Scheduling radar and communication functions in a frequency-division multiplexing (FDM) fashion is a classic solution to achieve mutual-interference-free integrated sensing and communication (ISAC) system. However, it is haunted by poor spectral efficiency (SE) in comparison to the spectrum-sharing scheme. In this paper, we propose a novel photonics millimeter-wave ISAC system reaping the FDM scheme while offering inspirations for significantly enhancing SE using super-resolution techniques. Using the coherent fusion processing (CFP) of sparse sub-band linear frequency modulated (LFM) radar signals, a small fraction of the total bandwidth can enable a full-band-equivalent radar range resolution and more spectrum resources can be allocated for communications, leading to an ultrahigh SE for ISAC system. Moreover, an improved sparse sub-band fusion algorithm is developed based on particle swarm optimization to enhance the accuracy of synthesized radar range profiles. In experiments, a 60-GHz super-resolution FDM ISAC system using photonics-assisted millimeter-wave signal generation and dual-parallel coherent de-chirping processing is established, wherein the idle spectrum between two widely-spaced LFM sub-bands is loaded with broadband orthogonal frequency division multiplexing (OFDM) communication signals. Our system can simultaneously achieve up to 18-Gbit/s communication data rate and a radar range resolution as high as 2.14 cm, by exploring the globally unlicensed 7-GHz spectrum around 60-GHz with 4.5-GHz communication and 2-GHz radar bandwidths. In addition, a 0.97-cm radar range resolution corresponding to 16-GHz equivalent bandwidth (EB) is experimentally obtained by using two 1.5-GHz LFM sub-bands. This renders only ∼9.4% EB loss (3 GHz/32 GHz), without significant performance tradeoff between sensing and communications.