• Ozan (16/16) Feb 26 2016 Hi
• maik klein (3/19) Feb 26 2016 It seems to be at the bottom because of "vibe.d — Did not
• =?UTF-8?Q?S=c3=b6nke_Ludwig?= (8/23) Feb 26 2016 Unfortunately they didn't look at my PR

Ozan <ozan.sueel gmail.com> writes:

Hi
I found a benchmark page for web frameworks.
http://www.techempower.com/benchmarks/#section=data-r12&hw=peak&test=query

Here, vibe.d was lost somewhere in bottom of slowest frameworks, 
what surprise me.
I tried some test out, compared jetty, servlets, grails, php and 
vibe.d with completely different result.
Running on my laptop and measured by Chrome's network time 
function vibe.d was in ms area, jetty close to seconds and php 
between. Servlets and grails are the slowest on my machine (Linux 
Mint 17.3 / 64bit / 16 GB RAM / i7)

And from the technical point of view, I'm sure some results must 
be wrong.
So I shake my head in surprise and ask the community what they 
think about techempower's benchmarks.

Regards, Ozan

maik klein <maikklein googlemail.com> writes:

On Friday, 26 February 2016 at 12:14:03 UTC, Ozan wrote:
 Hi
 I found a benchmark page for web frameworks.
 http://www.techempower.com/benchmarks/#section=data-r12&hw=peak&test=query

 Here, vibe.d was lost somewhere in bottom of slowest 
 frameworks, what surprise me.
 I tried some test out, compared jetty, servlets, grails, php 
 and vibe.d with completely different result.
 Running on my laptop and measured by Chrome's network time 
 function vibe.d was in ms area, jetty close to seconds and php 
 between. Servlets and grails are the slowest on my machine 
 (Linux Mint 17.3 / 64bit / 16 GB RAM / i7)

 And from the technical point of view, I'm sure some results 
 must be wrong.
 So I shake my head in surprise and ask the community what they 
 think about techempower's benchmarks.

 Regards, Ozan

It seems to be at the bottom because of "vibe.d — Did not 
complete"

=?UTF-8?Q?S=c3=b6nke_Ludwig?= <sludwig outerproduct.org> writes:

Am 26.02.2016 um 13:14 schrieb Ozan:
 Hi
 I found a benchmark page for web frameworks.
 http://www.techempower.com/benchmarks/#section=data-r12&hw=peak&test=query

 Here, vibe.d was lost somewhere in bottom of slowest frameworks, what
 surprise me.
 I tried some test out, compared jetty, servlets, grails, php and vibe.d
 with completely different result.
 Running on my laptop and measured by Chrome's network time function
 vibe.d was in ms area, jetty close to seconds and php between. Servlets
 and grails are the slowest on my machine (Linux Mint 17.3 / 64bit / 16
 GB RAM / i7)

 And from the technical point of view, I'm sure some results must be wrong.
 So I shake my head in surprise and ask the community what they think
 about techempower's benchmarks.

 Regards, Ozan

Unfortunately they didn't look at my PR 
(https://github.com/TechEmpower/FrameworkBenchmarks/pull/1916), which 
was submitted before their deadline for changes that still go into round 
12. 0.7.27 had a performance bug where it basically didn't scale to 
multiple threads at all. They are using a 24-core machine or so, so the 
penalty is huge. I'll have to have another look at why some of the tests 
didn't complete, though.