Adobe claims 2x speed increases as Photoshop and Premiere added to NVIDIA RTX Spark optimization

Adobe and NVIDIA extend RTX Spark optimization to Premiere and Photoshop, as well as Frame.io workflows and Substance 3D.

Adobe and NVIDIA have expanded their partnership first announced in March to optimize Photoshop, Premiere, and Substance 3D for NVIDIA RTX Spark-powered machines such as the latest ASUS P16 and P14, as well as ramp up GPU acceleration for Frame.io.

The initial partnership set the groundwork for all this and included GPU acceleration for media ingestion and intelligence in Frame.io, as well as starting a work program around Substance 3D, Firefly Foundry, and 3D digital twin workflows for enterprise customers. The introduction of the NVIDIA RTX Spark superchip, which NVIDIA started shipping earlier this month, takes it all up a notch. Indeed, Adobe says that the rearchitected Premiere and Photoshop can deliver up to 2x faster AI processing, editing, color, and effects when running on the new hardware.

The updates will start rolling out later this year. Photoshop has been rearchitected around GPU-accelerated compositing. This should make it faster and more responsive when working on a number of creative tasks such as removing distractions with the Remove Tool or applying effects like Gaussian Blur and Light Leaks directly on canvas.

In Premiere, playback, rendering, effects, color, and AI-powered tools have been RTX-accelerated and optimized for RTX Spark, including Enhance Speech and Object Mask. Again, this should make working on the timeline even faster, while in Substance 3D Painter, RTX Spark speeds up the heaviest texturing steps.

Elsewhere in the Adobe portfolio, Frame.io’s cloud pipeline now uses NVIDIA NVDEC for hardware-accelerated decoding and TensorRT with CUDA for content preprocessing and embedding generation. Adobe says early results from selected transcode and search workloads show approximately 5x faster embedding generation, a 4-5x improvement in P95 processing latency, and a rough 80% smaller search proxy.