Floating-Point Precision, Rounding, and Exceptions
Understand IEEE 754 representation, rounding modes, and how the FPU and SSE units signal exceptions like overflow, underflow, and invalid operations.
Floating-Point Precision, Rounding, and Exceptions is a free Assembly Language & x86 Low-Level Systems Programming lesson on CoddyKit — lesson 4 of 4. You can read the complete lesson below for free — then practise it hands-on in the browser with a built-in code editor and a 24/7 AI tutor. It is part of the Assembly Language & x86 Low-Level Systems Programming learning path, one of 4 lessons in the course, and your progress syncs across the web and the CoddyKit app.
IEEE 754 Format
x86 floating-point follows the IEEE 754 standard. A number is stored as sign, exponent, and mantissa:
- Single (32-bit): 1 + 8 + 23
- Double (64-bit): 1 + 11 + 52
- Extended (80-bit): used internally by the x87 FPU
Why Precision Matters
Most decimal fractions cannot be represented exactly in binary. For example 0.1 + 0.2 does not equal exactly 0.3. Accumulated rounding error is a core hazard in numerical code.
The Four Rounding Modes
IEEE 754 defines four rounding modes:
- Round to nearest, ties to even (default)
- Round toward negative infinity (floor)
- Round toward positive infinity (ceil)
- Round toward zero (truncate)
Controlling Rounding on x87
The x87 control word holds the rounding-control (RC) bits. You read it with fstcw, modify, then load with fldcw.
sub esp, 4
fstcw [esp] ; store control word
or word [esp], 0x0C00 ; RC = 11 -> round toward zero
fldcw [esp] ; load it back
add esp, 4SSE MXCSR Register
SSE has its own control/status register, MXCSR. It holds rounding-control bits plus exception masks and flags. Manage it with ldmxcsr and stmxcsr.
sub rsp, 4
stmxcsr [rsp] ; save MXCSR
ldmxcsr [rsp] ; restore MXCSR
add rsp, 4Floating-Point Exceptions
The standard defines six exceptions:
- Invalid (e.g. 0/0, sqrt of negative)
- Denormal operand
- Divide by zero
- Overflow
- Underflow
- Inexact (rounding occurred)
Masked vs Unmasked
Each exception can be masked. A masked exception produces a default result (like infinity or NaN) and sets a flag. An unmasked exception raises a CPU trap so your handler can intervene.
Special Values: NaN and Infinity
IEEE 754 reserves bit patterns for special values:
- +Inf / -Inf from overflow or divide-by-zero
- NaN (Not a Number) from invalid operations
Any comparison with NaN is false — even NaN == NaN.
Detecting NaN
Because NaN is unordered, you detect it with an unordered compare. In SSE, ucomiss sets the parity flag when an operand is NaN.
ucomiss xmm0, xmm0 ; compare value with itself
jp is_nan ; PF set => NaN detectedA Runnable C Demonstration
This self-contained C program shows that floating-point addition is not exact and detects a NaN.
#include <stdio.h>
#include <math.h>
int main(void) {
double a = 0.1 + 0.2;
printf("0.1+0.2 = %.17f\n", a);
printf("equals 0.3? %d\n", a == 0.3);
double nan_val = 0.0 / 0.0;
printf("isnan? %d\n", isnan(nan_val));
return 0;
}Practical Advice
To write robust floating-point code:
- Never compare floats with
==; use an epsilon tolerance - Be aware of accumulation order in sums
- Keep the default round-to-nearest unless you have a reason
- Check status flags after risky operations
Quick Check
Test your floating-point knowledge.
Recap
You explored floating-point behavior:
- IEEE 754 defines single, double, and 80-bit extended formats
- Four rounding modes are configured via the x87 control word or MXCSR
- Six exceptions can be masked (default result + flag) or unmasked (trap)
- NaN and infinity are special values; NaN compares false to everything
Frequently asked questions
Is the “Floating-Point Precision, Rounding, and Exceptions” lesson free?
Yes — the full text of “Floating-Point Precision, Rounding, and Exceptions” is free to read here on the web, and the Assembly Language & x86 Low-Level Systems Programming course includes 4 lessons in total. To practise it interactively (a built-in code editor and a 24/7 AI tutor) and unlock the rest of the Assembly Language & x86 Low-Level Systems Programming course, upgrade to CoddyKit PRO.
What will I learn in “Floating-Point Precision, Rounding, and Exceptions”?
Understand IEEE 754 representation, rounding modes, and how the FPU and SSE units signal exceptions like overflow, underflow, and invalid operations. You practise Assembly Language & x86 Low-Level Systems Programming with hands-on code you run directly in the browser, and a 24/7 AI tutor answers your questions as you work through the lesson.
Do I need any experience to start Assembly Language & x86 Low-Level Systems Programming?
No prior experience is required. Assembly Language & x86 Low-Level Systems Programming on CoddyKit is structured for beginners through advanced learners; this is — lesson 4 of 4, so you can start here or from the beginning and move at your own pace.
How long does the “Floating-Point Precision, Rounding, and Exceptions” lesson take?
Most CoddyKit lessons take about 5–10 minutes. Each one is bite-sized and interactive, so you make steady progress and pick up exactly where you left off across the web and the app.
Can I write and run code in this Assembly Language & x86 Low-Level Systems Programming lesson?
Yes. Every Assembly Language & x86 Low-Level Systems Programming lesson includes a built-in code editor, so you write and run real code right in your browser and get instant AI feedback — no local setup required.
All lessons in this course
- x87 FPU Programming Basics
- SSE/AVX Instruction Sets Introduction
- Vectorizing Code with SIMD
- Floating-Point Precision, Rounding, and Exceptions