Duration of work split over N cores, not ignoring "add-on" overheads
6
表达式7: "y" greater than or equal to left parenthesis, "s" Subscript, "e" "r" "i" "a" "l" , Baseline positive StartFraction, "p" Subscript, "a" "r" "a" "l" "l" "e" "l" , Baseline Over max left parenthesis, 1 , "x" , right parenthesis , EndFraction positive "a" Subscript, "d" "d" "o" "n" , Baseline , right parenthesisy≥serial+parallelmax1,x+addon
Speedup's glass-ceiling, even ∞ many free CPUcores will never touch...
9
表达式10: "y" less than or equal to StartFraction, left parenthesis, "s" Subscript, "e" "r" "i" "a" "l" , Baseline positive "p" Subscript, "a" "r" "a" "l" "l" "e" "l" , Baseline , right parenthesis Over left parenthesis, "s" Subscript, "e" "r" "i" "a" "l" , Baseline positive StartNestedFraction, "p" Subscript, "a" "r" "a" "l" "l" "e" "l" , Baseline NestedOver max left parenthesis, 1 , "x" , right parenthesis , EndNestedFraction plus "a" Subscript, "d" "d" "o" "n" , Baseline , right parenthesis , EndFractiony≤serial+parallelserial+parallelmax1,x+addon