Does FROUNDNX.Q write an integer register?
No. The result is written to floating-point register f[rd]; the floating-point value is integral.
FROUNDNX.Q
RISC-V FROUNDNX.Q Instruction Details
Instruction ManualR-typeRound a quad-precision floating-point value to an integral value in the same FP format; set NX only when the input is not NaN and differs from the rounded result.
Instruction Syntax
froundnx.q rd, rs1, rm
Operand Breakdown
rd: destination floating-point register receiving the rounded result in the same format.
rs1: the only source floating-point register; the encoded rs2 field is fixed to 00101 (decimal 5), not a second source floating-point register.
rm: rounding-mode field; rm=111 uses the dynamic frm rounding mode.
ZfaQFP Precision
Instruction Behavior
When the Q extension is implemented, FROUNDNX.Q rounds the quad-precision floating-point value in f[rs1] to an integral value according to rm, but writes the result back to f[rd] in the same quad-precision floating-point format. It sets NX only when the input is not NaN and differs from the rounded result. A signaling NaN sets NV; a quiet NaN produces the canonical NaN under the general FP NaN rule.
FROUNDNX.Q Decode And Execute Animation
Shows the Zfa + Q-extension quad-precision OP-FP flow: decode fields, read FP operands, perform FROUNDNX semantics, then write rd.
rd
rs1
rm
froundnx.q
,
,
Execution Context
FLEN=128 / fmt=Q (11)
rounding: RNE nearest-even
fflags: NX
31..27
26..25
24..20
19..15
14..12
11..7
6..0
01000
FROUNDNX
11
fmt=Q
00101
rs2=5 (roundToIntegralExact)
01011
rs1
000
rm
01010
rd
1010011
OP-FP
Execution Data Path
instruction
0x46558553
opcode
1010011 -> OP-FP
funct5/fmt
01000 + Q(11) -> FROUNDNX.Q
rm
000 -> RNE nearest-even
rd/rs
fa0(f10) / fa1(f11) / rs2=00101 (fixed)
read
1.5
FP op
roundToIntegralExact(1.5, RNE nearest-even) = 2.0
write
fa0(f10) = 2.0; NX
Current Step
Fetch: show the 32-bit OP-FP encoding
The word is split as an R-type OP-FP instruction; quad-precision OP-FP instructions use opcode 1010011.
encoding: 0x46558553
syntax : froundnx.q fa0(f10), fa1(f11), rne
result : fa0(f10) = 2.0; NX
This bounded teaching model computes whitelisted finite examples plus zero, infinity, and NaN; FROUNDNX.Q writes the integral value in Q format. Zero and infinity are copied. NX is set when the input differs from the rounded result and is not NaN.
Quad-precision Calculation
rs1
fa1(f11) = 1.5
rs2
ft5(f5) = 0.0
result
roundToIntegralExact(1.5, RNE nearest-even) = 2.0
This bounded teaching model computes whitelisted finite examples plus zero, infinity, and NaN; FROUNDNX.Q writes the integral value in Q format. Zero and infinity are copied. NX is set when the input differs from the rounded result and is not NaN.
This animation shows only Zfa + Q-extension ISA-visible OP-FP encoding, example numeric results, and fflags relationship; it does not model FPU pipelines, latency, every IEEE 754 corner case, or microarchitecture.
Quick Understanding & Search Notes
FROUNDNX.Q differs from FCVT by output type: FROUNDNX.Q still writes a floating-point register, with the value rounded to an integral value. It reports NX when a non-NaN input changes.
rm selects the rounding direction, with frm supplying dynamic rounding; rm=101/110 and reserved frm values under DYN have reserved behavior.
Special values such as zero and infinity follow the Zfa rules; signaling NaNs still set NV.
Common Usage Scenarios
Floating-Point Rounding
Understand this scenario with real code like «froundnx.q f0, f1, rne # NX only when non-NaN f1 changes after rounding».
Numeric Format Handling
Understand this scenario with real code like «froundnx.q f0, f1, rne # NX only when non-NaN f1 changes after rounding».
Pre-Use Checklist
Syntax Check
- rd: destination floating-point register receiving the rounded result in the same format.
- rs1: the only source floating-point register; the encoded rs2 field is fixed to 00101 (decimal 5), not a second source floating-point register.
- rm: rounding-mode field; rm=111 uses the dynamic frm rounding mode.
Semantic Check
- Confirm the floating-point format suffix (.H/.S/.D/.Q) matches the data width.
- Decide whether software must inspect fflags for NV, DZ, OF, UF, or NX.
Pitfalls / Common Confusions
FROUNDNX sets NX only when the input is not NaN and differs from the rounded result; do not confuse it with FROUND, which does not set NX.
The result is still floating-point, not an integer-register result; use FCVT for an integer result.
rm controls the direction for rounding to an integral value; static rm=101/110 and reserved frm values under DYN have reserved behavior.
FAQ
How does FROUNDNX.Q differ from the other FROUND form?
FROUNDNX sets NX only when the input is not NaN and differs from the rounded result; FROUND does not set NX.