Engin Deniz Cengiz

After my last blog post was about a super specific feature of Unreal Engine, I decided to go for a broader topic this time: Pragmatic tips for optimizing the CPU performance of a game, based on my optimization experience of the last years:

71

This year's Game Industry Conference I'll be talking about ECS in UE5, struggles of multiplatform programming and many more. See you there!

53

Every month, Epic Games chooses 20 items from their Marketplace to showcase where they have up to +47K items they have there. Today, on the Marketplace, Epic Games saw my product Optimization Debug Tool, and they chose to showcase my item after a short period of being there, in addition to that I'm giving you a discount of 30%! Maybe another good update will be that now there is a live Discord server for support. https://lnkd.in/d8duDKwB #unrealengine #unreal #unrealengine5 #unreal5 #optimization #debug #quality #marketplace #indiegamedev #indiedev Epic Games Unreal Engine

16

4 comentarios

The open source Unreal Engine Rest API plugin, VaRest (initially developed by Vladimir Alyamkin), was discontinued and delisted from Fab due to his personal reasons. I made contact with him, and with his consent, took over rights of further maintenance. The plugin is published again with Unreal Engine 5.5 version is now available, with Fab link: https://lnkd.in/dBjXVXhR Previous Unreal Engine versions will be added later. They're still available on GitHub, under this fork: https://lnkd.in/dUpC2Jgr

21

Hi, today I'm writing about fighting the fear of refactoring finished parts of the GDD. When working with game design documentation, it is important to remember that it is a continuous process of refactoring and support. The GDD can only be called final if the game has already been released and will not receive any more updates. Otherwise, reworking the documentation as decisions are changed is just as much work as writing it in the first place. In order for a valuable game designer not to get discouraged by such work, he should remember a simple basic rule, which is described by the principle of "clean documentation" DRY (do not repeat yourself). Any piece of documentation should not be repeated more than once. Reworking of already written documentation goes from the starting point of changes in the project through all affected related documents. When you use DRY, two positive things happen:  1. Duplicate documentation sections disappear 2. Logically related documents are actually cross-referenced. You can't repeat an idea, you have to link to it. Under these conditions, making changes to a document doesn't look scary. All you have to do is systematically go through the linked documents and make pinpoint changes. Moreover, this process ensures complete coverage. On the organizational side, you shouldn't be afraid of the refactoring process either, because it's perfectly fine to budget time for document maintenance. The problems start when you don't, because as part of the iteration, time starts to leak as you keep going back to what has already been written. This leads to stress, and the whole system of work starts to break down.  Remember:  * The GDD is a living structure. * If you don't keep the GDD in order, there will be problems. * Refactoring is a planned part of documentation work. * Link documents, and don't repeat yourself, so you can make changes quickly. Send this useful post to the game designers you know who have nightmares about refactoring documentation. 

18

8Bit Gamedev Salary Pulse 2024: Interestingly, game and level designers in lead/expert positions have an average salary of $4,717, which is lower than the average $6,253 for senior designers. 😱 This represents an 18% decrease in salary when advancing to higher levels. How is that possible? 🤔 This surprising trend raises questions about salary structures and career progression in the industry. What do you think could be causing this anomaly? #NumberOfTheDay

9

🎥 Catch All the Exciting Talks from the Graphics Programming Conference! 🎥 Missed the recent Graphics Programming Conference? No worries! You can now watch all the inspiring sessions on the official YouTube playlist: Graphics Programming Conference Talks via this YouTube Link: https://lnkd.in/dMm4b-r2 Highlights include: ✨ GPU Work Graphs: Explore groundbreaking GPU programming features with practical applications. ✨ Neural Networks in Rendering: Learn how machine learning integrates seamlessly into modern render loops. ✨ Ray Tracing in Game Engines: See how engines like Unreal Engine 5 and Snowdrop are pushing boundaries in visual fidelity. ✨ Non-Photorealistic Rendering: Discover innovative approaches to rendering stylised visuals in games like The Wolf Among Us. Whether you're into real-time graphics, ray tracing, volumetric rendering, or leveraging the power of WebGPU, there's something here for everyone passionate about graphics technology. 📅 Topics Covered: ✨ Advanced fluid simulations ✨ Large-world precision challenges ✨ GPU-optimised occlusion techniques ✨ Insights into tools used for Baldur's Gate 3 #GraphicsProgramming #GameDev #Rendering #MachineLearning #RayTracing #3DGraphics  #RealTimeRendering

16

This weekend got me head over heels with videogame programming... I joined the #MiniJam 159, first time participating in a jam as a programmer, while my colleague Florio Del Pino had the opportunity to test their game design skills for the first time in a jam context. The theme was "Ocean" so we made a clicker game where you hunt sea creatures and they look exactly like in the old maps! We decided to use the limiter "as small as possible" in several ways: *Scope limitations, something simple and doable in the two days we had to produce the game *Minimal game input (left click to attack, right click to dodge, that's it) *Claustrophobic aesthetic (you're going down an oppressive underwater tunnel) Most game assets were taken from the internet (credits on our itch.io page) to fully focus on the programming and game design. It's also my first time programming a combat system and, albeit a tad broken cause I haven't gotten to state machines yet, it's quite funky. Besides I could use the minimal tech art skills I learned in the horror game we're making at uni to boost this game's aesthetics. All in all, when it comes to programming, I'm beginning to believe. https://lnkd.in/ezRuVsSJ

11

2 comentarios

💡 [TIP] How do you handle unexpected news as an Engineering Manager? In day-to-day work, you and your teams sometimes receive unexpected news that can shake the team's motivation. As Engineering Manager, how are you handling this? Here is how I’m approaching these situations in case it helps you reinforce your own mindset: 1️⃣ Understand this reality: companies evolve. One day, your best teammate might leave. Another day, you might decide to move on. And on another day, the company might make changes. That’s just how it works. 2️⃣ So, I ask myself this question: What opportunities can my teams and I find in this situation? 3️⃣ If you're following step number 3️⃣ , I can ensure you that you'll find at least 1 opportunity to grow. 4️⃣ Communicate this to your teams and remain open to deeply listening to their thoughts and proposals. 5️⃣ Create your own plan to seize the opportunity. The ideal world doesn’t exist. And perhaps it doesn’t because “ideal” means something different to everyone. 🙂 This may sound obvious, but when we receive unexpected news, we tend to ask ourselves why and try to understand it, investing energy in something that probably isn’t up to us. This leads to frustration and exhaustion when it’s unnecessary. I choose to keep fighting to protect my #energy and #motivation. What about you? More content related to engineering management at https://lnkd.in/dNePvQJV #team #growth #engineeringmanager #manager #productivity

11

Junior developers, always want simple answers when there are none. You ask me: Vladimir, how many lines should a single component have? You want to hear: 173 I want to say, being a senior developer: it depends. I know it's annoying, so instead I do my best: As many as necessary, but not too many. Not much better. I'm trying: We have this 5000 line component. Now that's too much. But no blame, we had to ship that one real fast. Oh, and over there, that folder with 50 5-line components that all depend on each other, no good either. So, somewhere between 5 and 5000. Bear with me. Look, if we have simple code, like a config or something, we can have more of it, no problem. If some part of your code falls right out, let it fall. Cohesion. Coupling. Testability. Tree shaking. Lifecycle. I see you slowly go mad with information overload. I stop right there. Maybe I should have gone with "it depends". Maybe you should have asked "hey, I wrote this component, it has 150 lines, is it too big?" No, my little friend, relax, it's not too big.

41

15 comentarios

Hey, I wanted to share my little project – a multiplayer Arkanoid game built in Rust. I was in the mood to try something cool in this language and also see what HTTP/3 and MessagePack are capable of. Here’s the repo with the code -> https://lnkd.in/dV6EFF48 Overall, it came together pretty smoothly without any major issues – basically a mashup of ping-pong and Arkanoid. WebTransport (kind of like WebSocket, but using HTTP/3) works great. I didn’t run any serious benchmarks, but it only takes a few lines of code to set up. I used MessagePack to handle the game state data between the server and clients, and that was also a breeze with just a few lines, no extra hassle. I wouldn’t recommend copy-pasting my code into your own projects since it’s admittedly not the best quality, but it might come in handy as a reference for anyone looking to kick off a similar project in Rust and isn’t sure where to start. (Oh, and the lag is in the video, not the game itself. I didn’t bother recording it again.) #rust #rustlang #gamedev #multiplayergames #webtransport #messagepack #indiedev #opensource #coding #devcommunity

8

2 comentarios

🛠️ Unity Cancels the Runtime Fee – But is Trust Restored? 🛠️ Unity's decision to cancel the Runtime Fee, one year after its controversial introduction, shows they're listening to the community. But has this reversal rebuilt trust with developers? While the move is a positive step, trust takes time to rebuild. Developers need long-term transparency and stability to feel confident in their engine choice. Unity has the potential to regain its position as a go-to platform, but consistent engagement with the community will be crucial. Actions speak louder than words – and trust is earned over time. Let’s see what’s next! #Unity #GameDevelopment #Trust #Transparency #GamingIndustry

7

6 comentarios

Here's a #tip for #Unity game devs planning to take on something serious and commit to it. 1) It's going to take time. The codebase would become huge and you would forget your own code implementations. So make sure you follow some standard practices so that even if you do forget your own code the standard makes it easy to get back quickly. 2) Contain your systems with proper assemblies and if possible package them. Because it's never going to be perfect and you will continuously change your code. as such it might start to look like a spaghetti at some point. Instead of having one big pile of mess, make small piles of mess that way it's easily navigatable. it's a mess. But a small mess. And usually, we do change small sections at a time. 3) Data-oriented systems should be developed via the TDD process. No need to write tests for everything. Just the public methods should do the trick. This way even if you change the underlying implementations the tests would make sure nothing is breaking. Utilizing unity package would be a great way to ensure this. 4) Clean up your tech debts periodically. Don't let it grow too big. Every two months take a look at the entirety of the project and try to guess where things can go wrong. Make sure you comment todo if you do a hacky code somewhere. This would greatly help to reduce tech debts later.

3

I recently came across another C++ book. 750+ pages of useful tutorials on modern C++ programming. So if you are interested in learning C++ seriously, I recommend this book! https://lnkd.in/dcJvg84h

12

1 comentario

#gameengine #pathtracing Streaming RIS using weighted reservoir sampling of ReSTIR, seems already correct... this is not easy, hahaha SPP one, yes, only 1 sample per pixel! no any kind of reuse yet. 2000+ emissive triangles. by the way, the noise seems not correct, because I tried to use blue noise to replace white noise, and it's my first try, seems failed :P

2

𝐓𝐡𝐨𝐮𝐠𝐡𝐭𝐬 𝐨𝐧 𝐭𝐡𝐞 𝐆𝐚𝐦𝐞𝐍𝐆𝐞𝐧 𝐩𝐚𝐩𝐞𝐫 𝐮𝐬𝐢𝐧𝐠 𝐃𝐈𝐌𝐌 𝐃𝐢𝐟𝐟𝐮𝐬𝐢𝐨𝐧 Man, I love a RL/NLP/Diffusion paper about Video games. So I could not hold back from commenting on the GameNGen aka Doom paper that came out last week (https://lnkd.in/d4k5xfUN) kindy shared by my mates Joshua Chittick and Chris Benjaminsen. 🤔For those who have not read it yet, it's another way of emulating a DOOM gameplay, but this time, the heavylifting is done by a Diffusion model. More specifically, a 𝐃𝐃𝐈𝐌 (𝐃𝐞𝐧𝐨𝐢𝐬𝐢𝐧𝐠 𝐃𝐢𝐟𝐟𝐮𝐬𝐢𝐨𝐧 𝐈𝐦𝐩𝐥𝐢𝐜𝐢𝐭 𝐌𝐨𝐝𝐞𝐥𝐬). 🚩DDIM is just a variation of a Diffusion model, where the steps and the outcome are 𝘥𝘦𝘵𝘦𝘳𝘮𝘪𝘯𝘪𝘴𝘵𝘪𝘤, 𝘪𝘯𝘴𝘵𝘦𝘢𝘥 𝘰𝘧 𝘴𝘵𝘰𝘤𝘩𝘢𝘴𝘵𝘪𝘤. 🔡In this case, the text part of the t̳e̳x̳t̳-̳to-image model is removed and replaced by embeddings of the agent actions and latents of the frame history. Kind of an 𝐚𝐜𝐭𝐢𝐨𝐧+𝐟𝐫𝐚𝐦𝐞𝐬-𝐭𝐨-𝐢𝐦𝐚𝐠𝐞 The way it gathers data is interesting. 🚶🚶‍➡️It trains an 𝐑𝐞𝐢𝐧𝐟𝐨𝐫𝐜𝐞𝐦𝐞𝐧𝐭 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠 𝘢𝘨𝘦𝘯𝘵 𝘱𝘰𝘭𝘪𝘤𝘺 with the objective of collecting a big amount of observations, not necessarily to be a good AI player. 🧑‍🏫 The actions are guided by a 𝐭𝐞𝐚𝐜𝐡𝐞𝐫 𝐟𝐨𝐫𝐜𝐢𝐧𝐠 𝐨𝐛𝐣𝐞𝐜𝐭𝐢𝐯𝐞, that emulates what a human player would do. The gameplay collects episodes of 🖼️𝐟𝐫𝐚𝐦𝐞𝐬 and agent's 🦿𝐚𝐜𝐭𝐢𝐨𝐧𝐬 as the training of the Agent's policy runs. 1️⃣ It 𝘦𝘮𝘣𝘦𝘥𝘴 𝘵𝘩𝘦 𝘈𝘨𝘦𝘯𝘵 𝘢𝘤𝘵𝘪𝘰𝘯𝘴 𝘢𝘯𝘥 𝘶𝘴𝘦𝘴 𝘢 𝘤𝘳𝘰𝘴𝘴-𝘢𝘵𝘵𝘦𝘯𝘵𝘪𝘰𝘯 𝘮𝘦𝘤𝘩𝘢𝘯𝘪𝘴𝘮 as one of the inputs into the DDIM. 2️⃣ It 𝘦𝘮𝘣𝘦𝘥𝘴 𝘱𝘳𝘦𝘷𝘪𝘰𝘶𝘴 𝘢𝘯𝘥 𝘤𝘶𝘳𝘳𝘦𝘯𝘵 𝘧𝘳𝘢𝘮𝘦 𝘪𝘯𝘵𝘰 𝘭𝘢𝘵𝘦𝘯𝘵𝘴 (encoding the image pixels), as the other input. 🎨Finally, With those encodings, it uses the 𝐃𝐞𝐧𝐨𝐢𝐬𝐢𝐧𝐠 𝐍𝐞𝐭𝐰𝐨𝐫𝐤 (DDIM) to 𝘯𝘰𝘪𝘴𝘦 𝘢𝘯𝘥 𝘥𝘦𝘯𝘰𝘪𝘴𝘦 - create the next frame, in the Stable Diffusion way. It uses a 𝘥𝘪𝘧𝘧𝘶𝘴𝘪𝘰𝘯 𝘭𝘰𝘴𝘴 function to calculate the loss between what it predicted and what was the actual next frame. And voilà. Basically, what you've got is an interactive video player of DOOM.

11

3 comentarios

💥 Wow. What an inspiring webinar with Valentina (Cupać) Jemuović 🤖 on AI in TDD! A big thank 🙏 you to Valentina for sharing her experience on using Copilot for TDD. 👉 Key takeaways: 1- Creating failing tests: This is the task with the highest cognitive load in TDD, and it requires a lot of back-and-forth prompts to get the right tests. The generated tests often had quality issues, such as not challenging poor requirements or failing to express good interface design. 2- Writing code to make the test pass: Copilot performed ok but often needed multiple attempts. 3- Refactoring step: Copilot provided refactoring ideas but also required very specific guidance when performing challenging refactoring (which is why you need to be intentional about your coding practices - 🚀 go Packmind). Overall, Copilot falls short in the same areas as humans, which is not surprising. The 'red step' is particularly challenging for developers; it's where you decompose the requirements and drive the design. This is typically where Valentina spends the majority of her time coaching teams on TDD. Recording available in comment 👇 #TDD #AI #Copilot

41

11 comentarios

Did you know that in UE 5.5 you can do a full muscle simulation all inside UE, so without Houdini or Maya, and train an ML deformer directly from that? This is still experimental, but one step closer to making muscle based ML deformer easier for users. This is thanks to the awesome muscle simulation work by Yushan Han https://lnkd.in/en5Yj26S #mldeformer #ue5 #unrealengine

245

1 comentario

Players complain about laggy inputs and reactions in your online game. The good news 👇 Implementing the right network authority model can change that overnight! Few teams have devs with experience in making multiplayer games from the beginning... And laggy gameplay is one of the results. Every multiplayer games has different needs: 1. Depending on the genre 2. Depending on the pace 3. Depending on the usage of physics And many other factors. You'll get smooth, synchronized gameplay only if you'll pick the right authority model and input systems. This is a foundational choice: 🧱 The more you wait, the more expensive it becomes to switch later on. Sometimes the switch is so painful that teams decide to "release anyway", getting huge negative feedback on day 1, and throwing away thousands of $$$ of investment. 💸 Do you really want to take that risk? If not, fix your authority model today. You can spend a couple weeks of development time doing your own research, evaluating all the factors and authority models (client-driven, server-authoritative, etc...) Or you can ask for my help, and use that time to make your game instead 😉 >>>>>>>>>>>>>>>>>>>>>> P.S: I post about Game Development 5x/week. Follow me for more insights🔔 And check the comments for more tips! #JustDevIt #gamedevelopment #multiplayer #netcode #online  

9

2 comentarios