Go as an alternative to Node.js for very fast servers (bbs.studygolang.com) |
Go as an alternative to Node.js for very fast servers (bbs.studygolang.com) |
If you send 3KB responses then you would see both setups are much closer in performance.
Factor in some actual I/O and the difference will be even less. Then you'll eventually realize using either one makes little difference when it comes to performance.
This is why micro benchmarks are pure jokes. A real world site has a mix of response sizes, database I/O and caching. It's the only way to test something properly and you'll see if you properly test both there will be little difference in performance.
Please note that the microbench was "rigged" by the author of Node when he was first presenting it several years ago (spelled out in the article).
If you need a tl;dr, it's this: I don't care for JavaScript as a language. Many make the argument that JavaScript should be adopted widely server-side because of its speed. I assert that languages should be evaluated not only for performance but for maintainability, feature sets, standard library, etc. Go provides a great combination of execution speed, development speed, and ease of maintainability.
The default template language is also really archaic and no one has created a solid alternative yet that's actually well tested and used by the masses.
It might have good execution speed and the language itself might be nice but the only thing that matters is going from point A to point B. Go will not get you there faster than other languages and the execution speed is a non-issue for pretty much every platform (even rails) if you use tools available to you to their fullest.
P.S., I compared Go to Node almost a year ago and even wrote a mini framework for Go to resemble a smaller version of Express. I eventually just said fk it and stopped because the gains were not even close to being worth it.
Just something to be aware of if anyone reading has Go 1.0.2 installed and has not yet upgraded. 1.1 is worth the upgrade.
[1] https://code.google.com/p/go/source/detail?r=45c12efb46
[2] http://www.techempower.com/blog/2013/05/02/frameworks-round-...
That explains the use of Go 1.0.2.
Summary of results:
1. Node.js v0.10.15, single worker: 46.2 seconds
2. Node.js v0.10.15, cluster 8 workers using naught: 17.2 seconds
3. Go 1.0.2, GOMAXPROCS left default: 3.5 seconds
4. Go 1.0.2, GOMAXPROCS=8: 3.7 seconds
Detailed results below:
1. Node.js v0.10.15, single worker
Concurrency Level: 100
Time taken for tests: 46.217 seconds
Complete requests: 10000
Failed requests: 0
Write errors: 0
Total transferred: 10486510000 bytes
HTML transferred: 10485760000 bytes
Requests per second: 216.37 [#/sec] (mean)
Time per request: 462.168 [ms] (mean)
Time per request: 4.622 [ms] (mean, across all concurrent requests)
Transfer rate: 221580.08 [Kbytes/sec] received
Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 0 0.2 0 3
Processing: 193 461 36.2 450 944
Waiting: 16 235 127.3 235 534
Total: 193 461 36.2 450 944
Percentage of the requests served within a certain time (ms)
50% 450
66% 467
75% 470
80% 486
90% 492
95% 514
98% 517
99% 545
100% 944 (longest request)
Concurrency Level: 100
Time taken for tests: 17.199 seconds
Complete requests: 10000
Failed requests: 0
Write errors: 0
Total transferred: 10486510000 bytes
HTML transferred: 10485760000 bytes
Requests per second: 581.41 [#/sec] (mean)
Time per request: 171.995 [ms] (mean)
Time per request: 1.720 [ms] (mean, across all concurrent requests)
Transfer rate: 595408.80 [Kbytes/sec] received
Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 0 0.2 0 3
Processing: 7 171 116.4 149 739
Waiting: 5 96 81.9 71 710
Total: 8 171 116.5 150 740
Percentage of the requests served within a certain time (ms)
50% 150
66% 197
75% 236
80% 266
90% 324
95% 397
98% 438
99% 493
100% 740 (longest request)
Concurrency Level: 100
Time taken for tests: 3.542 seconds
Complete requests: 10000
Failed requests: 0
Write errors: 0
Total transferred: 10486730000 bytes
HTML transferred: 10485760000 bytes
Requests per second: 2823.16 [#/sec] (mean)
Time per request: 35.421 [ms] (mean)
Time per request: 0.354 [ms] (mean, across all concurrent requests)
Transfer rate: 2891181.71 [Kbytes/sec] received
Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 1 0.3 1 3
Processing: 9 35 2.2 34 56
Waiting: 0 1 1.3 1 22
Total: 12 35 2.3 35 57
Percentage of the requests served within a certain time (ms)
50% 35
66% 36
75% 36
80% 36
90% 37
95% 38
98% 39
99% 41
100% 57 (longest request)
Concurrency Level: 100
Time taken for tests: 3.657 seconds
Complete requests: 10000
Failed requests: 0
Write errors: 0
Total transferred: 10486730000 bytes
HTML transferred: 10485760000 bytes
Requests per second: 2734.54 [#/sec] (mean)
Time per request: 36.569 [ms] (mean)
Time per request: 0.366 [ms] (mean, across all concurrent requests)
Transfer rate: 2800429.67 [Kbytes/sec] received
Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 1 0.4 1 3
Processing: 19 36 2.5 35 57
Waiting: 0 1 1.1 1 16
Total: 20 37 2.5 36 58
Percentage of the requests served within a certain time (ms)
50% 36
66% 37
75% 37
80% 37
90% 38
95% 39
98% 42
99% 51
100% 58 (longest request)http://www.techempower.com/benchmarks/
In addition to various JVM technologies, there are faster technologies in C++, PHP, and Lua.
Further, once you apply a framework to paper over some language warts, performance is often terrible and can only be rescued by extremely liberal caching.
If this were not so, it would be hard to see the motivation for HipHop.
Which is often the huge array of JARs and folder structure you typically need to bootstrap.
It compiles a binary, just send it to the server and it will run, no need to even have go installed on the server (or any libs really).
Am curious to find out the results. There was a recent, don't remember which, comparison that put Scala way ahead of Go possibly because of a superior GC.
Wonder if the results correlate
My eloquent post was eaten by an expired link on the original article but here are some counter-points:
The commutative property applies to all code that has semicolons to end lines. Use jslint/jshint/an IDE. The symbol is "optional" not marked as "leave it out because it makes scripts nicer."
Use jslint/jshint/an IDE to prevent globals. Seriously. It's the same as running go fmt on your code.
Forcing people to use "channels" as a best practice to accomplish scaling is the same as the best practice of callbacks - but yes, channels are nicer to use. Node standard modules require callbacks by default as Go standard library implements channels by default.
32-bit integers in JS don't have float problems because the precision doesn't break
Typing in Go can still be annoying for some situations. If you're dealing with external content (creating an API with mutable content that you still need to read) it can be annoying at best (e.g. reminds me of writing C). Types in Go are a nice implementation though.
npm is way better than go get and there are at least 3 projects in Go trying to replicate npm's ease-of-use
The vim/emacs syntax highlighter is nice but it's awfully frustrating if you don't use vim/emacs. This is due to Go's young age, but you shouldn't be forced into using a certain editor to get syntax highlighting.
The alternative to "go fmt" for Javascript is a good IDE or jslint/jshint.
In Go, you get a lot in the standard library, but you miss out on a lot in the community. Yet another young language problem, you wind up having to roll your own for a lot of things that should just exist. It can be frustrating looking for answers because you might be the first person working on the problem in the language.
Also, just as a general thing I've perceived, people seem to argue static types vs dynamic types more than that a particular language is "better." Go and JS/node are both great!!!! Go is typed, JS is not and you deal with the consequences in both situations. Static typed languages have big faults with external data handling that sometimes cripple features (or makes them far more difficult to accomplish). Dynamic typed languages can fall victim to variables being used incorrectly (especially when not using an IDE). That's the biggest bulk of the difference in my experience.
It seems like this site confirm there must be a pretty dedicated Golang following, but I cannot figure out for the life of me why. Does anyone actually know?
Just my two cents.
Yes.
> If so it seems it would be possible to write highly functional code given the flexibility of interface{}
Not really, you'd have to add type assertions everywhere as go has neither generic functions nor user-defined generic types (only a handful of special-status types get to have type parameters, IIRC they're chans, arrays, slices and maps). That makes higher-order operations extremely cumbersome.
I'm also uncertain whether scalars (e.g. integral types) can be used through interface{}.
They can (and there are optimizations to avoid heap allocation in some cases when you do use them), but you still have to write downcasts everywhere.
It's a really small library that was made in response to my own intuition (similar to yours) that the Go standard library wasn't really embracing a functional style. I haven't used it in any major projects but for small one-off things it's proved pretty useful. It basically works like pipes on the command line.
go version go1.0.2
runghc 7.6.2
cabal packages of today
import qualified Network.Wai as Wai
import qualified Network.Wai.Handler.Warp as Warp
import qualified Network.HTTP.Types as HTTP
import qualified Data.ByteString as ByteString
import Blaze.ByteString.Builder.ByteString (fromByteString)
main = do
let port = 8000
Warp.run port app
app req = do
let n = 1024*1024
let bytes = fromByteString $ ByteString.replicate n 100
return $ Wai.ResponseBuilder HTTP.status200 [] bytesAlternatively you can implement everything on interface{} and force the user of the higher-order function to use type assertions everywhere.
[1] - http://blog.burntsushi.net/type-parametric-functions-golang
1. They are not using the latest Go (atm 1.1.2)
2. GOMAXPROCS is not = the number of CPUs
3. They are using ab rather than something more scalable like wrk
On a side note, Go is already known to be much faster at web-serving than node.js: http://www.techempower.com/benchmarks/#section=data-r6&hw=i7...
Also in this bench Go danced circles around node. I dunno what you're complaining about.
For example, when I go to StackOverFlow I see that Node has far more questions asked:
Being developed by Thompson and Pike makes you gain something like 30 years of "maturity". Plus, running in production in the Google infrastructure is far more of a proof of maturity than running the chat service of every hackathon project for 2 years.
// When calling .end(buffer) right away, this triggers a "hot path"
// optimization in http.js, to avoid an extra write call.
//
// However, the overhead of copying a large buffer is higher than
// the overhead of an extra write() call, so the hot path was not
// always as hot as it could be.
//
// Verify that our assumptions are valid.
Indeed, that's probably the biggest flaw I've found working with GO's type system. Trying to work with unknown n-level JSON is real pain.
If your JSON is effectively "strongly typed" (most APIs are), this is going to be a huge win for you.
If your JSON is not, then you'll have a problem in any statically typed language (not just Go), because you need some way to reason about the type. You'll also have the same problem with dynamically typed languages as well - the main difference is that Go will never do implicit casts (I would view this as a good thing).
I've done a lot of work in Go involving JSON (that's originally why I wrote the above tool - to save myself time), and in practice, it's rare that I have to do anything more than decode, check for an error[0], and then move on.
[0] Which is something everyone should do in all languages, not just Go - once you've confirmed that there is no error, you rid yourself of a lot of possible bugs that could pop up later on in harder-to-discover places.
Unless you try multiplying them.
I'd say Go isn't an impressive programming language, but a very impressive software engineering language.
Curious, what is the distinction?
Giving one guy from the Ukraine the typical choices at a restaurant here in the US would freeze him in place.
What would you like to drink? Tea
Sweetened or Unsweetened? Sweet. frustration += 1
Sugar or artificial sweetener. Sugar. frustration += 2
Soup or Salad? Soup. frustration += 7
Which soup? Minestrone, Pasta Fagioli, Italian Wedding? frustration += 14
Then there was the guy from Iceland... He took 10+ minutes to decide what to order in a new place and once he figured out what he liked in a particular place, he never gave the waitstaff a chance to offer him a choice. This, well done, salad with x dressing, with unsweetened tea. Iceland liked making the choices but only a few times.
Some people like to be dictated to: Code will be formatted like this. Braces will be like this. etc. The author of the article thinks these are a feature while I find it obnoxious.
My brain requires braces to line up in the left column, anything else slows me down. Maybe I'm just old?
I use the ML approach of combining other people's algorithms. :)
My first time in a new place, I'm usually with friends, and I imitate their orders. On the rare times when I'm not, I ask the waiter what the favorite plates of he/the other people in the restaurant are, what they order when they eat there.
This way even my initial orders are lightning-fast and already judged as 'tasty' in comparison to other offerings. And over time, I can investigate other plates for myself.
> My brain requires braces to line up in the left column,
> anything else slows me down. Maybe I'm just old?
http://bbs.studygolang.com/thread-278-1-1.html
Notice that, despite English content, the forum looks Chinese-language oriented. I think it is cool, but not normal to me.
I expect the choice of Chinese as the language is probably the fact that you can't play code page games very easily with Chinese, which you could with Korean, Arabic, etc. It is probably the most widely spoken language with a non-Latin character set. Also, chances are very good that a Chinese-writing colleague of the example writer was readily available. Although, who knows, it would be fun if it were because of the board game.
> They seem to be making a concerted effort to make Go Chinese friendly https://code.google.com/p/go-zh/
The Chinese fork seems to be a project started by Minux. Minux (Shenghou Ma) is a Go contributer and is very active in the Go community, but he does not work for Google, as far as I know.
Go is a clean, intuitive language with a very robust supporting ecosystem and a strong concurrency model. It builds fast, small code and makes quick work of big problems. Why shouldn't there be a big following of one of the better platforms to come out in a long time?
The runtime kind of makes it a silo; ie: hard to bind other languages to it through an FFI.
Of course if I am mistaken or there's something being done to address such a scenario then I will be much happier seeing more and more infrastructure code shipping in Go.
I think this "benefit" of using the same language for front and backend is pretty over-hyped, as well. In theory, I can agree that it sounds good. In practice, use the best tool for the job on both ends to fit your team's abilities and strengths.
The other neat thing is that if your frontend and UI are cleanly separated, you can swap them out individually without the other even noticing. If our backend gets to be too sluggish, I can replace it with something lower level gradually over time (Go?) without leaving a bunch of deprecated backend JS cruft to clean out.
Just to give one example: any data query can get answered locally or from the server, depending on what's already cached and whether you're online. Before, I had to implement every call twice and make sure they stayed in sync. Now I can implement once and run the same code in both places.
And other interesting possibilities open up. The server can just run the client's data synchronization code to pull changes from another server, giving realtime server-to-server replication nearly "for free".
That said, you hire "backend-focused" developers, "front-end" developers, etc., but to me that still means they should be able to do whatever tasks are necessary (Go, JS) to get their job done.
However, one of the more frustrating pieces of this is that I never dive deep into one paradigm. I'm continually having to work bilingually. When I was working with node.js, I didn't feel nearly as much bilingual stress (even though I've written backends in python, ruby, java, groovy, and scala!)
As such, my single person one-off and side projects will be in node.js. It's about cognitive overhead in the moment. I agree that it's a mediocre solution for larger teams and enterprise is not its sweet spot, but it's great for projects in the small.
I really don't value this type of response where the benchmark is brought into question, yet nothing is substituted as evidence. In the few cases were someone does claim that they rewrote a sufficiently complex system in Node from some other language, it's impossible to discount that the performance changes come from architectural choices rather than language choices.
Playing devils advocate, I've heard the same claim you put forth here about Node.js: it gets much slower in real apps because you have more slow JS code being run vs the very fast C libraries that back the core of Node.js.
To me, benchmarks like the one I posted above are the most compelling form of evidence we have. What I take from it is that a great many "boring" languages and frameworks are really very fast. It's not the answer that most people want to hear of course; it goes against the current popular hype.
Much better at 100 reqs/sec and up.
I'm working on a webapp that mimics functionality of an existing desktop application -- mostly as a self-educational project. As I'm writing it and adding features, I can push processing from server-side to client-side and vice versa with hardly more than a copy and paste of the relevant chunk of code.
Say I don't really care about the security of some particular process, and I really don't want to use up any additional resources on the server for it, I can just push that load to the client browser in a couple clicks.
Did I just write some feature into the client that I suddenly realize presents a security problem? With node, there's no need to change my frame of mind or rethink how to do something in a different language -- just copy and paste from client.js to server.js and do some slight cleanup.
I've been paid for work in roughly 12 different programming languages over a 28+ year career, among them python (no braces at all, significant white-space), Rexx (keywords as braces), C, C++, Perl, Pascal, Bourne shell, Awk, C#, (braces braces and more braces), x86 assembler and several employer specific scripting tools.
Guess which ones are most comfortable? Certainly not python though it is the most bang for the buck of the languages I use and I'd rather eat hospital food for the rest of my life than write (or, deity forbid read) one more line of x86 assembler.
What was it we were talking about again?
Maybe, anecdotal etc...
It was simply something I noticed from lots of years working with lots of people from lots of different countries. No offense was intended.
In the same sense that a language that "in theory" has tons of cool stuff might be unusuable in practice (due to complexity, opaqueness, lack of libs, strange syntax etc).
The word "rigged" in your original comment implied dishonesty on my part. I just wanted to clarify that the JS code used in the microbench was Ryan Dahl's, and mine was just a port to Go. I was merely giving Go the same task that he did.
I don't want to spend most of my time solving boring issues. I want to spend most of my time writing features for apps I make. Being productive makes me happy but everyone has different happiness triggers I suppose.
I see performance as a somewhat solved problem in 99.999% of the cases so to me a language is not even worth looking at anymore if that's all it offers. I'm not fortunate to be involved in the other 0.001% where easy to do caching might not be enough.
I didn't mean to imply you were dishonest. I just wanted to point out that sending out 1MB responses is kind of silly.
It took me like 2.5 years of messing around to finally realize all I care about is being able to take an idea and turn it into working code that's maintainable and scales good enough for the time being.
Luckily, Go has a decent head of steam behind it, so when I revisit it in a few years that should be a solved problem ;).
In this case, language maturity means the quantity of reference material available to help troubleshoot problems as the arise.
This is not a stab at Go or its maturity, but rather a realistic assessment of the importance of answers being (nearly) instantly available when you're working on a project or troubleshooting a critical issue.
And nobody cares if Go is being used for some tiny, insignificant part of the Google infrastructure. Get back to me when it is used for a stock exchange, betting site or complex web app.
complex web app: Cloudfare https://www.cloudflare.com/railgun BBC http://www.quora.com/Go-programming-language/Is-Google-Go-re... Soundcloud http://backstage.soundcloud.com/2012/07/go-at-soundcloud/
I'm complaining because you and other people will go on to:
1. Use a version of Go in your development that is slower (as has much worse memory characteristics) and lacks new features.
2. End up running all your programs on single core until you understand GOMAXPROCS
3. Use ab to bench real things which is bad
Go 1.1 > Go 1.0.2
wrk > ab
Wrk uses a slightly different argument syntax from ab and WeigHTTP but has some upsides:
1. Wrk is also multi-threaded.
2. In our experience, wrk is slightly higher-performance than WeigHTTP (~5 to 10%).
3. Wrk provides average, maximum, and stdev for latency.
4. Wrk provides a time-limited mode (rather than request-count limited), which is appealing for some test types.
In my experience, as long as you configure Go and node to use all of your cores, Go will benchmark better than Node in any permutation of these configuration variables:
Go 1.0.2 and node tested with ab.
Go 1.1 and node tested with ab.
Go 1.0.2 and node tested with wrk.
Go 1.1 and node tested with wrk.
Sorry, that was not intended.
>but his intentions were good. :)
They really were...
Can you please expand on why? I recently bumped on wrk and am in process of evaluating switch from ab, thank you
For example you said: GOMAXPROCS left default. I don't know how you set your environment vars, are they unset so default = 1? You didn't mention in your post that GOMAXPROCS=1/single node worker test cases are really toy test cases (useful only for benchmarking). So if you know everything below, then great! Maybe other people can learn:
GOMAXPROC is the number of OS level threads that the Go runtime is multiplexing Go tasks (goroutines) over.
So if GOMAXPROCS = 1, When one goroutine blocks, another will run, BUT, you will never use more than one OS thread and thus you will never use more than one logical core.
Setting GOMAXPROCS correctly is per application. For example GOMAXPROCS=1 might be right for a commandline tool or a program that was designed to have multiple instances started on the same machine. That being said, a vast majority of the time any high load application I have written is best with GOMAXPROCS=<# of logical CPU cores>. So Go always has concurrency, but GOMAXPROCS gives it parallelism. GOMAXPROCS > 1 also will allow the garbage collector to have more parallelism too.
So if we are talking about a benchmark like this, ideally we want to process requests made in parallel in a parallel fashion. A clear sign is that if you use Node.js worker cluster you should probably test with Go at the same number.
All this being said, depending on your CPUs implementation details, you would sometimes be better off setting both your node worker count and GOMAXPROC to the number of physical rather than logical cores. Sometimes simultaneous multi-threading (SMT, aka hyperthreading) actually creates more overhead than any concurrency gains it offers.
In short when testing something like this I would always test. 1. n = 1 (with a disclaimer note) 2. n = physical CPU count 3. n = logical CPU count Where n is the number of GOMAXPROCS/node worker threads.
It's not. It's a new Java, not a new C++.
> ie: hard to bind other languages to it through an FFI.
More or less impossible: GC and goroutines are not optional so you'd need to cleanly setup and shutdown the Go runtime. You'd have to embed Go as you do Lua or Python.
> It's not. It's a new Java, not a new C++.
Not according to the Go developers. "Go is a general-purpose language designed with systems programming in mind." http://golang.org/ref/spec#Introduction
Go has absurdly simple FFI through, as mratzloff mentioned, C. This is self-promotion and apologies, but see my submission history for two examples.
* Download Go source
* Extract, cd go/src/
* GOOS=linux GOARCH=386 ./make.bash #this will build the linux_386 toolchain
* GOOS=linux GOARCH=amd64 ./make.bash #linux_amd64
* GOOS=linux GOARCH=arm ./make.bash #you get the picture
GOOS can be windows, darwin, freebsd, netbsd, plan9.
Then when you want to cross-compile your app, you do:
GOOS=linux GOARCH=arm go build myApp.go
That's it. Now you have a statically linked binary that you can drop on whatever your target is. As someone who has had to cross-compile a lot of C and C++ code, I find this simplicity to be a huge win.
That would give you a better estimate of the risk you're taking by using that language. And that's where the 30 years of PL design from GO's author gets some importance.
You could consider my argument a bit far fetched, but GO's design has been explicitly focused on keeping things simple and no-surprise, and so far all the reviews seem to agree on that point. That could compensate for the lack of results in Google (especially compared to what node.js forces you to do)
However, it would probably only help if you use -threaded (which you should, anyways).
Also "-threaded" gives the usual penalty of bookkeeping and bad caching and stuff. You don't need threads in such a benchmark. Though I am mildly surprised that "node" runs faster with multiple processes :)
Use: ghc -threaded -O2 main.hs && ./main
I suppose maybe it uses the optimized compiled packages code, and the main code doesn't really do anything, so -O2 doesn't matter.
However, I would still expect "-threaded" to have an effect.
I did use GOMAXPROCS with the number of logical cores that I have, and I did test the node cluster with the same number.
In the domain I'm in (internal enterprise cough) I find that most people are expected to be full stack. On the other hand, these tend to be more simple UIs in general, so the front end knowledge isn't as vital (cue 'good' full stack developer comments...)
Don't forget your form needs to be comprised of up and down submit-buttons; wouldn't want the progressive-CSS lords coming down on us.
Those fearful of JS come along... Collapse comments? Refresh the page, but don't forget to pass previous state of comments previously collapsed! Submit an up arrow? Don't forget that blob of state! Loaded deep-nested replies? Push it into state! Careful; none of this matters to 99.9% of your users, because they are understanding they'll lose the statefulness on refresh, while everything stateful happens inline and without refreshing the page and bringing in sections of the page as dictated by our stateful blob.
For example, if you ever tried PE for real you would know that there is no point in implementing collapsible comments in pure HTML. That's exactly the kind of functionality you "enhance" via JavaScript. Of course, like most modern developers you've never really tried the practice, so you don't have any experience with such things.
I have a main screen with a query whose processing takes over a minute to perform it on all of the data, but seconds for incremental changes after that. Each row returned has a detail screen associated with it, and the users have a requirement to be able to move to a detail screen, update a record, and return to the main screen without continually incurring the heavy query penalty each time they return to the main screen. Business also has a requirement that the database of record is normalized. (Denormalizing and precomputing is good, but you run into problems when the feed updates.) Do you...
A. keep a Nosql (say... redis) cache, precompute, and perform two commits? This adds a non-trivial amount of complexity keeping them in sync, not including watching for jobs from outside feeds that can update at any time. This isn't bad for a simple system, or a system with a lot of engineers, but it's non-trivial to get right and easy to find out a bug after the fact.
B. Keep it all in the session? This has its own problems: if using an ORM, you're going to be increasing your memory load per user substantially. Otherwise, you're having to keep a separate data structure around and marshal/unmarshal, creating an additional complexity layer, and I haven't seen a great way to do this.
C. Keep a single page application and implement a polling mechanism behind the scenes to keep the information fresh? With a system like Ember or Angular, this may not be trivial but it's not terrible either, and I'd argue it's superior to a redis cache.
I know that I can provide a better user experience by prefetching. I don't know how to do this with progressive enhancement.
http://en.wikipedia.org/wiki/Hyper-threading
The speedup is very work dependent and in practice for things like web pages and api servers you generally only get another 20-40% of performance from them rather than a full 100%.
A hyperthreaded Intel CPU has M functional units and N decode/issue pipelines.
A non-hyperthreaded Intel CPU has M' functional units and N' decode/issue pipelines.
A hyperthreaded Intel CPU with hyperthreading disabled has M functional units and N/2 decode/issue pipelines.
A hyperthread core is a virtual core -- it is not actually a core at all but is a re-purposed, possibly stalled physical core. While it can improve some scenarios, in some cases (particularly core-saturating benchmarks) it can actually hurt performance.
This is hardly an out there or controversial statement. Further I didn't say to disable hyperthreading, I said to try setting parallelism to the physical cores. Again, nothing, whatsoever, controversial about that.
If you don't disable hyperthreading, but instead run four threads on an 8-thread CPU, it is extremely likely that the threads will be scheduled on the first two cores/four threads and the other two cores will be shut down, especially on the newer intel CPUs with "turbo" features where this strategy can have large benefits.
This is common knowledge, and your laughable obnoxiousness, which anyone who has ever worked with multithreaded code on a HT processor knows is farce, rings pretty ridiculous.
That's the reason intel tells you to shut hyperthreading off if your operating system doesn't support it.
The gorilla toolkit's APIs are inadequate and it seemed like quite a few people agreed too because most of them said they rolled their own solutions to do things that certain gorilla libs did but with a more intuitive and friendly API.
Go really isn't that young either. It's been what at least 4 years now? There's no excuse. It's not like the language is 6 months old.
As for re-inventing stuff, it's more so go's ecosystem rather than revel's shortcomings although revel does have its own shortcomings if you were to compare it to something like rails and not express.
Revel seems to be somewhere in between rails and express in terms of opinions which is fine but if it's going to make me less productive then I'm simply not going to use it.
1.5 if you count from the first stable release. Which IMHO is what matters, before was just a experiment with a lot of uncertainties. It took ruby 9 years to arrive from 1.0 to Rails. They were other times, sure, but still.
I still consider it very young, or at least I don't know of any other younger language with a better ecosystem.
> Revel seems to be somewhere in between rails and express in terms of opinions which is fine but if it's going to make me less productive then I'm simply not going to use it.
Well, everyone has their preferences, certainly in Go there are not many choices so is not for everyone. But I'd keep an eye on it. Things can change very quickly.
9 months to me is a huge amount of time. I don't want to have to wait years to be super productive. I want to be super productive right now and by using other platforms I can be.
For a new viable web platform to be accepted it needs to really explode in popularity. It has to offer MASSIVE gains.
Look at node, it offers performance and also offers the benefit of using the same language on both ends. That's pretty neat... maybe, but I think you would at least agree with me that node's popularity and growth has been unmatched. Even so, it's still quite far behind rails and I don't think it will catch up.
I'm not some massive rails fan boy either. I only started using it when 4.0 came out because the ease of caching seemed interesting to me and I was looking for an excuse to go from node/express to something more opinionated just to see if it was more productive.
And most importantly, you've now created two code paths, both of which need testing, both of which require maintenance, and one of which (in a web application) will never be used. What's the ROI?
Now, for some domains (say, e-commerce), progressive enhancement is the right way to go. I've done it. For web applications, which are the extension of client-server applications designed for a limited captive audience (employees), I don't see the ROI.
Of course I agree. But node is an unique case, since javascript is not exactly new, and had a trillion people using it when node appeared. You can not expect that to happen again any time soon, unless all main browsers start supporting client-side PHP or something crazy like that.
Go is not 'there' yet, sure, and maybe it'll never be. But its growth can not be judged by node's standards, no really new language could compete then.
For a new language is growing quite good: http://www.google.com/trends/explore?q=golang#q=golang&cmpt=...
Of course this doesn't mean you have to use it, but IMHO it can not be regarded as a failure in anyway.
Unless, of course, you've set a power profile to prioritize power efficiency, but that would be an absolutely ridiculous assumption given that we're talking about benchmarks.