Oolala

Member
  • Content count

    124
  • Joined

  • Last visited

Community Reputation

1102 Excellent

About Oolala

  • Rank
    Member
  1. Gpu Stress

    Can we just skip the question for a moment, and assume you have whatever data you'd get from this process.  What knowledge do you hope to extract from this data, and why would you be confident that this knowledge would be inferred by that data?
  2. Gpu Stress

    You're not going to find such a thing because modern GPUs aren't architecturally segregated based on "pipeline stage".  As such, there is no possible answer to your question.  Architecturally, GPUs are just banks of processors, and new "pipeline stages" are added by extending new architectural capability to all these processors.   With that said, there are well understood GPU benchmarking suites, such as GPUBench, that are widely used in research, and are designed to stress different architectural elements.
  3. Wyoming's Eyeball Smuggler

    Oh you adorable naive person.  I just want to hug you and tell you that you've reached the bottom of the barrel with regard to human behavior and that eyeballs up the ass are as bad as reality gets.
  4. Safety vs Efficiency

    The general rule I go by is to assert on all exposed mutable state that may be left in a unexpected form by some combination of method calls.  These assertions are usually clumped up at the top of functions, or sometimes top of loop bodies, in a big brick.  Even if the assertion never gets tripped, it's still a nice thing to have there for the purpose of establishing a set of assumptions that are made during the authorship of the function body, and serve a a functional form of documentation.  In languages that feature code contracts, I replace those assert-bricks with contracts instead.  There is actual english-style documentation as well, but the assertions are for added specificity   In all, somewhere in the range of 20-25% of function-body lines of code are assertions on average.  Often times the assertions are in place before the useful part of the  function body is even written.   I can tell you though, spending the time bloating my code by 20-25% takes a bit of extra time, but that time is recouped many fold over by saved time in debugging.
  5. Projects to learn things.

    I totally know what you're talking about.  If a deadline is far enough into the future, I still find though that being able to learn with a small sample project helps me improve my design approach on a larger project that needs whatever the new technology is.   Maybe it's just me, but it gives me an opportunity to get the "oh man I'm just learning this, and boy is it sloppy" trial-run out of the way on something that matters less.  If deadlines are near-term though, or the project is short term, I totally agree with your approach.
  6. Projects to learn things.

    I suspect I'm not the only one, but often times when I'm trying to learn something new, it helps to have a project to act as a driver, and to establish a goal.  I was wondering if others would like to share some of their "made this for the purpose of learning that" kind of projects.   For example....   - When I wanted to learn how to deal with Java's "swing" GUI system, I made a multi-window version of "Qwirkle", which is a tile-laying table top game kind of similar in flavor to dominoes.  The game is super simple, but it's actually kinda a fun little thing, and is very GUI-heavy and logic-light.   - When first learning about neural networks, I made a top-down tank "game" (which in reality was more of a screen saver, since all characters were computer-controlled by default) that gave me an interesting avenue to tinker with parameters.   - When learning the windows FS driver systems, I made a C# program that allows for software-intercept of file system activity via plug-ins (which is actually still something I use).  I still use a matured version of this software to do stuff like redirecting file IO writes of legacy software result files to standard in of results parsing software, without having to write to disk first.   What cool stuff have you made for the purpose of learning something?
  7. Honestly it seems a bit pointless to target JUST flash, when there are so many other problem programs out there.  I agree that flash needs to go, but flash needs to be the first of many rejected technologies if the intention is to build secure systems.  Been arguing for years with people that even things like memory mapped IO device drivers need to go, and any program deployed as binary needs to go, and "secure" programs need to all be formally validated vs code contracts, and '"unsecure" programs need to be sand-boxed to the hilt, and and and.   Burning adobe to the ground is a good first effort, but it shouldn't be the last effort.
  8. Basic wind simulation in a large building

    Calculate it off-line.  This is what I did for snow fall in a previous project.  When you're done with your entire map, go the fluid simulation route, and use that to generate cyclic "wind swing" loops, like a couple seconds long.  Bake them into your map, and you get the results of a highly detailed simulation with 0 CPU cost.  The other nice thing is that this can also give you a lot of options on how to tweak things to actually get the look you want, without having to have your artists go in and fix stuff.  Store the forces instead of positions, so you can sum forces that may not be considered offline.
  9. This question is going to be half curiousity, and half start of a discussion.   I've always found it interesting that PC games often go full-screen, and immediately present GUI-like features that are closely similar to the OS that just got hidden away.  Things like buttons, drop-down windows, check boxes, even sub-windows.  Why do game authors do this?  Everyone starts with a heavily tested, optimized, and feature rich GUI suite, and use it primarily to hide the original GUI and implement a new one, that is less tested, has fewer features, and is generall all-around shittier and less stable.   Springing off of that, I was hoping to have a discussion about embracing the native GUI, and actually using windows.  Windows for inventory screens, windows for the main game screen, maybe windows for dialog with characters, windows for the minimap.  You have this super strong tool that allows the player to customize their experience (draggable windows, minimize buttons, close buttons, UI alerts, etc), why not take it?   So, as game designers, have you considered dropping the full-screen paradigm?  Are you considering it now?  What's some cool user-experience stuff you can do easily with native GUI's that are otherwise painful?
  10. An Open World Idea

    I always find it interesting the lack of appreciation that so many people have for how truly huge the physical world is.  Like if you randomly distributed all the WOW players onto a to-scale USA, on foot, and gave them an hour to find even a single other "player", nearly all of them would fail.  Even randomly distributing them around a big city, like chicago or los angeles, and give them an hour to find other "players", chances are most of them would find just a few.  Chances are many of them wouldn't even find their way out of the building they started in.   Things are set in small towns so that players actually get a chance to interact with each other.  Besides just a new map for Day Z, to borrow your comparison, what would this project creatively contribute?  If your answer is "nothing", then just make a mod to Day Z.  Really though, this just sounds like heaps of busy-work for artists.
  11. Best loading screen.

    Loading screens are pretty much a necessary evil.  At the very least, when the game is booting up for the very first time.  Some do it better than others though.  I was wondering what are some of the projects that you think really did loading screens right, and what you think makes a kick ass loading screen?   Two of my personal favorites, are "world of goo", which took the opportunity to tell jokes, and "hotline miami", which made loading screens so psychadelic and colorful that they just didn't feel painful in the slightest.   So, what are your favs, and what do you think makes an awesome loading screen?
  12. DRM, piracy and more

    The money I pay for games these days no longer means as much as the time I would otherwise spend searching for them & playing them.  I can remember back when I was in high school, 40$ meant a whole lot.  I still didn't pirate very often, but mostly because I was paranoid about getting caught and/or catching viruses.   These days, steam is super convenient, and I can spend 40$ on a game without missing it.  Pirating a game takes a lot of time, and carries the risk of taking even more time & effort if it goes badly.  The 40$ just isn't worth the hassle.  Honestly, games targeted at older demographics could probably get away with charging a whole lot more than they do now, for exactly this reason.  I'm sure everyone has a cut-off where the effort to pirate outweighs the cost of going the legit route, but chances are for professional adults, that cut off is likely way over 40$.
  13. Drawing a city in 3D

    A whole lot of the answer to this has to do with what "the town" means.  Are you authoring an actual town, or rendering a hypothetical town that appears period correct?   It's a whole lot easier to describe to a program how to author a town than it is to author a town.  Authoring a building by hand is easy, but the computer is better at doing things as stuff scales up.  Render bits and pieces, like period correct doors and windows, and walls, and what a road looked like in different periods, and maybe even how roads are laid out in different periods.  Describe to the machine how to snap all the little pieces together, and let it rip.
  14. Voxel disk operations?

    How large are these "collection of 2d voxels" (images)?  Depending on your platform, memory might be super massive compared to your data structures.  You may very well be trying to solve a non-problem.
  15. If it's a single-player game, how about a button that just fast-forwards the game simulation to the next meaningful event?  If nothing is happening, then just skip ahead to when something is happening.