Seeking new test ideas for Framework Benchmarks Round 4+

https://github.com/TechEmpower/FrameworkBenchmarks/issues/133

by bhauer • 13 years ago

38 39 13 years ago

bhauer10• 13 years ago

The community has contributed even more pull requests since Round 2 of our web application framework benchmarks. We're planning to start Round 3 tests on Monday 4/15 using a new build of Wrk that allows time-limited tests (rather than request-count limited) so that all frameworks are tested for a uniform amount of time.

In the meantime, this Github issue request is seeking thoughts you may have concerning additional simple tests that we can introduce in Round 4 and beyond. We want to define tests that continue to exercise typical web application functionality but remain fairly simple to implement on a ever-widening field of frameworks.

If you have thoughts, please add them here or at Github. Thanks!

wheaties• 13 years ago

I have got to do a fork/pull before then in my infinite spare time but a quick question, why only ORM? Excuse my ignorance but there are almost no ORM that I've used which are not materially slower than straight SQL or json, the exception being g Salat on MongoDB. I wonder what not having an ORM would do to some of the poor performance experienced.

bhauer10• 13 years ago

We expect that most modern webapps are (still) developed with an ORM, but we're not ORM hard-liners. A few of the tests are run without an ORM and are identified with the suffix "raw."

Terretta4• 13 years ago

Thoughts of incremental web app functionality to test:

1. Exercising a randomized mix of reading and writing. I think you already said you were planning a CRUD test. Consider a tunable ratio here, something like 10000 R to 100 U to 10 C to 1 D.

2. Exercising synchronous web service (JSONP) calls in two modes: (a) to some web service that is consistently fast and low latency, say, the initial JSON example from this test suite running in servlet mode, and (b) to a web service written in the same framework as the one being tested, again using the initial JSON example.

(The idea here is that many frameworks fall on their faces when confronted with latency. This is why synthetic tests are usually so poorly predictive of real world behavior -- people forget that latency causes backlogs and backlogs cause all parts of the stack to misbehave in interesting ways.)

3. Test async ability if the framework has it, with a system call (sleep?) that takes a randomized 0 - 60 seconds to return. Would help understand when a framework is likely to blow up calling out to a credit card processor, doing server side image processing, etc.

4. Exercising authentication (standardize on bcrypt, but only create passwords on 1 in 10K requests), authorization, and session state, if offered.

5. Exercising any built-in support for caching, where 1 in rand(X) requests invalidates the DB query cache, 1 in rand(X) requests invalidates the WS call cache, 1 in rand(X) requests invalidates the long term async system call cache, and 1 in rand(Y) requests blows away the whole cache.

For the enterprise legacy integrators, it would also be interesting to test XML as well (in particular, SOAP), anywhere we're testing JSON.

bhauer10• 13 years ago

This is great input, Terretta. Exactly the kind of thinking I wanted to tap into.

Some quick thoughts in response:

For #1, the conceptual test included reading and writing in a 1:1 R:W ratio (well, to be more accurate 1:1 R:U). I like the idea of extending this a little bit to include C and D. For the sake of benchmark run-time, I'm looking to restrain the growth of permutations. But on the other hand, I like your idea of a tunable ratio. Something to think about!

I like #2 and #3 as well. I'll think about those some more too.

I really like the idea of incorporating some bcrypt and session state (#4).

We have a few caching tests in mind, but like elsewhere, we'll start out simple and then add complexity.

Thanks for the great input! This planning is to have some good long-term ideas in mind.

coolsunglasses6• 13 years ago

Can't wait to see how Go 1.1 does in round 3 :)

For anybody else interested in nonsense like this I have a Clojure template benchmark repository on Github here:

https://github.com/bitemyapp/clojure-template-benchmarks

I should probably update the clabango benchmark, some changes were made not too long ago.

bhauer10• 13 years ago

My oh my those are some awesome sunglasses on that bear.

We've got Go 1.1 in Round 3 and it's amazing.

darkchasma• 13 years ago

ASP.Net MVC?

bhauer10• 13 years ago

Sadly, we still haven't fit this in and no pull requests to-date. A Mono pull request would cause heartfelt cheers.

polskibus• 13 years ago

+1 for the .NET , both mono and microsoft versions. It would be great to compare ASP.NET MVC against the most popular web frameworks out in the open source world. If it's a matter of licenses then perhaps you could apply for bizspark and get windows licenses for dev purposes for free?

CoffeeDregs• 13 years ago

As a passive consumer of the benchmarks/info, I've got to say a huge "Thanks" for this! I've been a watcher of the Debian benchmarks game for a long time (since its inception), but, after using Python, Ruby, Haskell, etc, had written off JVM languages as either slow (e.g. Groovy) or non-expressive (e.g. Java)... This benchmark has got me seriously interested in JVM languages again. (Note: I don't want to use Scala^H^H^H^H^HFrankenstein)

Given the computational ability of browsers, the "front end" is rich and the "back end" is shrinking down to just the API. I'm not sure that the extra verbosity of Java is too large a cost for the performance it offers.

I currently have an application doing 1500 requests a second on Django and it's using 3 servers. I could have used only one server if I'd used Java or Clojure? Hmmm...

coolsunglasses6• 13 years ago

> I could have used only one server if I'd used Java or Clojure? Hmmm...

Use Clojure, but keep in mind that if you're not caching aggressively you can saturate persistence-side quickly - making your choice of application language irrelevant.

spullara3• 13 years ago

Sadly that isn't very true anymore. Databases (and I/O — think SSD) really are faster than most of the dynamic languages these days.

bhauer10• 13 years ago

We've got more Haskell, some Erlang, Lua, more Scala, a bunch of PHP, and myriad other goodies in Round 3 thanks to a bunch of pull requests. I think Round 3 is going to show a lot more competition at the top tiers.

meddlepal• 13 years ago

The Groovy is slow argument has sort of run out of steam in my opinion. Support for invok-dynamic and the static compilation annotation provide a lot of speed. I use Groovy daily and my experience is anecdotal but maybe it is time to do some benchmarks of my own.

vorg• 13 years ago

Dynamic-mode Groovy is as slow as other dynamic languages. The recently added static-compilation mode for Groovy still regularly spews up bugs, perhaps because it was written by a single programmer with little beta testing whereas Scala and Java have many high-pedigree developers behind them, and heaps of documentation and testing.

happyhappy007• 13 years ago

I would like to see a simple CRUD blog built with different frameworks. Building a blog is like the "Hello world!" for dynamic web development.

coolsunglasses6• 13 years ago

Given the scope and breadth involved in these benchmarks, that's a helluva tall order. I'm sure nothing's stopping anybody from doing it themselves though.

kbenson6• 13 years ago

I think that this could be specced out in stages, and implemented an a number of rounds. First would be a schema for a blog, with authors, posts an comments. Next would be a rest API for posts and comments. Finally mock pages to be used for posting, reading and commenting in HTML to test the included templating system, if there is one.

You really need to go at least this far. This will also give you an approximate code size for thus sample project as well, which is is at least as important as performance to some people.

crypto5• 13 years ago

Do we have any chance to look at round 3 results?

bhauer10• 13 years ago

Of course! We are starting the tests next Monday and you can expect them to be complete some time around the middle of next week.

ckluis• 13 years ago

@bhauer - this may be the best marketing I have ever seen a company do in the tech space (and by marketing I mean marketing to developers). This is a recruiting goldmine.

Sheer genius.

voidlogic• 13 years ago

What about tracking memory usage? Peak, average, etc.

Previously you tested EC2 vs Local HW. What about adding local KVM virtual machine as well?

I also think a graph showing latency as a function of concurrency would be very interesting.

bhauer10• 13 years ago

Thanks for the ideas, voidlogic. We do want to capture server statistics and have that as an issue in Github [1]. I am particularly interested in capturing CPU and I/O utilization because in spot checks, we've observed some frameworks do not fully saturate the 8 HT cores on our i7 hardware, suggesting lock or resource contention.

As for a variety of other hardware and VM environments, the data would be interesting. Related: we plan to migrate the charts and tables to a stand-alone page. Right now, the blog entries are hard-coded to fetch two specific results.json files for rendering the charts/tables. But when we build a stand-alone page, I would like to enhance the script to allow selection of one or two results.json files from a menu for comparing side-by-side. And to your point, the community could then contribute their own results files. Imagine being able to compare EC2 large vs xlarge or vs Xeon E5s or ...

Right now, as you noticed, the latency is only displayed at 256 concurrency. I'll make a note to myself to include a chart for latency versus concurrency when we move to a stand-alone page [2].

[1] https://github.com/TechEmpower/FrameworkBenchmarks/issues/10...

[2] https://github.com/TechEmpower/FrameworkBenchmarks/issues/14...

Seeking new test ideas for Framework Benchmarks Round 4+

Related Stories