• Hey Guest. Check out your NeoGAF Wrapped 2025 results here!

Why are PC ports so bad? Digitaltrends asked some devs.

A game trying to use 16-25 GB VRAM is totally an issue of hardware flexibility. I mean how could they ever anticipate there being users with less than that.....
journey of course GIF
 
Last edited:
You wanna hear why they're bad? Its because these games are sloppy, engines balanced on a handful sticks, their systems and game logic glued together with duct and tape.

See Jedi Survivor, or even FFXVI. Their performance is terrible even on consoles. This isn't an issue of many PC configurations, but rather that variety of configurations exposing how poor their R&D is. Its merely easier to mask that with consoles since most players will brush a game running at sub 30 fps or stuttering being "just how it is".
 
This should surprise no one. PC porting eats up 80% of your QA budget and causes 90% of your support tickets, but only accounts for 40% of your sales. Most devs will budget QA based on expected revenue percentage from PC sales, and what you get is a shitty port. The good ports come from devs who know that you're going to have to over budget QA and dev bugfix time for the PC port, even if it means gross profit is lower than the console space.
 
The engines are put together by hope, dreams and sticks, they use bubblegum and dog shit to glue them. No one has a clue on how to optimize them because the people that knew left the industry.

And the publishers really dont give a fuck.
 
The one thing PC had in its favour is now what is hurting it.
Brute force.
Devs didn't optimise for PC, they just relied on more raster to get better results. Now there are more cards than ever, and alot of them not much more powerful than a PS5, they arnt getting enough optimisation.

The bigger issue for me as an Xbox gamer is that it looks like the Xbox is becoming just another PC configuration.
 
Most issues seem to be cpu/memory related and have little to do with having different hardware configurations. If your engine is hopelessly single-threaded, you can't really blame it on anything else.
 
Most issues seem to be cpu/memory related and have little to do with having different hardware configurations. If your engine is hopelessly single-threaded, you can't really blame it on anything else.
There is only a few really CPU optimised game engines around on PC.
Idtech
Forzatech
IW7
???
 
The one thing PC had in its favour is now what is hurting it.
Brute force.
Devs didn't optimise for PC, they just relied on more raster to get better results. Now there are more cards than ever, and alot of them not much more powerful than a PS5, they arnt getting enough optimisation.

The bigger issue for me as an Xbox gamer is that it looks like the Xbox is becoming just another PC configuration.
And honestly it's basically a PC, to say that they don't have an x86 baseline is legit laughable lol.


Also, apparently everyone under devolver digital can make a PC port that scales amazingly well but EA can't, I think it speaks for itself.
 
Last edited by a moderator:
That's what happens when you run a gig economy with devs. Everyone has to relearn the shit spaghetti code at every new job, then add in some more band aid spaghetti code, and rinse and repeat.

No wonder why everything is broken.
 
Last edited:
Basically there's too many different configurations of PCs for them to do a good job of them all.
That's the old line, and of course it is a challenge, but it isn't really the reason PC ports are bad as of late, because the games we're talking about are poorly optimized on every configuration.

I think the issue is really that console architecture has evolved to where it's PC like in a lot of good ways but also has some major advantages and porting from console to PC is a step backwards in certain ways (like not having unified RAM, for example), and devs don't want to put in the additional effort to optimize.
 
16gb is alot of freedom and too good to notice how good or bad the stuff they are working on, alot of freedom doesn't tell how good or talented they are, it hides alot of personal weaknesses, that's just how it is.
 
Two memory pools on PC against unified memory on Ps5 and Series SX.

Games are developed first for consoles and its unified memory. Then, these games are ported to PC with a different memory configuration and all the perfomance and optimizations the ganes had get obliterated when ported directly to PC.

PC´s Architecture is getting old and obsolete.

Also, there is no hardware card like Kraken on Pc getting through all data decompressión fron the NVme.
 
Last edited:
Not a great article. It's basically saying that there is a configuration that can run some games great which is just factually incorrect. Problems that arise in one game happens on all hardware.

It's a API problem (and thus a utilization problem), that's pretty well known by now, but I guess they don't want to burn bridges?
 
Last edited:
Two memory pools on PC against unified memory on Ps5 and Series SX.

Games are developed first for consoles and its unified memory. Then, these games are ported to PC with a different memory configuration and all the perfomance and optimizations the ganes had get obliterated when ported directly to PC.

PC´s Architecture is getting old and obsolete.

Also, there is no hardware card like Kraken on Pc getting through all data decompressión fron the NVme.
Yes and no. Unified pools is not a magic bullet. Unified memory was already a thing since the PS360 era, and PC's blew past that by a mile. It's a utilization problem. PC hardware is already leagues more powerful (especially on the CPU GPU front). But it's the underlying API's and low-level approach that creates problems for devs (that are solvable by the API developers).

'Direct-storage' should be a thing quicker though. It directly gives the GPU access to assets without the CPU's involvement.
 
Last edited:
if you are a game dev, would you prefer to optimizes a game console with the same spec and architecture, or a gaming PC with more than 10 different configurations?
 
The bigger issue for me as an Xbox gamer is that it looks like the Xbox is becoming just another PC configuration.
Not quite as devs can optimise ray tracing far better (more customisation options, the HW has more console like optimisations than PC's do) and engines do not deal with shader compilation at runtime, they can pre build them, but otherwise yeah well that is the point of moving from XDK to GDK.
 
Not a great article. It's basically saying that there is a configuration that can run some games great which is just factually incorrect. Problems that arise in one game happens on all hardware.

It's a API problem (and thus a utilization problem), that's pretty well known by now, but I guess they don't want to burn bridges?
Please elaborate more this API problem. Because if some games work fine, the problem is not the same API but how e who utilize it
 
Not a great article. It's basically saying that there is a configuration that can run some games great which is just factually incorrect. Problems that arise in one game happens on all hardware.

It's a API problem (and thus a utilization problem), that's pretty well known by now, but I guess they don't want to burn bridges?
It is also a complexity mix that burdens you through development and support: different Windows versions, different HW setups, different driver mixes… the current issue with shader compilations at install time or runtime (again as the HW and SW might change after the initial install) is one but not the only one.

Not saying it is not also an API problem, but I am not sure there is a magic bullet: consoles and PC's have different approaches and different pros and cons.
 
In the end it all boils down to developer deadlines. Sure some are better that others in optimizing but I can't really expect all of them to perform well. It's a double edged sword in a way. Deadlines make the game not go out of scope but if the scope is already large to begin with, a longer dev time is required.
 
Please elaborate more this API problem. Because if some games work fine, the problem is not the same API but how e who utilize it
Quick and dirty answer: it mainly because of API's like DX12 and Vulkan, these are great for performance reasons over older API's like DX11, but the low-level access granted by DX12 makes things so devs have more freedom to get the best performance out of what they're making, but also it's way more complex for devs now that there are less safe guards. I'll explain:

The core problem of modern video games has mostly to do with memory management and Multicore CPU's and how to utilize them in runtime code. Previous API did have these safeguards baked in, but ate inherently limited. DX9 up to 11 had problems utilizing more than 4 threads, because game-logic running on more threads is incredibly difficult because of deterministic dependencies. That's why every game uses one main thread and separate jobs over multiple threads and brings them back to the main thread to sync them up, but separating these is tricky and per game different. And separating them badly causes stalling which slows down performance.

That's where Vulkan and DX12 come in and give the opportunity for devs to have more low-level access to the API layers and customize it proper for their game. But this created issues for devs that aren't well equipped for this low-level access. So that's why DX12 is a blessing and a curse. DX12 (and Vulkan) can be leagues more powerful than DX11 (or consoles) in the right hands.
 
Last edited:
Sure, but ultimately you succeed by making more devs more effective.
If one is successful everyone one could be, the API is the same for everyone(and there are even easier alternatives), if something is fixed (API) than the problem are the variables (dev and/or engine and/or money)
 
Quick and dirty answer: it mainly because of API's like DX12 and Vulkan, these are great for performance reasons over older API's like DX11, but the low-level access granted by DX12 makes things so devs have more freedom to get the best performance out of what they're making, but also it's way more complex for devs now that there are less safe guards. I'll explain: the core problem of modern video games has mostly to do with memory management and Multicore CPU's and how to utilize them in runtime code. DX9 up to 11 had problems utilizing more than 4 threads, because game-logic running on more threads is incredibly difficult because of deterministic dependencies. That's why every game uses one main thread and separate jobs over multiple threads and brings them back to the main thread to sync them up, but separating these is tricky and per game different. And separating them badly causes stalling which slows down performance.

That's where Vulkan and DX12 come in and give the opportunity for devs to have more low-level access to the API layers and customize it proper for their game. But this created issues for devs that aren't well equipped for this low-level access. So that's why DX12 is a blessing and a curse. DX12 (and Vulkan) can be leagues more powerful than DX11 in the right hands
I agree with that, but as you wrote the problem is not the API but something else. Devs got more freedom and power but many can't fully utilise it. Just use dx11 and lose some potential performance but with a proper game running without issues and not by forcing people to use overpowered configurations to have something barely playable
 
I agree with that, but as you wrote the problem is not the API but something else. Devs got more freedom and power but many can't fully utilise it. Just use dx11 and lose some potential performance but with a proper game running without issues and not by forcing people to use overpowered configurations to have something barely playable
I get ya, but it's not that simple. The API's from these latest consoles are equivalents of the DX12 architecture (GNM I believe it's called for PS4 and 5)

So downgrading your code for an older obsolete API can cause even more problems when taking newer (or upcoming) hardware into account. There are translation layers but that's not very efficient. Some games that are highly deterministic/physics based (TEARDOWN, Amnesia The Bunker) are even still based on OpenGL, these have big utilization problems, like Teardown only really functioning on 1 CPU core. But because OpenGL is spec free and open sourced, it's easier to write a translation layers or to write a port for different hardware.

Lot of technical talk, and it way more complex/nuanced than I just described. But the point is, DX11 has it's own problems that may or may not work with you specific underlying game-logic.
 
Last edited:
Because they think they can get away with it financially.
more like its because they are getting away with it financially.
they are reviewed just as well.
and people are jumping from one broken hyped release to another by millions.
(and then waiting patches for months on forums)
 
more like its because they are getting away with it financially.
they are reviewed just as well.
and people are jumping from one broken hyped release to another by millions.
(and then waiting patches for months on forums)

Well some are and some are not... depends on if its bad enough to put people off...

People need to stop pre-ordering.

 
Just look at game sales on PC compared to console. Publishers make more money from console sales than PC so they don't prioritize PC ports the way they should.

PC ports are treated like a quick cash grab most of the time.
 
It's just that consoles and PC are now close enough in term of architecture that they can just make a PC port almost freely, and then try to get away with it to maximize their profits.
Considering a lot of PC ports are released utterly broken, and then are magically fixed a month later after the mediatic backlash, there are no technical excuses to it.
 
Last edited:
Top Bottom