Category: programming

Have you ever bought a game based upon the minimum reqs, and found out it doesn’t run? or it’s a slideshow? have you ever pheared the min reqs, but bought it anyway, and be surprised how well it runs?

Here is the inside track:

WE  (the devs) HAVE NO IDEA. WE MAKE IT UP.

Now obviously as an indie dev, I have less ability than Actiblizzard to tell you if the game will run on a 1 Gig RAM, 64 MB video card rig, because quite frankly, I don’t own one. I currently own 3 working PC’s, 1 Sony vaio laptop, 1 ultracheap dell netbook, and my main Mesh uber-dev PC.  That’s it.

Now when I release a game, I check it on all 3, then on my mates PC, my brothers PC, and a few other peoples PC’s. (other indie devs mainly). Then I take a rough guess, and we call it the min req. You might be thinking that this means it’s just me making it up, and that big publishers have 100 different PC’s and they can be scientific about it. I call bullshit.

When people ask me if GSB will run on their PC (A very understandable question right now, as there is no demo), I give as honest an answer as I can, and here it is:

I’ve heard some surprisingly good things about GSB performance on MACs running emulators, even under WINE, and on very low-spec PCs, but I honestly have no real idea of min spec. So here, as a service to people on the fence considering the pre-order, if you have bought and played GSB, feel free to post your specs, and the games performance in the comments. (Note that its frame-locked to 60FPS so you won’t see it go faster. It uses any spare ‘headroom’ to do a few fancier effects if it has time)

Patching a game people are playing is a nightmare. On the one hand, you have all these people experiencing problems with the game, and you desperately want to fix their problems.

On the other hand, is the fear that changes to the code may actually change something somewhere else and cause problems for people who are currently playing the game fine. Everyone has a different setup, with different hardware, and plays differently, so its’ very hard to have an absolute test case when you know you fixed bug A without introducing bug B.

Worse, is the nightmare that without a huge room full of QA experts like the big retail studios, you can either do some basic tests and release a new patch, or wait 3 days while you methodically test the new patch yourself. Neither option is easy, or popular, and whatever you choose you uspet some people.

I know some people have performance issues with the new improved 1.03, the one that fixed the sound issues (AFAIK). So today, rather than working on fun gameplay stuff (area of effect damage) and AI orders, I’ll be profiling the sound system code.

Balls.

It’s amazing what bugs you find when you just sit and play through the game for hours (I’m playtesting and balancing the game so that the AI fleets give you a bit of a challenge).

Two bugs squashed today are these:

Firstly, the code that detected that an enemy target was being ‘painted’ by a target painting laser (and thus easier to hit for missiles) was just not working at all. I’d checked and double checked it many times, but because the ‘is it painted’ code is just one of about 6 factors determining optimum target selection, it was hard to tell if it was working right without stepping through everything each time.

It turns out that code worked fine, but some *other* code which told a targeted module whether or not it even *was* a missile launcher, was buggy, and so the code was never even being run. doh!

The second bug was investigated after I noticed some very slow missiles which looked like multiple-warhead missiles but never split into their sub-munitions. It turns out that this code has never worked correctly! The code works fine the first time a missile splits into sub-munitions, but if it gets to live long enough to shoot another missile later, the sub-munitions are never triggered. That was an easy fix.

Unfortunately, that means I’ve been underestimating the destructive power of multiple-warhead missiles all the time, which means I need to re-play through every battle, on every difficulty, and both re-balance the fleets, and maybe re-balance the cost and strength of the appropriate missile modules too. This will take all of today at least.

Damn :(

One thing I did notice, that was pretty cool was an imperial AI fleet I’d put together that sends its fighters in ahead, and the rest of the fleet shows up afterwards (moving slower). By lucky hap, although the fighters were useless against the heavily armoured federation cruisers, they did manage to blow the crap out of the row of point-defense equipped escort frigates that flew alongside them. So when the Imperial cruisers got into range, they could bombard the feds with missiles without opposition. This is exactly the kind of design and thinking I hoped the game would reward, so it’s cool to see it is all about strategy and not just luck :D (Obviously when you play against that fleet, make sure you don’t have a line of unarmed frigates sitting waiting to be blown up :D)

I’ve never seen anyone share data like this, but I don’t see why they wouldn’t, it’s hardly source code. PIX is a free program from Microsoft that lets you analyze your games directx performance. I haven’t analyzed this data or changed any code yet, but here is a quick 2 minute start of a battle in the game (release build) analyzed with pix. Click the image for a larger version

This is great for working out why you get occasional slow frames, or working out where bottlenecks are. TYhe FPS seems to never go above 60 FPS, because the game caps it there, making a note of the ‘headroom’ available for potentially enabling higher-demand effect (not all done yet).

This is under Vista, 2 gig RAM, Core 2 Duo 6600 2.4 gig CPU and a Geforce 8800 GTS with latest drivers.

Please feel free to comment on how the numebr compare to agmes you might have worked on or develop now. I’d lvoe to see other peoples pix data as a sanity check. Is 83 set textures a frame a lot?  it seems to depend on what else is going on.