233 reputation
15
bio website jasperbekkers.nl
location Netherlands
age 26
visits member for 4 years, 2 months
seen May 29 at 20:35

Currently I'm working at EA/DICE in Stockholm as part of the Frostbite team.

Previously I was enrolled in the Internation Game Architecture and Design (IGAD) program at NHTV university in the Netherlands.

I've done an internship at Nixxes software, http://www.nixxes.com where I've worked on the video game Deus Ex: Human Revolution where I mainly did work on the Xbox 360, PS3 and PC.

I've done an internship at DICE, http://www.dice.se where my main focus was extending and optimizing Xbox 360, PS3 & PC code for Battlefield 3.


Feb
23
comment Does Big O really matter?
@MaulingMonkey Right, and on the IA-32 box they test on, the SIMD results and the bin results are pretty similar. On PPC these results would be pretty similar, on SPU they would not (due to more difficult to predict out-of-core access patterns that won't hide latency as well as a linear search would). Besides as you can see in the article you've quoted, binary search falls flat as soon as the vector size > 32 and the rest of the times its on-par with the SIMD implementation.
Feb
22
comment Does Big O really matter?
@MaulingMonkey No. I think you're pretty much only looking at IA-32 code that doesn't use SIMD. There are a lot of other game platforms out there ;-)
Feb
21
comment Does Big O really matter?
@MaulingMonkey that was exactly what you were doing with that link though. They proved it's possible conceive a scenario where it's possible in a microbenchmark. I've seen linear searches outperform sublinear searches on multiple platforms due to less pointer chasing and better locality. And the obvious lower algorithmic constants.
Feb
21
comment Does Big O really matter?
@MaulingMonkey Except, not all data is in the L1 cache all the time. There is data to be fetched from RAM too and guess what the prefetchers are better at? Anyway, measuring is always the best option :-)
Dec
3
awarded  Nice Answer
Dec
3
awarded  Yearling
Sep
2
revised Why are MVC & TDD not employed more in game architecture?
added 158 characters in body
Sep
2
awarded  Editor
Sep
2
revised Why are MVC & TDD not employed more in game architecture?
added 308 characters in body
Sep
2
comment Why are MVC & TDD not employed more in game architecture?
Joe is correct, most bugs in a video game are not algorithmic bugs; they are asset, or gameplay related bugs that you can only really test for with excessive Q&A and playtesting. Nobody (except for the graphics programmer) is going to be particularly bothered by the pixelated shadow edges, but people are going to be frustrated when the evil monster at the end of level 3 is unbeatable because it's not vulnerable to the magic spell you collected two levels earlier and there isn't another way to defeat it.
Sep
1
awarded  Teacher
Sep
1
answered Why are MVC & TDD not employed more in game architecture?
Aug
31
comment Does Big O really matter?
David, do you have any real-life examples? I mean, I can also Big-O the number of apples a group of people can carry, but that doesn't mean that it's used or useful. From my experience, most of the time gamedev choose their (parallel) algorithms based on raw performance, not on their growth functions.
Aug
31
answered Does Big O really matter?
Aug
17
awarded  Autobiographer
Aug
17
awarded  Supporter