The paper doesn’t calculate the radius of the star’s Roche limit, instead opting to calculate the orbital period of the Roche limit. I’ve never done a Roche limit calculation for stars, but I have for planets/moons, and I’m not seeing anything that suggests it’s different than for planets. So, I think I did this correctly (excepting typos):
The star’s Roche limit is about 1.5 million km from its centre (~1 million km above its surface), and the planet’s orbit is about 2 million km from the star’s centre. Assuming a circular orbit, which should be the case at these distances, the orbit has a circumference of about 12.7 million km, and the planet is whipping around at a speed of about 2.3 million km/h, or 0.2% the speed of light.