, send email to erlyweb@googlegroups.com
To unsubscribe from this group, send email to erlyweb-unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/erlyweb?hl=en
-~----------~----~----~----~------~----~------~--~---
Post recived from mailinglist |
|
|
| Back to top |
|
| Guest |
| Posted: Mon Dec 10, 2007 7:44 am |
|
|
|
Guest
|
You have funny timing -- I spent most of the weekend running real
benchmarks between two EC2 instances and I found that ErlyWeb
obliterates Rails in performance, with a peak response rate that's 47x
higher than Rails (it may be even higher, because it looks like
ErlyWeb saturated EC2's internal bandwidth).
http://yarivsblog.com/articles/2007/12/09/erlyweb-vs-ruby-on-rails-ec2-performance-showdown
Try to do hot code swapping in Haskell
Cheers,
Yariv
On Dec 9, 2007 9:30 PM, davber <davber@gmail.com> wrote:
>
> Ok, I have long suspected that ErlyWeb is faster than Ruby on Rails,
> even on a single server. But I wanted proof. Now I have proof; well,
> at least clear indications: http://blog.davber.com/2007/12/10/web-server-performance-shoot-out-simple-pages/
>
> Interesting to note is that ErlyWeb (just like the Haskell cousin,
> HAppS) is FASTER with dynamic content than with static content (from
> file...)
>
> From a non-performance aspect, it is interesting to note the
> difference in meta programming between ErlyWeb - with its reflective
> Smerl - and that of HAppS/Haskell with compilation-based code
> generation; ok, HAppS is not using Haskell proper (i.e., 9 but
> instead Template Haskell, which gives ALMOST as much flexibility as
> reflection/introspection.
>
> Ok, see you soon!
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "erlyweb" group.
To post to this group, send email to erlyweb@googlegroups.com
To unsubscribe from this group, send email to erlyweb-unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/erlyweb?hl=en
-~----------~----~----~----~------~----~------~--~---
Post recived from mailinglist |
|
|
| Back to top |
|
| Guest |
| Posted: Mon Dec 10, 2007 8:40 am |
|
|
|
Guest
|
Oops, turned out I set up Mongrail wrong and it looks like the
difference is only 6x (or lower)
Yariv
On Dec 9, 2007 11:43 PM, Yariv Sadan <yarivsadan@gmail.com> wrote:
> You have funny timing -- I spent most of the weekend running real
> benchmarks between two EC2 instances and I found that ErlyWeb
> obliterates Rails in performance, with a peak response rate that's 47x
> higher than Rails (it may be even higher, because it looks like
> ErlyWeb saturated EC2's internal bandwidth).
>
> http://yarivsblog.com/articles/2007/12/09/erlyweb-vs-ruby-on-rails-ec2-performance-showdown
>
> Try to do hot code swapping in Haskell
>
> Cheers,
> Yariv
>
>
>
> On Dec 9, 2007 9:30 PM, davber <davber@gmail.com> wrote:
> >
> > Ok, I have long suspected that ErlyWeb is faster than Ruby on Rails,
> > even on a single server. But I wanted proof. Now I have proof; well,
> > at least clear indications: http://blog.davber.com/2007/12/10/web-server-performance-shoot-out-simple-pages/
> >
> > Interesting to note is that ErlyWeb (just like the Haskell cousin,
> > HAppS) is FASTER with dynamic content than with static content (from
> > file...)
> >
> > From a non-performance aspect, it is interesting to note the
> > difference in meta programming between ErlyWeb - with its reflective
> > Smerl - and that of HAppS/Haskell with compilation-based code
> > generation; ok, HAppS is not using Haskell proper (i.e., 9 but
> > instead Template Haskell, which gives ALMOST as much flexibility as
> > reflection/introspection.
> >
> > Ok, see you soon!
> > > >
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "erlyweb" group.
To post to this group, send email to erlyweb@googlegroups.com
To unsubscribe from this group, send email to erlyweb-unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/erlyweb?hl=en
-~----------~----~----~----~------~----~------~--~---
Post recived from mailinglist |
|
|
| Back to top |
|
| Guest |
| Posted: Mon Dec 10, 2007 4:24 pm |
|
|
|
Guest
|
Yariv,
Yeah, that was weird. Well, two great minds often think alike (but so
do two feeble minds...)
I almost fell into that non-production trap myself, but remembered to
include "-e production" for Rails in the last minute.
Yariv, you know what? You and I should sit down and create a benchmark
test for the frameworks I provided, along with Tomcat and .NET, to
have a semi-official analysis cross-posted later. What do you say?
/Davd
On Dec 10, 2007, at 2:43 AM, Yariv Sadan wrote:
>
> You have funny timing -- I spent most of the weekend running real
> benchmarks between two EC2 instances and I found that ErlyWeb
> obliterates Rails in performance, with a peak response rate that's 47x
> higher than Rails (it may be even higher, because it looks like
> ErlyWeb saturated EC2's internal bandwidth).
>
> http://yarivsblog.com/articles/2007/12/09/erlyweb-vs-ruby-on-rails-ec2-performance-showdown
>
> Try to do hot code swapping in Haskell
>
> Cheers,
> Yariv
>
>
> On Dec 9, 2007 9:30 PM, davber <davber@gmail.com> wrote:
>>
>> Ok, I have long suspected that ErlyWeb is faster than Ruby on Rails,
>> even on a single server. But I wanted proof. Now I have proof; well,
>> at least clear indications: http://blog.davber.com/2007/12/10/web-server-performance-shoot-out-simple-pages/
>>
>> Interesting to note is that ErlyWeb (just like the Haskell cousin,
>> HAppS) is FASTER with dynamic content than with static content (from
>> file...)
>>
>> From a non-performance aspect, it is interesting to note the
>> difference in meta programming between ErlyWeb - with its reflective
>> Smerl - and that of HAppS/Haskell with compilation-based code
>> generation; ok, HAppS is not using Haskell proper (i.e., 9 but
>> instead Template Haskell, which gives ALMOST as much flexibility as
>> reflection/introspection.
>>
>> Ok, see you soon!
>>>
>>
>
> >
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "erlyweb" group.
To post to this group, send email to erlyweb@googlegroups.com
To unsubscribe from this group, send email to erlyweb-unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/erlyweb?hl=en
-~----------~----~----~----~------~----~------~--~---
Post recived from mailinglist |
|
|
| Back to top |
|
| Guest |
| Posted: Tue Dec 11, 2007 9:00 am |
|
|
|
Guest
|
On Dec 10, 2007 8:23 AM, David Bergman <davber@gmail.com> wrote:
>
> Yariv,
>
> Yeah, that was weird. Well, two great minds often think alike (but so
> do two feeble minds...)
>
> I almost fell into that non-production trap myself, but remembered to
> include "-e production" for Rails in the last minute.
>
> Yariv, you know what? You and I should sit down and create a benchmark
> test for the frameworks I provided, along with Tomcat and .NET, to
> have a semi-official analysis cross-posted later. What do you say?
That doesn't sound like a bad idea, but I must warn you that it's a
lot of work. Trying out different load levels against different
configurations of a framework/server can take a lot of time. Just for
Erlang I want to try with/without HiPE and with/without kernel poll,
and now with Rails I have to try different Mongrel cluster sizes...
And every time you tweak something you realize there was something you
forgot or something else you should try out... (e.g. a multi-core test
would be interesting, too).
Yariv
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "erlyweb" group.
To post to this group, send email to erlyweb@googlegroups.com
To unsubscribe from this group, send email to erlyweb-unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/erlyweb?hl=en
-~----------~----~----~----~------~----~------~--~---
Post recived from mailinglist |
|
|
| Back to top |
|
| Guest |
| Posted: Tue Dec 11, 2007 9:10 am |
|
|
|
Guest
|
On Dec 11, 2007, at 3:59 AM, Yariv Sadan wrote:
>
> On Dec 10, 2007 8:23 AM, David Bergman <davber@gmail.com> wrote:
>>
>> Yariv,
>>
>> Yeah, that was weird. Well, two great minds often think alike (but so
>> do two feeble minds...)
>>
>> I almost fell into that non-production trap myself, but remembered to
>> include "-e production" for Rails in the last minute.
>>
>> Yariv, you know what? You and I should sit down and create a
>> benchmark
>> test for the frameworks I provided, along with Tomcat and .NET, to
>> have a semi-official analysis cross-posted later. What do you say?
>
> That doesn't sound like a bad idea, but I must warn you that it's a
> lot of work. Trying out different load levels against different
> configurations of a framework/server can take a lot of time. Just for
> Erlang I want to try with/without HiPE and with/without kernel poll,
> and now with Rails I have to try different Mongrel cluster sizes...
> And every time you tweak something you realize there was something you
> forgot or something else you should try out... (e.g. a multi-core test
> would be interesting, too).
Yes, it *could* be a lot of work. What I am talking about is adding
two frameworks and perhaps one more setting for ErlyWeb (HiPE) and one
more for Rails (multiple Mongrels on same box) and off we go. I have
no problem adding a few disclaimers and then people can choose to
ignore the results if they want. I.e., I do not want to create an
exhaustive (and exhausting...) benchmark test with strict conditions.
I want to give people an idea of performance differences and - quite
probably - get aggravated and explain that we were biased against
*their* specific favorite framework
/David
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "erlyweb" group.
To post to this group, send email to erlyweb@googlegroups.com
To unsubscribe from this group, send email to erlyweb-unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/erlyweb?hl=en
-~----------~----~----~----~------~----~------~--~---
Post recived from mailinglist |
|
|
| Back to top |
|
|
|
All times are GMT
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You cannot download files in this forum
|
|
|
Digg It
Del.icio.us
Reddit
Facebook
Stumble Upon
Technorati
