RISC-V: Allow all const_vec_duplicates as constants.
Checks
Commit Message
Hi,
as we can always broadcast an integer constant to a vector register
allow them in riscv_const_insns. We need as many instructions as
it takes to generate the constant and one vmv.vx.
Regards
Robin
gcc/ChangeLog:
* config/riscv/riscv.cc (riscv_const_insns): Allow
const_vec_duplicates.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/autovec/vmv-imm-fixed-rv32.c: Add vmv.v.x
tests.
* gcc.target/riscv/rvv/autovec/vmv-imm-fixed-rv64.c: Dito.
* gcc.target/riscv/rvv/autovec/vmv-imm-run.c: Dito.
* gcc.target/riscv/rvv/autovec/vmv-imm-rv32.c: Dito.
* gcc.target/riscv/rvv/autovec/vmv-imm-rv64.c: Dito.
* gcc.target/riscv/rvv/autovec/vmv-imm-template.h: Dito.
---
gcc/config/riscv/riscv.cc | 22 ++++--
.../riscv/rvv/autovec/vmv-imm-fixed-rv32.c | 3 +-
.../riscv/rvv/autovec/vmv-imm-fixed-rv64.c | 3 +-
.../riscv/rvv/autovec/vmv-imm-run.c | 8 ++
.../riscv/rvv/autovec/vmv-imm-rv32.c | 3 +-
.../riscv/rvv/autovec/vmv-imm-rv64.c | 3 +-
.../riscv/rvv/autovec/vmv-imm-template.h | 74 ++++++++++---------
7 files changed, 73 insertions(+), 43 deletions(-)
Comments
Lgtm
Robin Dapp <rdapp@ventanamicro.com> 於 2023年5月26日 週五 22:10 寫道:
> Hi,
>
> as we can always broadcast an integer constant to a vector register
> allow them in riscv_const_insns. We need as many instructions as
> it takes to generate the constant and one vmv.vx.
>
> Regards
> Robin
>
> gcc/ChangeLog:
>
> * config/riscv/riscv.cc (riscv_const_insns): Allow
> const_vec_duplicates.
>
> gcc/testsuite/ChangeLog:
>
> * gcc.target/riscv/rvv/autovec/vmv-imm-fixed-rv32.c: Add vmv.v.x
> tests.
> * gcc.target/riscv/rvv/autovec/vmv-imm-fixed-rv64.c: Dito.
> * gcc.target/riscv/rvv/autovec/vmv-imm-run.c: Dito.
> * gcc.target/riscv/rvv/autovec/vmv-imm-rv32.c: Dito.
> * gcc.target/riscv/rvv/autovec/vmv-imm-rv64.c: Dito.
> * gcc.target/riscv/rvv/autovec/vmv-imm-template.h: Dito.
> ---
> gcc/config/riscv/riscv.cc | 22 ++++--
> .../riscv/rvv/autovec/vmv-imm-fixed-rv32.c | 3 +-
> .../riscv/rvv/autovec/vmv-imm-fixed-rv64.c | 3 +-
> .../riscv/rvv/autovec/vmv-imm-run.c | 8 ++
> .../riscv/rvv/autovec/vmv-imm-rv32.c | 3 +-
> .../riscv/rvv/autovec/vmv-imm-rv64.c | 3 +-
> .../riscv/rvv/autovec/vmv-imm-template.h | 74 ++++++++++---------
> 7 files changed, 73 insertions(+), 43 deletions(-)
>
> diff --git a/gcc/config/riscv/riscv.cc b/gcc/config/riscv/riscv.cc
> index c2acab08b89..5e0b515fe7a 100644
> --- a/gcc/config/riscv/riscv.cc
> +++ b/gcc/config/riscv/riscv.cc
> @@ -1297,12 +1297,22 @@ riscv_const_insns (rtx x)
> }
> }
>
> - /* Constants from -16 to 15 can be loaded with vmv.v.i.
> - The Wc0, Wc1 constraints are already covered by the
> - vi constraint so we do not need to check them here
> - separately. */
> - if (satisfies_constraint_vi (x))
> - return 1;
> + rtx elt;
> + if (const_vec_duplicate_p (x, &elt))
> + {
> + /* Constants from -16 to 15 can be loaded with vmv.v.i.
> + The Wc0, Wc1 constraints are already covered by the
> + vi constraint so we do not need to check them here
> + separately. */
> + if (satisfies_constraint_vi (x))
> + return 1;
> +
> + /* A const duplicate vector can always be broadcast from
> + a general-purpose register. This means we need as many
> + insns as it takes to load the constant into the GPR
> + and one vmv.v.x. */
> + return 1 + riscv_integer_cost (INTVAL (elt));
> + }
>
> /* TODO: We may support more const vector in the future. */
> return x == CONST0_RTX (GET_MODE (x)) ? 1 : 0;
> diff --git
> a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-fixed-rv32.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-fixed-rv32.c
> index 631ea3bf268..e8d017f7339 100644
> --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-fixed-rv32.c
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-fixed-rv32.c
> @@ -3,4 +3,5 @@
>
> #include "vmv-imm-template.h"
>
> -/* { dg-final { scan-assembler-times "vmv.v.i" 32 } } */
> +/* { dg-final { scan-assembler-times {vmv.v.i} 32 } } */
> +/* { dg-final { scan-assembler-times {vmv.v.x} 8 } } */
> diff --git
> a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-fixed-rv64.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-fixed-rv64.c
> index 7ded6cc18d2..f85ad4117d3 100644
> --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-fixed-rv64.c
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-fixed-rv64.c
> @@ -3,4 +3,5 @@
>
> #include "vmv-imm-template.h"
>
> -/* { dg-final { scan-assembler-times "vmv.v.i" 32 } } */
> +/* { dg-final { scan-assembler-times {vmv.v.i} 32 } } */
> +/* { dg-final { scan-assembler-times {vmv.v.x} 8 } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-run.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-run.c
> index 22b1958af81..79099a37bdd 100644
> --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-run.c
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-run.c
> @@ -53,4 +53,12 @@ int main ()
> TEST_POS(uint64_t, 13)
> TEST_POS(uint64_t, 14)
> TEST_POS(uint64_t, 15)
> + TEST_POS(uint32_t, 16)
> + TEST_POS(uint32_t, 123)
> + TEST_POS(uint32_t, 255)
> + TEST_POS(uint32_t, 999)
> + TEST_POS(uint32_t, 32701)
> + TEST_POS(uint32_t, 65535)
> + TEST_POS(uint32_t, 65536)
> + TEST_POS(uint32_t, 923423)
> }
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-rv32.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-rv32.c
> index c419256cd45..6843bc6018d 100644
> --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-rv32.c
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-rv32.c
> @@ -3,4 +3,5 @@
>
> #include "vmv-imm-template.h"
>
> -/* { dg-final { scan-assembler-times "vmv.v.i" 32 } } */
> +/* { dg-final { scan-assembler-times {vmv.v.i} 32 } } */
> +/* { dg-final { scan-assembler-times {vmv.v.x} 8 } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-rv64.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-rv64.c
> index e386166f95e..39fb2a6cc7b 100644
> --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-rv64.c
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-rv64.c
> @@ -3,4 +3,5 @@
>
> #include "vmv-imm-template.h"
>
> -/* { dg-final { scan-assembler-times "vmv.v.i" 32 } } */
> +/* { dg-final { scan-assembler-times {vmv.v.i} 32 } } */
> +/* { dg-final { scan-assembler-times {vmv.v.x} 8 } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-template.h
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-template.h
> index 855343d7e3e..84b26e0f1c2 100644
> --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-template.h
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vmv-imm-template.h
> @@ -17,38 +17,46 @@
> dst[i] = -VAL; \
> }
>
> -#define TEST_ALL() \
> -VMV_NEG(int8_t,16) \
> -VMV_NEG(int8_t,15) \
> -VMV_NEG(int8_t,14) \
> -VMV_NEG(int8_t,13) \
> -VMV_NEG(int16_t,12) \
> -VMV_NEG(int16_t,11) \
> -VMV_NEG(int16_t,10) \
> -VMV_NEG(int16_t,9) \
> -VMV_NEG(int32_t,8) \
> -VMV_NEG(int32_t,7) \
> -VMV_NEG(int32_t,6) \
> -VMV_NEG(int32_t,5) \
> -VMV_NEG(int64_t,4) \
> -VMV_NEG(int64_t,3) \
> -VMV_NEG(int64_t,2) \
> -VMV_NEG(int64_t,1) \
> -VMV_POS(uint8_t,0) \
> -VMV_POS(uint8_t,1) \
> -VMV_POS(uint8_t,2) \
> -VMV_POS(uint8_t,3) \
> -VMV_POS(uint16_t,4) \
> -VMV_POS(uint16_t,5) \
> -VMV_POS(uint16_t,6) \
> -VMV_POS(uint16_t,7) \
> -VMV_POS(uint32_t,8) \
> -VMV_POS(uint32_t,9) \
> -VMV_POS(uint32_t,10) \
> -VMV_POS(uint32_t,11) \
> -VMV_POS(uint64_t,12) \
> -VMV_POS(uint64_t,13) \
> -VMV_POS(uint64_t,14) \
> -VMV_POS(uint64_t,15)
> +#define TEST_ALL() \
> +VMV_NEG(int8_t,16) \
> +VMV_NEG(int8_t,15) \
> +VMV_NEG(int8_t,14) \
> +VMV_NEG(int8_t,13) \
> +VMV_NEG(int16_t,12) \
> +VMV_NEG(int16_t,11) \
> +VMV_NEG(int16_t,10) \
> +VMV_NEG(int16_t,9) \
> +VMV_NEG(int32_t,8) \
> +VMV_NEG(int32_t,7) \
> +VMV_NEG(int32_t,6) \
> +VMV_NEG(int32_t,5) \
> +VMV_NEG(int64_t,4) \
> +VMV_NEG(int64_t,3) \
> +VMV_NEG(int64_t,2) \
> +VMV_NEG(int64_t,1) \
> +VMV_POS(uint8_t,0) \
> +VMV_POS(uint8_t,1) \
> +VMV_POS(uint8_t,2) \
> +VMV_POS(uint8_t,3) \
> +VMV_POS(uint16_t,4) \
> +VMV_POS(uint16_t,5) \
> +VMV_POS(uint16_t,6) \
> +VMV_POS(uint16_t,7) \
> +VMV_POS(uint32_t,8) \
> +VMV_POS(uint32_t,9) \
> +VMV_POS(uint32_t,10) \
> +VMV_POS(uint32_t,11) \
> +VMV_POS(uint64_t,12) \
> +VMV_POS(uint64_t,13) \
> +VMV_POS(uint64_t,14) \
> +VMV_POS(uint64_t,15) \
> +VMV_POS(uint32_t,16) \
> +VMV_POS(uint32_t,123) \
> +VMV_POS(uint32_t,255) \
> +VMV_POS(uint32_t,999) \
> +VMV_POS(uint32_t,32701) \
> +VMV_POS(uint32_t,65535) \
> +VMV_POS(uint32_t,65536) \
> +VMV_POS(uint32_t,923423) \
>
> TEST_ALL()
> --
> 2.40.1
>
@@ -1297,12 +1297,22 @@ riscv_const_insns (rtx x)
}
}
- /* Constants from -16 to 15 can be loaded with vmv.v.i.
- The Wc0, Wc1 constraints are already covered by the
- vi constraint so we do not need to check them here
- separately. */
- if (satisfies_constraint_vi (x))
- return 1;
+ rtx elt;
+ if (const_vec_duplicate_p (x, &elt))
+ {
+ /* Constants from -16 to 15 can be loaded with vmv.v.i.
+ The Wc0, Wc1 constraints are already covered by the
+ vi constraint so we do not need to check them here
+ separately. */
+ if (satisfies_constraint_vi (x))
+ return 1;
+
+ /* A const duplicate vector can always be broadcast from
+ a general-purpose register. This means we need as many
+ insns as it takes to load the constant into the GPR
+ and one vmv.v.x. */
+ return 1 + riscv_integer_cost (INTVAL (elt));
+ }
/* TODO: We may support more const vector in the future. */
return x == CONST0_RTX (GET_MODE (x)) ? 1 : 0;
@@ -3,4 +3,5 @@
#include "vmv-imm-template.h"
-/* { dg-final { scan-assembler-times "vmv.v.i" 32 } } */
+/* { dg-final { scan-assembler-times {vmv.v.i} 32 } } */
+/* { dg-final { scan-assembler-times {vmv.v.x} 8 } } */
@@ -3,4 +3,5 @@
#include "vmv-imm-template.h"
-/* { dg-final { scan-assembler-times "vmv.v.i" 32 } } */
+/* { dg-final { scan-assembler-times {vmv.v.i} 32 } } */
+/* { dg-final { scan-assembler-times {vmv.v.x} 8 } } */
@@ -53,4 +53,12 @@ int main ()
TEST_POS(uint64_t, 13)
TEST_POS(uint64_t, 14)
TEST_POS(uint64_t, 15)
+ TEST_POS(uint32_t, 16)
+ TEST_POS(uint32_t, 123)
+ TEST_POS(uint32_t, 255)
+ TEST_POS(uint32_t, 999)
+ TEST_POS(uint32_t, 32701)
+ TEST_POS(uint32_t, 65535)
+ TEST_POS(uint32_t, 65536)
+ TEST_POS(uint32_t, 923423)
}
@@ -3,4 +3,5 @@
#include "vmv-imm-template.h"
-/* { dg-final { scan-assembler-times "vmv.v.i" 32 } } */
+/* { dg-final { scan-assembler-times {vmv.v.i} 32 } } */
+/* { dg-final { scan-assembler-times {vmv.v.x} 8 } } */
@@ -3,4 +3,5 @@
#include "vmv-imm-template.h"
-/* { dg-final { scan-assembler-times "vmv.v.i" 32 } } */
+/* { dg-final { scan-assembler-times {vmv.v.i} 32 } } */
+/* { dg-final { scan-assembler-times {vmv.v.x} 8 } } */
@@ -17,38 +17,46 @@
dst[i] = -VAL; \
}
-#define TEST_ALL() \
-VMV_NEG(int8_t,16) \
-VMV_NEG(int8_t,15) \
-VMV_NEG(int8_t,14) \
-VMV_NEG(int8_t,13) \
-VMV_NEG(int16_t,12) \
-VMV_NEG(int16_t,11) \
-VMV_NEG(int16_t,10) \
-VMV_NEG(int16_t,9) \
-VMV_NEG(int32_t,8) \
-VMV_NEG(int32_t,7) \
-VMV_NEG(int32_t,6) \
-VMV_NEG(int32_t,5) \
-VMV_NEG(int64_t,4) \
-VMV_NEG(int64_t,3) \
-VMV_NEG(int64_t,2) \
-VMV_NEG(int64_t,1) \
-VMV_POS(uint8_t,0) \
-VMV_POS(uint8_t,1) \
-VMV_POS(uint8_t,2) \
-VMV_POS(uint8_t,3) \
-VMV_POS(uint16_t,4) \
-VMV_POS(uint16_t,5) \
-VMV_POS(uint16_t,6) \
-VMV_POS(uint16_t,7) \
-VMV_POS(uint32_t,8) \
-VMV_POS(uint32_t,9) \
-VMV_POS(uint32_t,10) \
-VMV_POS(uint32_t,11) \
-VMV_POS(uint64_t,12) \
-VMV_POS(uint64_t,13) \
-VMV_POS(uint64_t,14) \
-VMV_POS(uint64_t,15)
+#define TEST_ALL() \
+VMV_NEG(int8_t,16) \
+VMV_NEG(int8_t,15) \
+VMV_NEG(int8_t,14) \
+VMV_NEG(int8_t,13) \
+VMV_NEG(int16_t,12) \
+VMV_NEG(int16_t,11) \
+VMV_NEG(int16_t,10) \
+VMV_NEG(int16_t,9) \
+VMV_NEG(int32_t,8) \
+VMV_NEG(int32_t,7) \
+VMV_NEG(int32_t,6) \
+VMV_NEG(int32_t,5) \
+VMV_NEG(int64_t,4) \
+VMV_NEG(int64_t,3) \
+VMV_NEG(int64_t,2) \
+VMV_NEG(int64_t,1) \
+VMV_POS(uint8_t,0) \
+VMV_POS(uint8_t,1) \
+VMV_POS(uint8_t,2) \
+VMV_POS(uint8_t,3) \
+VMV_POS(uint16_t,4) \
+VMV_POS(uint16_t,5) \
+VMV_POS(uint16_t,6) \
+VMV_POS(uint16_t,7) \
+VMV_POS(uint32_t,8) \
+VMV_POS(uint32_t,9) \
+VMV_POS(uint32_t,10) \
+VMV_POS(uint32_t,11) \
+VMV_POS(uint64_t,12) \
+VMV_POS(uint64_t,13) \
+VMV_POS(uint64_t,14) \
+VMV_POS(uint64_t,15) \
+VMV_POS(uint32_t,16) \
+VMV_POS(uint32_t,123) \
+VMV_POS(uint32_t,255) \
+VMV_POS(uint32_t,999) \
+VMV_POS(uint32_t,32701) \
+VMV_POS(uint32_t,65535) \
+VMV_POS(uint32_t,65536) \
+VMV_POS(uint32_t,923423) \
TEST_ALL()