(RV64 Only) SIMD Q15 Saturating Multiply Instructions
- __STATIC_FORCEINLINE unsigned long __RV_KDMBB16 (unsigned long a, unsigned long b)
- __STATIC_FORCEINLINE unsigned long __RV_KDMBT16 (unsigned long a, unsigned long b)
- __STATIC_FORCEINLINE unsigned long __RV_KDMTT16 (unsigned long a, unsigned long b)
- __STATIC_FORCEINLINE unsigned long __RV_KDMABB16 (unsigned long t, unsigned long a, unsigned long b)
- __STATIC_FORCEINLINE unsigned long __RV_KDMABT16 (unsigned long t, unsigned long a, unsigned long b)
- __STATIC_FORCEINLINE unsigned long __RV_KDMATT16 (unsigned long t, unsigned long a, unsigned long b)
- __STATIC_FORCEINLINE unsigned long __RV_KHMBB16 (unsigned long a, unsigned long b)
- __STATIC_FORCEINLINE unsigned long __RV_KHMBT16 (unsigned long a, unsigned long b)
- __STATIC_FORCEINLINE unsigned long __RV_KHMTT16 (unsigned long a, unsigned long b)
- group NMSIS_Core_DSP_Intrinsic_RV64_SIMD_Q15_SAT_MULT
(RV64 Only) SIMD Q15 Saturating Multiply Instructions
there are 9 (RV64 Only) SIMD Q15 saturating Multiply Instructions
Functions
- __STATIC_FORCEINLINE unsigned long __RV_KDMBB16 (unsigned long a, unsigned long b)
KDMBB16 (SIMD Signed Saturating Double Multiply B16 x B16)
Type: SIMD (RV64 only)
Syntax:
KDMxy16 Rd, Rs1, Rs2 (xy = BB, BT, TT)
Purpose
:
Multiply the signed Q15 integer contents of two 16-bit data in the corresponding portion of the 32-bit chunks in registers and then double and saturate the Q31 results into the 32-bit chunks in the destination register. If saturation happens, an overflow flag OV will be set.
Description
:
Multiply the top or bottom 16-bit Q15 content of the 32-bit portions in Rs1 with the top or bottom 16-bit Q15 content of the 32-bit portions in Rs2. The Q30 results are then doubled and saturated into Q31 values. The Q31 values are then written into the 32-bit chunks in Rd. When both the two Q15 inputs are 0x8000, saturation will happen. The result will be saturated to 0x7FFFFFFF and the overflow flag OV will be set.
Operations:
// KDMBB16: (x,y,z)=(0,0,0),(2,2,1) // KDMBT16: (x,y,z)=(0,1,0),(2,3,1) // KDMTT16: (x,y,z)=(1,1,0),(3,3,1) aop[z] = Rs1.H[x]; bop[z] = Rs2.H[y]; If (0x8000 != aop[z] | 0x8000 != bop[z]) { Mresult[z] = aop[z] * bop[z]; resQ31[z] = Mresult[z] << 1; } else { resQ31[z] = 0x7FFFFFFF; OV = 1; } Rd.W[z] = resQ31[z];
- Parameters
a – [in] unsigned long type of value stored in a
b – [in] unsigned long type of value stored in b
- Returns
value stored in unsigned long type
- __STATIC_FORCEINLINE unsigned long __RV_KDMBT16 (unsigned long a, unsigned long b)
KDMBT16 (SIMD Signed Saturating Double Multiply B16 x T16)
Type: SIMD (RV64 only)
Syntax:
KDMxy16 Rd, Rs1, Rs2 (xy = BB, BT, TT)
Purpose
:
Multiply the signed Q15 integer contents of two 16-bit data in the corresponding portion of the 32-bit chunks in registers and then double and saturate the Q31 results into the 32-bit chunks in the destination register. If saturation happens, an overflow flag OV will be set.
Description
:
Multiply the top or bottom 16-bit Q15 content of the 32-bit portions in Rs1 with the top or bottom 16-bit Q15 content of the 32-bit portions in Rs2. The Q30 results are then doubled and saturated into Q31 values. The Q31 values are then written into the 32-bit chunks in Rd. When both the two Q15 inputs are 0x8000, saturation will happen. The result will be saturated to 0x7FFFFFFF and the overflow flag OV will be set.
Operations:
// KDMBB16: (x,y,z)=(0,0,0),(2,2,1) // KDMBT16: (x,y,z)=(0,1,0),(2,3,1) // KDMTT16: (x,y,z)=(1,1,0),(3,3,1) aop[z] = Rs1.H[x]; bop[z] = Rs2.H[y]; If (0x8000 != aop[z] | 0x8000 != bop[z]) { Mresult[z] = aop[z] * bop[z]; resQ31[z] = Mresult[z] << 1; } else { resQ31[z] = 0x7FFFFFFF; OV = 1; } Rd.W[z] = resQ31[z];
- Parameters
a – [in] unsigned long type of value stored in a
b – [in] unsigned long type of value stored in b
- Returns
value stored in unsigned long type
- __STATIC_FORCEINLINE unsigned long __RV_KDMTT16 (unsigned long a, unsigned long b)
KDMTT16 (SIMD Signed Saturating Double Multiply T16 x T16)
Type: SIMD (RV64 only)
Syntax:
KDMxy16 Rd, Rs1, Rs2 (xy = BB, BT, TT)
Purpose
:
Multiply the signed Q15 integer contents of two 16-bit data in the corresponding portion of the 32-bit chunks in registers and then double and saturate the Q31 results into the 32-bit chunks in the destination register. If saturation happens, an overflow flag OV will be set.
Description
:
Multiply the top or bottom 16-bit Q15 content of the 32-bit portions in Rs1 with the top or bottom 16-bit Q15 content of the 32-bit portions in Rs2. The Q30 results are then doubled and saturated into Q31 values. The Q31 values are then written into the 32-bit chunks in Rd. When both the two Q15 inputs are 0x8000, saturation will happen. The result will be saturated to 0x7FFFFFFF and the overflow flag OV will be set.
Operations:
// KDMBB16: (x,y,z)=(0,0,0),(2,2,1) // KDMBT16: (x,y,z)=(0,1,0),(2,3,1) // KDMTT16: (x,y,z)=(1,1,0),(3,3,1) aop[z] = Rs1.H[x]; bop[z] = Rs2.H[y]; If (0x8000 != aop[z] | 0x8000 != bop[z]) { Mresult[z] = aop[z] * bop[z]; resQ31[z] = Mresult[z] << 1; } else { resQ31[z] = 0x7FFFFFFF; OV = 1; } Rd.W[z] = resQ31[z];
- Parameters
a – [in] unsigned long type of value stored in a
b – [in] unsigned long type of value stored in b
- Returns
value stored in unsigned long type
- __STATIC_FORCEINLINE unsigned long __RV_KDMABB16 (unsigned long t, unsigned long a, unsigned long b)
KDMABB16 (SIMD Signed Saturating Double Multiply Addition B16 x B16)
Type: SIMD (RV64 only)
Syntax:
KDMAxy16 Rd, Rs1, Rs2 (xy = BB, BT, TT)
Purpose
:
Multiply the signed Q15 integer contents of two 16-bit data in the corresponding portion of the 32-bit chunks in registers and then double and saturate the Q31 results, add the results with the values of the corresponding 32-bit chunks from the destination register and write the saturated addition results back into the corresponding 32-bit chunks of the destination register. If saturation happens, an overflow flag OV will be set.
Description
:
Multiply the top or bottom 16-bit Q15 content of the 32-bit portions in Rs1 with the top or bottom 16-bit Q15 content of the corresponding 32-bit portions in Rs2. The Q30 results are then doubled and saturated into Q31 values. The Q31 values are then added with the content of the corresponding 32-bit portions of Rd. If the addition results are beyond the Q31 number range (-2^31 <= Q31 <= 2^31-1), they are saturated to the range and the OV flag is set to 1. The results after saturation are written back to Rd. When both the two Q15 inputs are 0x8000, saturation will happen and the overflow flag OV will be set.
Operations:
// KDMABB16: (x,y,z)=(0,0,0),(2,2,1) // KDMABT16: (x,y,z)=(0,1,0),(2,3,1) // KDMATT16: (x,y,z)=(1,1,0),(3,3,1) aop[z] = Rs1.H[x]; bop[z] = Rs2.H[y]; If (0x8000 != aop[z] | 0x8000 != bop[z]) { Mresult[z] = aop[z] * bop[z]; resQ31[z] = Mresult[z] << 1; } else { resQ31[z] = 0x7FFFFFFF; OV = 1; } resadd[z] = Rd.W[z] + resQ31[z]; if (resadd[z] > (2^31)-1) { resadd[z] = (2^31)-1; OV = 1; } else if (resadd[z] < -2^31) { resadd[z] = -2^31; OV = 1; } Rd.W[z] = resadd[z];
- Parameters
t – [in] unsigned long type of value stored in t
a – [in] unsigned long type of value stored in a
b – [in] unsigned long type of value stored in b
- Returns
value stored in unsigned long type
- __STATIC_FORCEINLINE unsigned long __RV_KDMABT16 (unsigned long t, unsigned long a, unsigned long b)
KDMABT16 (SIMD Signed Saturating Double Multiply Addition B16 x T16)
Type: SIMD (RV64 only)
Syntax:
KDMAxy16 Rd, Rs1, Rs2 (xy = BB, BT, TT)
Purpose
:
Multiply the signed Q15 integer contents of two 16-bit data in the corresponding portion of the 32-bit chunks in registers and then double and saturate the Q31 results, add the results with the values of the corresponding 32-bit chunks from the destination register and write the saturated addition results back into the corresponding 32-bit chunks of the destination register. If saturation happens, an overflow flag OV will be set.
Description
:
Multiply the top or bottom 16-bit Q15 content of the 32-bit portions in Rs1 with the top or bottom 16-bit Q15 content of the corresponding 32-bit portions in Rs2. The Q30 results are then doubled and saturated into Q31 values. The Q31 values are then added with the content of the corresponding 32-bit portions of Rd. If the addition results are beyond the Q31 number range (-2^31 <= Q31 <= 2^31-1), they are saturated to the range and the OV flag is set to 1. The results after saturation are written back to Rd. When both the two Q15 inputs are 0x8000, saturation will happen and the overflow flag OV will be set.
Operations:
// KDMABB16: (x,y,z)=(0,0,0),(2,2,1) // KDMABT16: (x,y,z)=(0,1,0),(2,3,1) // KDMATT16: (x,y,z)=(1,1,0),(3,3,1) aop[z] = Rs1.H[x]; bop[z] = Rs2.H[y]; If (0x8000 != aop[z] | 0x8000 != bop[z]) { Mresult[z] = aop[z] * bop[z]; resQ31[z] = Mresult[z] << 1; } else { resQ31[z] = 0x7FFFFFFF; OV = 1; } resadd[z] = Rd.W[z] + resQ31[z]; if (resadd[z] > (2^31)-1) { resadd[z] = (2^31)-1; OV = 1; } else if (resadd[z] < -2^31) { resadd[z] = -2^31; OV = 1; } Rd.W[z] = resadd[z];
- Parameters
t – [in] unsigned long type of value stored in t
a – [in] unsigned long type of value stored in a
b – [in] unsigned long type of value stored in b
- Returns
value stored in unsigned long type
- __STATIC_FORCEINLINE unsigned long __RV_KDMATT16 (unsigned long t, unsigned long a, unsigned long b)
KDMATT16 (SIMD Signed Saturating Double Multiply Addition T16 x T16)
Type: SIMD (RV64 only)
Syntax:
KDMAxy16 Rd, Rs1, Rs2 (xy = BB, BT, TT)
Purpose
:
Multiply the signed Q15 integer contents of two 16-bit data in the corresponding portion of the 32-bit chunks in registers and then double and saturate the Q31 results, add the results with the values of the corresponding 32-bit chunks from the destination register and write the saturated addition results back into the corresponding 32-bit chunks of the destination register. If saturation happens, an overflow flag OV will be set.
Description
:
Multiply the top or bottom 16-bit Q15 content of the 32-bit portions in Rs1 with the top or bottom 16-bit Q15 content of the corresponding 32-bit portions in Rs2. The Q30 results are then doubled and saturated into Q31 values. The Q31 values are then added with the content of the corresponding 32-bit portions of Rd. If the addition results are beyond the Q31 number range (-2^31 <= Q31 <= 2^31-1), they are saturated to the range and the OV flag is set to 1. The results after saturation are written back to Rd. When both the two Q15 inputs are 0x8000, saturation will happen and the overflow flag OV will be set.
Operations:
// KDMABB16: (x,y,z)=(0,0,0),(2,2,1) // KDMABT16: (x,y,z)=(0,1,0),(2,3,1) // KDMATT16: (x,y,z)=(1,1,0),(3,3,1) aop[z] = Rs1.H[x]; bop[z] = Rs2.H[y]; If (0x8000 != aop[z] | 0x8000 != bop[z]) { Mresult[z] = aop[z] * bop[z]; resQ31[z] = Mresult[z] << 1; } else { resQ31[z] = 0x7FFFFFFF; OV = 1; } resadd[z] = Rd.W[z] + resQ31[z]; if (resadd[z] > (2^31)-1) { resadd[z] = (2^31)-1; OV = 1; } else if (resadd[z] < -2^31) { resadd[z] = -2^31; OV = 1; } Rd.W[z] = resadd[z];
- Parameters
t – [in] unsigned long type of value stored in t
a – [in] unsigned long type of value stored in a
b – [in] unsigned long type of value stored in b
- Returns
value stored in unsigned long type
- __STATIC_FORCEINLINE unsigned long __RV_KHMBB16 (unsigned long a, unsigned long b)
KHMBB16 (SIMD Signed Saturating Half Multiply B16 x B16)
Type: SIMD (RV64 Only)
Syntax:
KHMxy16 Rd, Rs1, Rs2 (xy = BB, BT, TT)
Purpose
:
Multiply the signed Q15 integer contents of two 16-bit data in the corresponding portion of the 32-bit chunks in registers and then right-shift 15 bits to turn the Q30 results into Q15 numbers again and saturate the Q15 results into the destination register. If saturation happens, an overflow flag OV will be set.
Description
:
Multiply the top or bottom 16-bit Q15 content of the 32-bit portions in Rs1 with the top or bottom 16-bit Q15 content of the 32-bit portion in Rs2. The Q30 results are then right-shifted 15- bits and saturated into Q15 values. The 32-bit Q15 values are then written into the 32-bit chunks in Rd. When both the two Q15 inputs are 0x8000, saturation will happen. The result will be saturated to 0x7FFF and the overflow flag OV will be set.
Operations:
// KHMBB16: (x,y,z)=(0,0,0),(2,2,1) // KHMBT16: (x,y,z)=(0,1,0),(2,3,1) // KHMTT16: (x,y,z)=(1,1,0),(3,3,1) aop = Rs1.H[x]; bop = Rs2.H[y]; If (0x8000 != aop | 0x8000 != bop) { Mresult[31:0] = aop * bop; res[15:0] = Mresult[30:15]; } else { res[15:0] = 0x7FFF; OV = 1; } Rd.W[z] = SE32(res[15:0]);
- Parameters
a – [in] unsigned long type of value stored in a
b – [in] unsigned long type of value stored in b
- Returns
value stored in unsigned long type
- __STATIC_FORCEINLINE unsigned long __RV_KHMBT16 (unsigned long a, unsigned long b)
KHMBT16 (SIMD Signed Saturating Half Multiply B16 x T16)
Type: SIMD (RV64 Only)
Syntax:
KHMxy16 Rd, Rs1, Rs2 (xy = BB, BT, TT)
Purpose
:
Multiply the signed Q15 integer contents of two 16-bit data in the corresponding portion of the 32-bit chunks in registers and then right-shift 15 bits to turn the Q30 results into Q15 numbers again and saturate the Q15 results into the destination register. If saturation happens, an overflow flag OV will be set.
Description
:
Multiply the top or bottom 16-bit Q15 content of the 32-bit portions in Rs1 with the top or bottom 16-bit Q15 content of the 32-bit portion in Rs2. The Q30 results are then right-shifted 15- bits and saturated into Q15 values. The 32-bit Q15 values are then written into the 32-bit chunks in Rd. When both the two Q15 inputs are 0x8000, saturation will happen. The result will be saturated to 0x7FFF and the overflow flag OV will be set.
Operations:
// KHMBB16: (x,y,z)=(0,0,0),(2,2,1) // KHMBT16: (x,y,z)=(0,1,0),(2,3,1) // KHMTT16: (x,y,z)=(1,1,0),(3,3,1) aop = Rs1.H[x]; bop = Rs2.H[y]; If (0x8000 != aop | 0x8000 != bop) { Mresult[31:0] = aop * bop; res[15:0] = Mresult[30:15]; } else { res[15:0] = 0x7FFF; OV = 1; } Rd.W[z] = SE32(res[15:0]);
- Parameters
a – [in] unsigned long type of value stored in a
b – [in] unsigned long type of value stored in b
- Returns
value stored in unsigned long type
- __STATIC_FORCEINLINE unsigned long __RV_KHMTT16 (unsigned long a, unsigned long b)
KHMTT16 (SIMD Signed Saturating Half Multiply T16 x T16)
Type: SIMD (RV64 Only)
Syntax:
KHMxy16 Rd, Rs1, Rs2 (xy = BB, BT, TT)
Purpose
:
Multiply the signed Q15 integer contents of two 16-bit data in the corresponding portion of the 32-bit chunks in registers and then right-shift 15 bits to turn the Q30 results into Q15 numbers again and saturate the Q15 results into the destination register. If saturation happens, an overflow flag OV will be set.
Description
:
Multiply the top or bottom 16-bit Q15 content of the 32-bit portions in Rs1 with the top or bottom 16-bit Q15 content of the 32-bit portion in Rs2. The Q30 results are then right-shifted 15- bits and saturated into Q15 values. The 32-bit Q15 values are then written into the 32-bit chunks in Rd. When both the two Q15 inputs are 0x8000, saturation will happen. The result will be saturated to 0x7FFF and the overflow flag OV will be set.
Operations:
// KHMBB16: (x,y,z)=(0,0,0),(2,2,1) // KHMBT16: (x,y,z)=(0,1,0),(2,3,1) // KHMTT16: (x,y,z)=(1,1,0),(3,3,1) aop = Rs1.H[x]; bop = Rs2.H[y]; If (0x8000 != aop | 0x8000 != bop) { Mresult[31:0] = aop * bop; res[15:0] = Mresult[30:15]; } else { res[15:0] = 0x7FFF; OV = 1; } Rd.W[z] = SE32(res[15:0]);
- Parameters
a – [in] unsigned long type of value stored in a
b – [in] unsigned long type of value stored in b
- Returns
value stored in unsigned long type