Sport developer created a scene of his dogs with 10 billion polygons in Unreal Engine 5


What right took station? Unreal Engine 5 early salvage admission to right launched and builders are already testing its limits, or on the least making an try to search out them. Soon after launch, a sport developer loaded a 10 million polygon dogs into the engine and replicated it 1,000 cases, ensuing in a 10 billion polygon scene that the engine had no topic going thru.

To construct this experiment, Gary Freeman, lead developer of Ionized Games, began by taking 122 pictures of his dogs sound asleep on a mattress. Then, he threw the total pictures into Actuality Obtain, created 3D units out of them, and exported the generated property with a 4K texture into Unreal Engine 5.

As soon as the 10 million polygon mannequin was loaded into Unreal Engine 5, the developer historic Nanite meshes to construct 1000 cases of the mannequin, making a 10 billion polygon scene. As seen within the video, the engine did no longer fight, working the scene at 60FPS at all cases.

I was in a position to load a 10 million polygon photoscan of Ziggy within the @UnrealEngine. The use of #UE5‘s Nanite meshes I was in a position to load 1000 cases of it at 60fps earlier than I purchased bored. That is 10 billion polygons and it did not even blink. It would possibly perchance perchance possess handled loads more than this.

— Sgt. Gary Freeman of the UNC (@IonizedGames) Would possibly just 26, 2021

It be fee noting that Gary’s machine isn’t always genuinely exactly a toaster. He detailed his machine contains a Ryzen 7 5800X, an EVGA RTX 3070 XC, and 32GB of RAM. He plans to plan a same scene the utilization of a GTX 750, which he believes will likely be sufficient to manage with it.

Despite the terrifying complexity of the scene, the machine wasn’t finish to elephantine load. The developer furthermore claimed that “any GTX 10 sequence card or newer is gonna ogle some mind blowing enhancements.”

The developer was furthermore requested about the asset’s size, to which he spoke back it was 1.5GB. Gary added that games will “indubitably salvage bigger,” nonetheless the closing size of a sport continues to be the developer’s responsibility. If he had spent some more time around the asset, he believes he would possibly perchance perchance possess reduced its size to 50-100MB with out sacrificing its quality.