|
Slacker007 wrote: I know what you are doing here
I dont know what you mean! 
|
|
|
|
|
so i rolled my own B+ tree and it performed pretty well.
but then i ran Microsoft's KwData B+ tree side by side and it did slightly better.
i thought, that's weird, Microsoft's is usually a dog.
Well, it turns out, the SortedDictionary was causing so much GC churn that the b+tree ended up paying for it down the line.
Take the other dictionaries out of the test scenario and it speeds up nicely.
This is why I hate perf testing in .NET. The GC wildcard.
And I can't suspend the GC *and* hold a million items in memory, so that's out.
*headdesk*
a day of work for nothing, or at least, for nothing other than learning the GC was bombing my test numbers.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Have you tried calling GC.Collect() at strategic times, for better performance analysis, when memory is an issue....
But if memory muddle the issue, maybe the same thing will happen in the real world?
|
|
|
|
|
if in the real world people are filling several dictionaries with millions of items then what they really need is a database.
I've considered calling collect, but i first need to refactor so that it tests each dictionary individually and then collates the results (so the search times for each appear next to each other for example)
i just haven't done it yet, but it's on a long list of todos
what gets me right now, is the variation in performance run to run for that particular class.
sometimes it does great. sometimes it does poorly, under the exact same conditions and it has left me scratching my head, because I don't think the GC can account for all of that.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
I dunno how is your test.. but if you are using the unit test runner and look at the output, it could be multiple thread runs in parallel, adding a little bit of async randomness to the results?
But if you do (likely) your performance test in a console app, synchronously, then I dunno...
I usually get consistent perf results once I remove multithreading and memory issue.
Mm.. are you using a random number generator, for random data, per chance?
|
|
|
|
|
no, it's just plain old static functions .
i'm not doing perf with nunit or anything like that.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
are you using random number generator / randomly generated data?
|
|
|
|
|
I do both consecutive (optionally "feathered" to give the trees a better time) and random sequence generation using the internal random number generator which i incessantly reseed.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Mmmm.... 
Good luck with that, seems... mysteriously unfortunate! 
|
|
|
|
|
I'll figure it out. Or I'll get bored with it. It works as is in any case. This is more just picking it apart.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
|
I really don't think NUnit is magic.
It can't make the GC not be stressed when you're holding 50 million items across 5 dictionary classes.
anyway, I sorted it out. I just haven't implemented the solution yet.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
i am here to be disruptive, disgusting, awful and awesome. i am beautiful - hideous and perfect.
i am given over to madness, and the only truth teller in the only existence left.
the cleansed, and the joyfully dirty. the chaste w**re, or the malakoi priest - like the monster Jesus, but not like the man, Paul.
and all of these (but not Paul, f*ck Paul - i love Paul) are synonyms, in truth.
i speaks my truths as tools, to change my world, which done correctly
changes yours.
i love this chaos. but it hurts.
a clever monster learns how to enjoy hurt a little bit.
not so much that we eat it all the time.
like everything it's kata - not stance.
i'm a monster.
i'm here to dance in the shadows of the fires of rome.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Hiya, Kali, how ya' doin'?
|
|
|
|
|
While the poetry is beautiful, I think you are over exaggerating it! 
I'm just a random guy!
|
|
|
|
|
it's not so much a poem as it is a random dispatch from my particular flavor of crazy.
i essay sometimes when i need to degauss my brain.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Congratulations on revealing your true purpose here: to disrupt.
No, you are not a monster. Monsters exhibit uniqueness, originality: they terrify.
Your spate of drooling on the Lounge is boring, unoriginal: to such spew of personal aggrandizement one can only apply the kaopectate of indifference.
You are just a pathetic, insecure, troll, driven by garden variety narcissism, and pretentiousness. I consider it possible that many elements of your "life story," as trumpeted here, are fabrications, and I would not rule out the possibility of drug use.
On the other hand, who knows, some of your code might be useful, if it was ever finished, if it worked.
«One day it will have to be officially admitted that what we have christened reality is an even greater illusion than the world of dreams.» Salvador Dali
|
|
|
|
|
Well, Bill, it could be worse.
I could be you.
Have a great night.
(reported for abuse/trolling)
this is the first time I've had to report you, Bill. I don't like doing it, but at least you finally said what you've been dying to say. 
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Geez Bill, take a chill pill!
|
|
|
|
|
I knew you were a monster the first time I saw... Ah well, you know what I'm going to say 
|
|
|
|
|
|
umm wow.
I really want to dub "It's Raining Men" over video of that for some reason.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Thankfully, one is in Russia and the other in Australia!!
__________________
Lord, grant me the serenity to accept that there are some things I just can’t keep up with, the determination to keep up with the things I must keep up with, and the wisdom to find a good RSS feed from someone who keeps up with what I’d like to, but just don’t have the damn bandwidth to handle right now.
© 2009, Rex Hammock
|
|
|
|
|
Humble is offering CRYING SUNS DEMO for free. Not sure how much "demo" is going to be though
M.D.V.
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
I'd certainly hope the demo would be free. Haven't checked the site to see what actually is going but please don't give them any ideas! "Try our demo! It's only $19.99 and if you enjoy it you can buy the full game for $59.99!" ... "What? No of course you don't get a discount if you buy the demo first! Do you think this is a charity or something?"
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
