45
Just an observation on game engines
(lemm.ee)
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 23 Oct 2023
45 points (72.3% liked)
Games
30482 readers
1290 users here now
Welcome to the largest gaming community on Lemmy! Discussion for all kinds of games. Video games, tabletop games, card games etc.
Weekly Threads:
Rules:
-
Submissions have to be related to games
-
No bigotry or harassment, be civil
-
No excessive self-promotion
-
Stay on-topic; no memes, funny videos, giveaways, reposts, or low-effort posts
-
Mark Spoilers and NSFW
-
No linking to piracy
More information about the community rules can be found here.
founded 1 year ago
MODERATORS
The problem is that hardware has come a long way and is now much harder to understand.
Back in the old days you had consoles with custom MIPS processors, usually augmented with special vector ops and that was it. No out-of-order memory access, no DMA management, no GPU offloading etc.
These days, you have all of that on x86 plus branch predictors, complex cache architecture with various on-chip interconnects, etc... It's gotten so bad that most CS undergrad degrees only teach a simplified subset of actual computer architecture. How many people actually write optimized inline assembly these days? You need to be a crazy hacker to pull off what game devs in the 80-90s used to do. And crazy hackers aren't in the game industry anymore, they get paid way better working on high performance simulation software/networking/embedded programming.
Are there still old fashioned hackers that make games? Yes, but you'll want to look into the modding scene. People have been modifying the Java bytecode /MS cli for ages for compiled functions. A lot of which is extremely technically impressive (i.e. splicing a function in realtime). It's just that none of these devs who can do this wants to do this for a living with AAA titles. Instead, they're doing it as a hobby with modding instead.
Re-using engines has been around for basically as long as game development has existed. This idea of some mythical age when game development was more "pure" is a fantasy. What has changed is that expectations on AAA titles has grown to the point where it's extremely difficult to roll your own engine if you are committed to many, many years of work.
Not to mention, it certainly doesn't guarantee that the engine performs well. Look at Starfield or Baldur's Gate 3. Both have noticeable issues with performance, and both are built on in-house engines by their respective studios.
Yeah, this guy is basically harping on the concept of re-usable code. That's why we praise RollerCoaster Tycoon's dev, he wrote the entire thing in assembly. Beyond that, everything since 2d has used an engine. Hell, to not use an engine would be wasteful and delay games. What, every game should rewrite an engine?
Even Halo CE, 2002, used an engine. The Blam! engine. Dude's delusional if he thinks people were drawing individual pixels on the monitor.
It's ironic that we always seem to praise RollerCoaster Tycoon specifically, as that's one's based on the Transport Tycoon engine, which was also by Josh Sawyer and also in x86 assembly.
Nah this isn't true. If you gave the devs a free month I'm sure they could optimize the hell out of things. The issue is there are deadlines and higher priority items. You can technically play cities 2 unoptimized at a lower fps and graphics setting, you'll have a much worse time playing it if features are incomplete and full of bugs.
They simply didn't have the time to get to optimization
That's quite true. A friend told me that the red engine from cdpr is one of the most efficient and we'll made engine today. It didn't surprised me because cdpr has been working on it for 2 years after cyberpunk failed to meet the technical requirements they sold it for.
That's what it takes now for an optimized engine: 2 years on a decade old engine.