Inside Intel - The Future Of PC Performance, Panther Lake, Multi-Frame Gen - Tom Petersen Interview

In the interview, Tom Petersen from Intel discusses advancements in PC graphics performance, highlighting innovations like multi-frame generation, asynchronous shader compilation in Panther Lake, and the integration of AI to enhance smoothness and reduce stutter. He also emphasizes Intel’s focus on improving user experience through perceptual rendering, emerging display technologies, and expanding into handheld gaming markets with competitive performance and ecosystem development.

In this in-depth interview at CES, Tom Petersen from Intel discusses several cutting-edge technologies shaping the future of PC graphics performance, focusing heavily on Intel’s advancements with Arc GPUs and the new Panther Lake platform. Petersen opens by addressing frame generation technology, explaining the transition from two-frame to multi-frame generation (MFG) and the engineering challenges involved. He highlights the importance of frame pacing for delivering smooth visuals and envisions a future where rendering, frame generation, and projection are decoupled processes, potentially improving latency and visual quality, especially in VR and high-refresh-rate displays.

The conversation shifts to Intel’s efforts in reducing stuttering and improving frame smoothness, where Petersen introduces Intel’s PresentMon tool and the concept of animation error as a better metric than traditional FPS for assessing gaming performance. He emphasizes that frame generation should be seen as a visual enhancement rather than a pure performance boost, aiming to improve smoothness and reduce perceptible stutter. Petersen also discusses Intel’s approach to shader compilation stutter, revealing that Panther Lake supports asynchronous cloud-based precompiled shaders for DirectX 12 games, which helps minimize in-game stuttering by downloading shader caches compiled from Intel’s own gameplay runs.

Petersen touches on the evolving landscape of machine learning (ML) and AI in graphics, noting that while multi-frame generation and super-resolution techniques like XCSS are advancing, there remains complexity and fragmentation across vendors. He advocates for a common effects API, possibly through Microsoft’s DirectML or DirectSR initiatives, to simplify development and broaden cross-platform adoption. Additionally, he envisions AI playing a larger role in future graphics workflows, including denoising for ray tracing and improving smoothness by correcting temporal mismatches in frame rendering.

Regarding hardware and market positioning, Petersen highlights Panther Lake’s impressive performance, comparable to a mobile RTX 4050, and Intel’s ambition to innovate in emerging segments like handheld gaming devices. He discusses the need for a more mature ecosystem, including OS-level support and better UI frameworks tailored for handheld gaming, and notes Intel’s influence on laptop display technologies. Petersen also explores future display innovations, such as eye-tracking and foveated rendering, which could optimize rendering based on where the user is looking, although he acknowledges the technical challenges involved.

Finally, Petersen reflects on the broader future of graphics, emphasizing a shift away from raw frame rate metrics toward perceptual rendering focused on smoothness, latency, and image persistence. He stresses the importance of eliminating stutter and delivering a seamless experience as paramount goals, with AI, improved rasterization, and advanced display technologies all playing key roles. The interview closes with a hopeful outlook on competition and innovation in the mobile and handheld gaming markets, where Intel aims to bring fresh energy and performance improvements.