echoprint server speed

[Martin Quevedo]

Hi all,

I just took an MP3 file (Eminem- If I Had.mp3), cut a 30 seconds piece and ran it with codegen.

I used the the resulting code against echonest server and it took about 2 seconds to give me the answer.

I did the same thing with a local server I have here and it took about 15 seconds to decode.

My server has 2 Intel XEON 5140 processors with 8GM RAM.

Is it possible that this huge difference in response time is due to hardware capacity?

Is there any other possible reason?

Regards,

Martin

[David DesRoches]

Martin,

How many tracks are indexed in your local server?

The servers behind The Echo Nest's Echoprint identification API are quite powerful with fast IO to support many simultaneous requests. I wouldn't make too much of the disparity with your local server. If you want to improve performance on your single machine, take a look at caching the index in RAM or using a fast RAM- or SSD-based storage.

David

--
You received this message because you are subscribed to the Google Groups "echoprint" group.
To unsubscribe from this group and stop receiving emails from it, send an email to echoprint+...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--

David DesRoches

Lead Audio Engineer

The Echo Nest

deli...@echonest.com

[Martin Quevedo]

Hi David,

 

I ingested all json files in http://echoprint.me/data_download site.

It is approximately 1 million songs

 

Is there an special process to index the ingested songs?

 

I do have the feeling it should be faster.

 

Regards,

Martin

[Daniel Cukier]

You will have problems with a lot of tracks. I noticed some improvement when I changed solr index to an SSD hard drive.

Take a look at our branch in github:

https://github.com/playax/echoprint-server

We did some small performance improvements.

Since echonest is not updating nor accepting pull requests for the echoprint-server, I think the community will have to take this project on by itself.