Benchmarks for JPEG codec on NVIDIA GeForce GPU

We have built super fast JPEG codec on NVIDIA CUDA technology. That codec developed by Fastvideo is a blend of strict compliance with standards and shocking encoding and decoding speed comparing with the fastest existing commercial solutions. This is full, performance-oriented implementation of Baseline JPEG. We got ultra fast JPEG compression and decompression on the GPU due to full parallel implementation of Baseline JPEG algorithm. Our CUDA JPEG codec is the fastest in comparison with the best commercial multithreaded JPEG codecs for multicore CPUs.

Benchmarks for JPEG decoding

• Image resolution: 1920 × 1080 (2 MPix), 8-bit per channel, RGB
• Test description: all data in GPU memory, timing includes GPU computations only
• Software: OS Windows-10/11, CUDA-12.6
• Hardware: CPU AMD Ryzen9 7950X (16 cores, 4.5–5.7 GHz)
• Hardware: NVIDIA GeForce RTX 4090, PCI-Express 4.0 ×16
• Encoding parameters: quality q = 95, subsampling 4:4:4
• Performance for MJPEG encoding - 4600 fps
• Performance for MJPEG decoding - 1500 fps

The above performance benchmarks were done for one-thread application and they could be improved further with multithreading. For Linux with MPS one could get even better results.

Benchmarks for JPEG encoding and decoding on NVIDIA GeForce RTX 4090

We've measured JPEG encoding performance on the latest NVIDIA GeForce RTX 4090 for high resolution images from modern machine vision cameras. JPEG compression (image resolution 5328 × 4508, 24-bit, subsampling 4:2:0, quality 90, image compression ratio ~10:1) could be done within 0.65 ms which corresponds to performance ~117 GB/s. Assuming that real bandwidth of the PCIe-4.0 x16 interface is around 24 GB/s, this is almost 5 times faster.