The IBM 360 was delivered and came with standard sqrt and transcendental routines. They were pretty good.

Hirando Kuki at Univ of Chicago supplied another set about a year later. They were faster, shorter, and more accurate.

When IBM was designing the ACS machine in about 1966, they asked Kuki what he needed. He said only, round right.

The 360’s always rounded towards zero. Dynamic rounding options had not been proposed then.