Would you prefer a tighter or a prolonged exam schedule? Would you prefer to take Math before Reading or the other way around? We exploit variation in end-of-course exam schedules across years and grades to identify distinct effects of the number of days between exams, the number of days since the first exam, and the exam order on subsequent performance. We find substantially different scheduling effects between STEM and non-STEM subjects. First, we find a positive relationship between exam performance in STEM subjects and exam order, controlling for other influences of scheduling, suggesting that the later in the schedule an exam is taken the higher the average performance. We call this phenomenon, exam warm-up. Second, we find a negative relationship between the number of days from the very first exam and subsequent exam performance in STEM subjects, suggesting the existence of a fatigue effect. For STEM subjects, the fatigue effect is estimated to be less than half the size of the warm-up effect. For non-STEM subjects, an additional day between exams is significantly associated with lower performance in subsequent exams. Students of lower prior performance have lower fatigue effects and higher warm-up effects in STEM subjects compared to students of higher prior performance. Also, we find that exam productivity in STEM increases faster for boys than it does for girls as they take additional exams due to a higher warm-up effect. Our findings suggest that low-cost changes in the exam schedule may have salient effects on student performance gaps.