diff mbox series

[32/38] target/riscv: RV64 Only SIMD Q15 saturating Multiply Instructions

Message ID 20210212150256.885-33-zhiwei_liu@c-sky.com
State New
Headers show
Series target/riscv: support packed extension v0.9.2 | expand

Commit Message

LIU Zhiwei Feb. 12, 2021, 3:02 p.m. UTC 
Signed-off-by: LIU Zhiwei <zhiwei_liu@c-sky.com>
---
 target/riscv/helper.h                   |  10 ++
 target/riscv/insn32-64.decode           |  10 ++
 target/riscv/insn_trans/trans_rvp.c.inc |  11 ++
 target/riscv/packed_helper.c            | 141 ++++++++++++++++++++++++
 4 files changed, 172 insertions(+)
diff mbox series

Patch

diff --git a/target/riscv/helper.h b/target/riscv/helper.h
index 384b42ce90..f8521a5388 100644
--- a/target/riscv/helper.h
+++ b/target/riscv/helper.h
@@ -1440,4 +1440,14 @@  DEF_HELPER_3(umin32, tl, env, tl, tl)
 DEF_HELPER_3(smax32, tl, env, tl, tl)
 DEF_HELPER_3(umax32, tl, env, tl, tl)
 DEF_HELPER_2(kabs32, tl, env, tl)
+
+DEF_HELPER_3(khmbb16, tl, env, tl, tl)
+DEF_HELPER_3(khmbt16, tl, env, tl, tl)
+DEF_HELPER_3(khmtt16, tl, env, tl, tl)
+DEF_HELPER_3(kdmbb16, tl, env, tl, tl)
+DEF_HELPER_3(kdmbt16, tl, env, tl, tl)
+DEF_HELPER_3(kdmtt16, tl, env, tl, tl)
+DEF_HELPER_4(kdmabb16, tl, env, tl, tl, tl)
+DEF_HELPER_4(kdmabt16, tl, env, tl, tl, tl)
+DEF_HELPER_4(kdmatt16, tl, env, tl, tl, tl)
 #endif
diff --git a/target/riscv/insn32-64.decode b/target/riscv/insn32-64.decode
index a2b8831467..2e1c1817e4 100644
--- a/target/riscv/insn32-64.decode
+++ b/target/riscv/insn32-64.decode
@@ -135,3 +135,13 @@  umin32     1010000  ..... ..... 010 ..... 1111111 @r
 smax32     1001001  ..... ..... 010 ..... 1111111 @r
 umax32     1010001  ..... ..... 010 ..... 1111111 @r
 kabs32     1010110  10010 ..... 000 ..... 1111111 @r2
+
+khmbb16    1101110  ..... ..... 001 ..... 1111111 @r
+khmbt16    1110110  ..... ..... 001 ..... 1111111 @r
+khmtt16    1111110  ..... ..... 001 ..... 1111111 @r
+kdmbb16    1101101  ..... ..... 001 ..... 1111111 @r
+kdmbt16    1110101  ..... ..... 001 ..... 1111111 @r
+kdmtt16    1111101  ..... ..... 001 ..... 1111111 @r
+kdmabb16   1101100  ..... ..... 001 ..... 1111111 @r
+kdmabt16   1110100  ..... ..... 001 ..... 1111111 @r
+kdmatt16   1111100  ..... ..... 001 ..... 1111111 @r
diff --git a/target/riscv/insn_trans/trans_rvp.c.inc b/target/riscv/insn_trans/trans_rvp.c.inc
index ce144ee5c0..2b4418abd8 100644
--- a/target/riscv/insn_trans/trans_rvp.c.inc
+++ b/target/riscv/insn_trans/trans_rvp.c.inc
@@ -1179,4 +1179,15 @@  GEN_RVP_R_OOL(umin32);
 GEN_RVP_R_OOL(smax32);
 GEN_RVP_R_OOL(umax32);
 GEN_RVP_R2_OOL(kabs32);
+
+/* (RV64 Only) SIMD Q15 saturating Multiply Instructions */
+GEN_RVP_R_OOL(khmbb16);
+GEN_RVP_R_OOL(khmbt16);
+GEN_RVP_R_OOL(khmtt16);
+GEN_RVP_R_OOL(kdmbb16);
+GEN_RVP_R_OOL(kdmbt16);
+GEN_RVP_R_OOL(kdmtt16);
+GEN_RVP_R_ACC_OOL(kdmabb16);
+GEN_RVP_R_ACC_OOL(kdmabt16);
+GEN_RVP_R_ACC_OOL(kdmatt16);
 #endif
diff --git a/target/riscv/packed_helper.c b/target/riscv/packed_helper.c
index c8a92f5b7d..5636848aaf 100644
--- a/target/riscv/packed_helper.c
+++ b/target/riscv/packed_helper.c
@@ -3432,4 +3432,145 @@  static inline void do_kabs32(CPURISCVState *env, void *vd, void *va, uint8_t i)
 }
 
 RVPR2(kabs32, 1, 4);
+
+/* (RV64 Only) SIMD Q15 saturating Multiply Instructions */
+static inline void do_khmbb16(CPURISCVState *env, void *vd, void *va,
+                              void *vb, uint8_t i)
+{
+    int32_t *d = vd;
+    int16_t *a = va, *b = vb;
+
+    d[H4(i / 2)] = sat64(env, (int64_t)a[H2(i)] * b[H2(i)] >> 15, 15);
+}
+
+RVPR(khmbb16, 2, 2);
+
+static inline void do_khmbt16(CPURISCVState *env, void *vd, void *va,
+                              void *vb, uint8_t i)
+{
+    int32_t *d = vd;
+    int16_t *a = va, *b = vb;
+
+    d[H4(i / 2)] = sat64(env, (int64_t)a[H2(i)] * b[H2(i + 1)] >> 15, 15);
+}
+
+RVPR(khmbt16, 2, 2);
+
+static inline void do_khmtt16(CPURISCVState *env, void *vd, void *va,
+                              void *vb, uint8_t i)
+{
+    int32_t *d = vd;
+    int16_t *a = va, *b = vb;
+
+    d[H4(i / 2)] = sat64(env, (int64_t)a[H2(i + 1)] * b[H2(i + 1)] >> 15, 15);
+}
+
+RVPR(khmtt16, 2, 2);
+
+static inline void do_kdmbb16(CPURISCVState *env, void *vd, void *va,
+                              void *vb, uint8_t i)
+{
+    int32_t *d = vd;
+    int16_t *a = va, *b = vb;
+
+    if (a[H2(i)] == INT16_MIN && b[H2(i)] == INT16_MIN) {
+        d[H4(i / 2)] = INT32_MAX;
+        env->vxsat = 0x1;
+    } else {
+        d[H4(i / 2)] = (int64_t)a[H2(i)] * b[H2(i)] << 1;
+    }
+}
+
+RVPR(kdmbb16, 2, 2);
+
+static inline void do_kdmbt16(CPURISCVState *env, void *vd, void *va,
+                              void *vb, uint8_t i)
+{
+    int32_t *d = vd;
+    int16_t *a = va, *b = vb;
+
+    if (a[H2(i)] == INT16_MIN && b[H2(i + 1)] == INT16_MIN) {
+        d[H4(i / 2)] = INT32_MAX;
+        env->vxsat = 0x1;
+    } else {
+        d[H4(i / 2)] = (int64_t)a[H2(i)] * b[H2(i + 1)] << 1;
+    }
+}
+
+RVPR(kdmbt16, 2, 2);
+
+static inline void do_kdmtt16(CPURISCVState *env, void *vd, void *va,
+                              void *vb, uint8_t i)
+{
+    int32_t *d = vd;
+    int16_t *a = va, *b = vb;
+
+    if (a[H2(i + 1)] == INT16_MIN && b[H2(i + 1)] == INT16_MIN) {
+        d[H4(i / 2)] = INT32_MAX;
+        env->vxsat = 0x1;
+    } else {
+        d[H4(i / 2)] = (int64_t)a[H2(i + 1)] * b[H2(i + 1)] << 1;
+    }
+}
+
+RVPR(kdmtt16, 2, 2);
+
+static inline void do_kdmabb16(CPURISCVState *env, void *vd, void *va,
+                               void *vb, void *vc, uint8_t i)
+
+{
+    int32_t *d = vd;
+    int16_t *a = va, *b = vb;
+    int32_t *c = vc, m0;
+
+    if (a[H2(i)] == INT16_MIN && b[H2(i)] == INT16_MIN) {
+        m0 = INT32_MAX;
+        env->vxsat = 0x1;
+    } else {
+        m0 = (int32_t)a[H2(i)] * b[H2(i)] << 1;
+    }
+    d[H4(i / 2)] = sadd32(env, 0, c[H4(i)], m0);
+}
+
+RVPR_ACC(kdmabb16, 2, 2);
+
+static inline void do_kdmabt16(CPURISCVState *env, void *vd, void *va,
+                               void *vb, void *vc, uint8_t i)
+
+{
+    int32_t *d = vd;
+    int16_t *a = va, *b = vb;
+    int32_t *c = vc, m0;
+
+    if (a[H2(i)] == INT16_MIN && b[H2(i + 1)] == INT16_MIN) {
+        m0 = INT32_MAX;
+        env->vxsat = 0x1;
+    } else {
+        m0 = (int32_t)a[H2(i)] * b[H2(i + 1)] << 1;
+    }
+    d[H4(i / 2)] = sadd32(env, 0, c[H4(i)], m0);
+}
+
+RVPR_ACC(kdmabt16, 2, 2);
+
+static inline void do_kdmatt16(CPURISCVState *env, void *vd, void *va,
+                               void *vb, void *vc, uint8_t i)
+
+{
+    int32_t *d = vd;
+    int16_t *a = va, *b = vb;
+    int32_t *c = vc, m0;
+
+    if (a[H2(i + 1)] == INT16_MIN && b[H2(i + 1)] == INT16_MIN) {
+        m0 = INT32_MAX;
+        env->vxsat = 0x1;
+    } else {
+        m0 = (int32_t)a[H2(i + 1)] * b[H2(i + 1)] << 1;
+    }
+    *d = sadd32(env, 0, c[H4(i)], m0);
+}
+
+RVPR_ACC(kdmatt16, 2, 2);
+
+
 #endif