Intel Quark SoC X1000 Core Developer's Manual page 286

Table 94. Floating-Point Clock Count Summary (Sheet 4 of 8)
Instruction
FSUB = Subtract Real from ST(0)
ST(0)←ST(0) – 32-bit memory
ST(0)←ST(0) – 64-bit memory
ST(d)←ST(0) – ST(i)
FSUBP = Subtract real from ST(0) and Pop (ST(i)← ST(0) -
ST(i))
FSUBR = Subtract Real reversed (Subtract ST(0) from Real)
ST(0)←32-bit memory – ST(0)
ST(0)←64-bit memory – ST(0)
ST(d)←ST(i) – ST(0)
FSUBRP = Subtract Real reversed and Pop (ST(i)← ST(i) -
ST(0))
FMUL = Multiply Real with ST(0)
ST(0)←ST(0) X 32-bit memory
ST(0)←ST(0) X 64-bit memory
ST(d)←ST(0) X ST(i)
FMULP = Multiply ST(0) with ST(i) and Pop (ST(i)← ST(0)
XST(i))
Notes:
1. If operand is 0 clock counts = 27.
2. If operand is 0 clock counts = 28.
3. If CW.PC indicates 24 bit precision then subtract 38 clocks.
If CW.PC indicates 53 bit precision then subtract 11 clocks.
4. If there is a numeric error pending from a previous instruction, add 17 clocks.
5. If there is a numeric error pending from a previous instruction, add 18 clocks.
6. The INT pin is polled several times while this function is executing to ensure short interrupt latency.
7. If ABS(operand) is greater than π/4 then add n clocks, where n=(operand/(π/4)).
®
Intel Quark SoC X1000 Core
Developer's Manual
286
Format
11011 000 : mod 100 r/m : s-i-b/disp.
11011 100 : mod 100 r/m : s-i-b/disp.
11011 d00 : 11001 ST(i)
11011 110 : 11001 ST(i)
11011 000 : mod 101 r/m : s-i-b/disp.
11011 100 : mod 101 r/m : s-i-b/disp.
11011 d00 : 11001 ST(i)
11011 110 : 11100 ST(i)
11011 000 : mod 001 r/m : s-i-b/disp.
11011 100 : mod 001 r/m : s-i-b/disp.
11011 d00 : 11001 ST(i)
11011 110 : 11001 ST(i)
®
Intel Quark Core—Instruction Set Summary
Cache Hit Concurrent
Penalty
Avg (Lower Execution
if
Range... Avg (Lower
Cache
Upper Range- Upper
Miss
Range)
10(8-20) 2
10(8-20) 3
10(8-20)
10(8-20)
10(8-20) 2
10(8-20) 3
10(8-20)
10(8-20)
11 2
14 3
16
16
Order Number: 329679-001US
Notes
Range)
7(5-17)
7(5-17)
7(5-17)
7(5-17)
7(5-17)
7(5-17)
7(5-17)
7(5-17)
8
11
13
13
October 2013