Read this thread for an idea of what to expect running Syncthing on “underpowered” hardware like your phone or Core 2 Duo-based system.
The upshot is that the initial indexing of all the files is very computationally intensive on x86 systems without hardware crypto and on ARM systems because Syncthing cryptographically hashes all the files that you tell it to watch.
One of my Syncthing nodes is a dinky little ARM board (Allwinner A20) and the initial indexing of ~90 gigabytes took over 12 hours (I stopped counting). Since then, the performance has been reasonable considering the hardware.
When you make a change that needs to be synced, Syncthing will use as much CPU as you give it to hash the new data and encrypt with AES for transfer. I run Syncthing under nice and ionice to limit its effect on my workstation. Some have argued in these forums that Syncthing should “play nice” by default but I disagree. Any software should do its job as fast as possible. And cryptographic hashing and encryption takes a lot of CPU.