This adds some missing tests for vf[nw]cvt.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/autovec/conversions/vfncvt-ftoi-run.c:
Add tests.
* gcc.target/riscv/rvv/autovec/conversions/vfncvt-ftoi-rv32gcv.c:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfncvt-ftoi-rv64gcv.c:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfncvt-ftoi-template.h:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfncvt-itof-rv32gcv.c:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfncvt-itof-rv64gcv.c:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfncvt-itof-template.h:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfncvt-itof-zvfh-run.c:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfwcvt-ftoi-rv32gcv.c:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfwcvt-ftoi-rv64gcv.c:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfwcvt-ftoi-template.h:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfwcvt-ftoi-zvfh-run.c:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfwcvt-itof-run.c:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfwcvt-itof-rv32gcv.c:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfwcvt-itof-rv64gcv.c:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfwcvt-itof-template.h:
Ditto.
* gcc.target/riscv/rvv/autovec/conversions/vfwcvt-itof-zvfh-run.c:
Ditto.
RUN2 (float, uint16_t, 4096)
RUN2 (float, uint16_t, 5975)
+ RUN (float, int8_t, 3)
+ RUN (float, int8_t, 4)
+ RUN (float, int8_t, 7)
+ RUN (float, int8_t, 99)
+ RUN (float, int8_t, 119)
+ RUN (float, int8_t, 128)
+ RUN (float, int8_t, 256)
+ RUN (float, int8_t, 279)
+ RUN (float, int8_t, 555)
+ RUN (float, int8_t, 1024)
+ RUN (float, int8_t, 1389)
+ RUN (float, int8_t, 2048)
+ RUN (float, int8_t, 3989)
+ RUN (float, int8_t, 4096)
+ RUN (float, int8_t, 5975)
+
+ RUN2 (float, uint8_t, 3)
+ RUN2 (float, uint8_t, 4)
+ RUN2 (float, uint8_t, 7)
+ RUN2 (float, uint8_t, 99)
+ RUN2 (float, uint8_t, 119)
+ RUN2 (float, uint8_t, 128)
+ RUN2 (float, uint8_t, 256)
+ RUN2 (float, uint8_t, 279)
+ RUN2 (float, uint8_t, 555)
+ RUN2 (float, uint8_t, 1024)
+ RUN2 (float, uint8_t, 1389)
+ RUN2 (float, uint8_t, 2048)
+ RUN2 (float, uint8_t, 3989)
+ RUN2 (float, uint8_t, 4096)
+ RUN2 (float, uint8_t, 5975)
+
RUN (double, int32_t, 3)
RUN (double, int32_t, 4)
RUN (double, int32_t, 7)
RUN2 (double, uint32_t, 3989)
RUN2 (double, uint32_t, 4096)
RUN2 (double, uint32_t, 5975)
+
+ RUN (double, int16_t, 3)
+ RUN (double, int16_t, 4)
+ RUN (double, int16_t, 7)
+ RUN (double, int16_t, 99)
+ RUN (double, int16_t, 119)
+ RUN (double, int16_t, 128)
+ RUN (double, int16_t, 256)
+ RUN (double, int16_t, 279)
+ RUN (double, int16_t, 555)
+ RUN (double, int16_t, 1024)
+ RUN (double, int16_t, 1389)
+ RUN (double, int16_t, 2048)
+ RUN (double, int16_t, 3989)
+ RUN (double, int16_t, 4096)
+ RUN (double, int16_t, 5975)
+
+ RUN2 (double, uint16_t, 3)
+ RUN2 (double, uint16_t, 4)
+ RUN2 (double, uint16_t, 7)
+ RUN2 (double, uint16_t, 99)
+ RUN2 (double, uint16_t, 119)
+ RUN2 (double, uint16_t, 128)
+ RUN2 (double, uint16_t, 256)
+ RUN2 (double, uint16_t, 279)
+ RUN2 (double, uint16_t, 555)
+ RUN2 (double, uint16_t, 1024)
+ RUN2 (double, uint16_t, 1389)
+ RUN2 (double, uint16_t, 2048)
+ RUN2 (double, uint16_t, 3989)
+ RUN2 (double, uint16_t, 4096)
+ RUN2 (double, uint16_t, 5975)
+
+ RUN (double, int8_t, 3)
+ RUN (double, int8_t, 4)
+ RUN (double, int8_t, 7)
+ RUN (double, int8_t, 99)
+ RUN (double, int8_t, 119)
+ RUN (double, int8_t, 128)
+ RUN (double, int8_t, 256)
+ RUN (double, int8_t, 279)
+ RUN (double, int8_t, 555)
+ RUN (double, int8_t, 1024)
+ RUN (double, int8_t, 1389)
+ RUN (double, int8_t, 2048)
+ RUN (double, int8_t, 3989)
+ RUN (double, int8_t, 4096)
+ RUN (double, int8_t, 5975)
+
+ RUN2 (double, uint8_t, 3)
+ RUN2 (double, uint8_t, 4)
+ RUN2 (double, uint8_t, 7)
+ RUN2 (double, uint8_t, 99)
+ RUN2 (double, uint8_t, 119)
+ RUN2 (double, uint8_t, 128)
+ RUN2 (double, uint8_t, 256)
+ RUN2 (double, uint8_t, 279)
+ RUN2 (double, uint8_t, 555)
+ RUN2 (double, uint8_t, 1024)
+ RUN2 (double, uint8_t, 1389)
+ RUN2 (double, uint8_t, 2048)
+ RUN2 (double, uint8_t, 3989)
+ RUN2 (double, uint8_t, 4096)
+ RUN2 (double, uint8_t, 5975)
}
/* { dg-do compile } */
-/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv32gcv_zvfh -mabi=ilp32d --param=riscv-autovec-preference=scalable" } */
+/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv32gcv_zvfh -mabi=ilp32d -fno-trapping-math --param=riscv-autovec-preference=scalable" } */
#include "vfncvt-ftoi-template.h"
-/* { dg-final { scan-assembler-times {\tvfncvt\.rtz\.x\.f\.w} 3 } } */
+/* The vectorizer only performs int -> float conversions with
+ intermediate types with -fno-trapping-math. */
+/* { dg-final { scan-assembler-times {\tvfncvt\.rtz\.x\.f\.w} 9 } } */
/* { dg-final { scan-assembler-times {\tvfncvt\.rtz\.xu\.f\.w} 3 } } */
/* { dg-do compile } */
-/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv64gcv_zvfh -mabi=lp64d --param=riscv-autovec-preference=scalable" } */
+/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv64gcv_zvfh -mabi=lp64d -fno-trapping-math --param=riscv-autovec-preference=scalable" } */
#include "vfncvt-ftoi-template.h"
-/* { dg-final { scan-assembler-times {\tvfncvt\.rtz\.x\.f\.w} 3 } } */
+/* The vectorizer only performs int -> float conversions with
+ intermediate types with -fno-trapping-math. */
+/* { dg-final { scan-assembler-times {\tvfncvt\.rtz\.x\.f\.w} 9 } } */
/* { dg-final { scan-assembler-times {\tvfncvt\.rtz\.xu\.f\.w} 3 } } */
#define TEST_ALL() \
TEST (double, int32_t) \
TEST (double, uint32_t) \
+ TEST (double, int16_t) \
+ TEST (double, uint16_t) \
+ TEST (double, int8_t) \
+ TEST (double, uint8_t) \
TEST (float, int16_t) \
TEST (float, uint16_t) \
+ TEST (float, int8_t) \
+ TEST (float, uint8_t) \
TEST (_Float16, int8_t) \
TEST (_Float16, uint8_t) \
#include "vfncvt-itof-template.h"
+/* { dg-final { scan-assembler-times {\tvfcvt\.f\.x\.v} 2 } } */
/* { dg-final { scan-assembler-times {\tvfncvt\.f\.x\.w} 2 } } */
/* { dg-final { scan-assembler-times {\tvfncvt\.f\.xu\.w} 2 } } */
#include "vfncvt-itof-template.h"
-/* { dg-final { scan-assembler-times {\tvfncvt\.f\.x\.w} 2 } } */
-/* { dg-final { scan-assembler-times {\tvfncvt\.f\.xu\.w} 2 } } */
+/* { dg-final { scan-assembler-times {\tvfncvt\.f\.x\.w} 5 } } */
+/* { dg-final { scan-assembler-times {\tvfncvt\.f\.xu\.w} 1 } } */
TYPE1 *restrict a, int n) \
{ \
for (int i = 0; i < n; i++) \
- dst[i] = (TYPE2) a[i]; \
+ dst[i] = (TYPE2) (a[i] & 0x7ffffffful); \
}
+
#define TEST_ALL() \
TEST (int64_t, float) \
TEST (uint64_t, float) \
+ TEST (int64_t, _Float16) \
+ TEST (uint64_t, _Float16) \
TEST (int32_t, _Float16) \
TEST (uint32_t, _Float16) \
RUN2 (uint32_t, _Float16, 3989)
RUN2 (uint32_t, _Float16, 4096)
RUN2 (uint32_t, _Float16, 5975)
+
+ RUN (int64_t, _Float16, 3)
+ RUN (int64_t, _Float16, 4)
+ RUN (int64_t, _Float16, 7)
+ RUN (int64_t, _Float16, 99)
+ RUN (int64_t, _Float16, 119)
+ RUN (int64_t, _Float16, 128)
+ RUN (int64_t, _Float16, 256)
+ RUN (int64_t, _Float16, 279)
+ RUN (int64_t, _Float16, 555)
+ RUN (int64_t, _Float16, 1024)
+ RUN (int64_t, _Float16, 1389)
+ RUN (int64_t, _Float16, 2048)
+ RUN (int64_t, _Float16, 3989)
+ RUN (int64_t, _Float16, 4096)
+ RUN (int64_t, _Float16, 5975)
+
+ RUN2 (uint64_t, _Float16, 3)
+ RUN2 (uint64_t, _Float16, 4)
+ RUN2 (uint64_t, _Float16, 7)
+ RUN2 (uint64_t, _Float16, 99)
+ RUN2 (uint64_t, _Float16, 119)
+ RUN2 (uint64_t, _Float16, 128)
+ RUN2 (uint64_t, _Float16, 256)
+ RUN2 (uint64_t, _Float16, 279)
+ RUN2 (uint64_t, _Float16, 555)
+ RUN2 (uint64_t, _Float16, 1024)
+ RUN2 (uint64_t, _Float16, 1389)
+ RUN2 (uint64_t, _Float16, 2048)
+ RUN2 (uint64_t, _Float16, 3989)
+ RUN2 (uint64_t, _Float16, 4096)
+ RUN2 (uint64_t, _Float16, 5975)
}
/* { dg-do compile } */
-/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv32gcv_zvfh -mabi=ilp32d --param=riscv-autovec-preference=scalable" } */
+/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv32gcv_zvfh -mabi=ilp32d -fno-trapping-math --param=riscv-autovec-preference=scalable" } */
#include "vfwcvt-ftoi-template.h"
-/* { dg-final { scan-assembler-times {\tvfwcvt\.rtz\.x\.f\.v} 2 } } */
+/* { dg-final { scan-assembler-times {\tvfwcvt\.rtz\.x\.f\.v} 4 } } */
/* { dg-final { scan-assembler-times {\tvfwcvt\.rtz\.xu\.f\.v} 2 } } */
/* { dg-do compile } */
-/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv64gcv_zvfh -mabi=lp64d --param=riscv-autovec-preference=scalable" } */
+/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv64gcv_zvfh -mabi=lp64d -fno-trapping-math --param=riscv-autovec-preference=scalable" } */
#include "vfwcvt-ftoi-template.h"
-/* { dg-final { scan-assembler-times {\tvfwcvt\.rtz\.x\.f\.v} 2 } } */
+/* { dg-final { scan-assembler-times {\tvfwcvt\.rtz\.x\.f\.v} 4 } } */
/* { dg-final { scan-assembler-times {\tvfwcvt\.rtz\.xu\.f\.v} 2 } } */
}
#define TEST_ALL() \
+ TEST (_Float16, int64_t) \
+ TEST (_Float16, uint64_t) \
TEST (_Float16, int32_t) \
TEST (_Float16, uint32_t) \
TEST (float, int64_t) \
RUN2 (_Float16, uint32_t, 3989)
RUN2 (_Float16, uint32_t, 4096)
RUN2 (_Float16, uint32_t, 5975)
+
+ RUN (_Float16, int64_t, 3)
+ RUN (_Float16, int64_t, 4)
+ RUN (_Float16, int64_t, 7)
+ RUN (_Float16, int64_t, 99)
+ RUN (_Float16, int64_t, 119)
+ RUN (_Float16, int64_t, 128)
+ RUN (_Float16, int64_t, 256)
+ RUN (_Float16, int64_t, 279)
+ RUN (_Float16, int64_t, 555)
+ RUN (_Float16, int64_t, 1024)
+ RUN (_Float16, int64_t, 1389)
+ RUN (_Float16, int64_t, 2048)
+ RUN (_Float16, int64_t, 3989)
+ RUN (_Float16, int64_t, 4096)
+ RUN (_Float16, int64_t, 5975)
+
+ RUN2 (_Float16, uint64_t, 3)
+ RUN2 (_Float16, uint64_t, 4)
+ RUN2 (_Float16, uint64_t, 7)
+ RUN2 (_Float16, uint64_t, 99)
+ RUN2 (_Float16, uint64_t, 119)
+ RUN2 (_Float16, uint64_t, 128)
+ RUN2 (_Float16, uint64_t, 256)
+ RUN2 (_Float16, uint64_t, 279)
+ RUN2 (_Float16, uint64_t, 555)
+ RUN2 (_Float16, uint64_t, 1024)
+ RUN2 (_Float16, uint64_t, 1389)
+ RUN2 (_Float16, uint64_t, 2048)
+ RUN2 (_Float16, uint64_t, 3989)
+ RUN2 (_Float16, uint64_t, 4096)
+ RUN2 (_Float16, uint64_t, 5975)
}
int
main ()
{
+ RUN (int8_t, float, 3)
+ RUN (int8_t, float, 4)
+ RUN (int8_t, float, 7)
+ RUN (int8_t, float, 99)
+ RUN (int8_t, float, 119)
+ RUN (int8_t, float, 128)
+ RUN (int8_t, float, 256)
+ RUN (int8_t, float, 279)
+ RUN (int8_t, float, 555)
+ RUN (int8_t, float, 1024)
+ RUN (int8_t, float, 1389)
+ RUN (int8_t, float, 2048)
+ RUN (int8_t, float, 3989)
+ RUN (int8_t, float, 4096)
+ RUN (int8_t, float, 5975)
+
+ RUN2 (uint8_t, float, 3)
+ RUN2 (uint8_t, float, 4)
+ RUN2 (uint8_t, float, 7)
+ RUN2 (uint8_t, float, 99)
+ RUN2 (uint8_t, float, 119)
+ RUN2 (uint8_t, float, 128)
+ RUN2 (uint8_t, float, 256)
+ RUN2 (uint8_t, float, 279)
+ RUN2 (uint8_t, float, 555)
+ RUN2 (uint8_t, float, 1024)
+ RUN2 (uint8_t, float, 1389)
+ RUN2 (uint8_t, float, 2048)
+ RUN2 (uint8_t, float, 3989)
+ RUN2 (uint8_t, float, 4096)
+ RUN2 (uint8_t, float, 5975)
+
+ RUN (int8_t, double, 3)
+ RUN (int8_t, double, 4)
+ RUN (int8_t, double, 7)
+ RUN (int8_t, double, 99)
+ RUN (int8_t, double, 119)
+ RUN (int8_t, double, 128)
+ RUN (int8_t, double, 256)
+ RUN (int8_t, double, 279)
+ RUN (int8_t, double, 555)
+ RUN (int8_t, double, 1024)
+ RUN (int8_t, double, 1389)
+ RUN (int8_t, double, 2048)
+ RUN (int8_t, double, 3989)
+ RUN (int8_t, double, 4096)
+ RUN (int8_t, double, 5975)
+
+ RUN2 (uint8_t, double, 3)
+ RUN2 (uint8_t, double, 4)
+ RUN2 (uint8_t, double, 7)
+ RUN2 (uint8_t, double, 99)
+ RUN2 (uint8_t, double, 119)
+ RUN2 (uint8_t, double, 128)
+ RUN2 (uint8_t, double, 256)
+ RUN2 (uint8_t, double, 279)
+ RUN2 (uint8_t, double, 555)
+ RUN2 (uint8_t, double, 1024)
+ RUN2 (uint8_t, double, 1389)
+ RUN2 (uint8_t, double, 2048)
+ RUN2 (uint8_t, double, 3989)
+ RUN2 (uint8_t, double, 4096)
+ RUN2 (uint8_t, double, 5975)
+
RUN (int16_t, float, 3)
RUN (int16_t, float, 4)
RUN (int16_t, float, 7)
RUN2 (uint16_t, float, 4096)
RUN2 (uint16_t, float, 5975)
+ RUN (int16_t, double, 3)
+ RUN (int16_t, double, 4)
+ RUN (int16_t, double, 7)
+ RUN (int16_t, double, 99)
+ RUN (int16_t, double, 119)
+ RUN (int16_t, double, 128)
+ RUN (int16_t, double, 256)
+ RUN (int16_t, double, 279)
+ RUN (int16_t, double, 555)
+ RUN (int16_t, double, 1024)
+ RUN (int16_t, double, 1389)
+ RUN (int16_t, double, 2048)
+ RUN (int16_t, double, 3989)
+ RUN (int16_t, double, 4096)
+ RUN (int16_t, double, 5975)
+
+ RUN2 (uint16_t, double, 3)
+ RUN2 (uint16_t, double, 4)
+ RUN2 (uint16_t, double, 7)
+ RUN2 (uint16_t, double, 99)
+ RUN2 (uint16_t, double, 119)
+ RUN2 (uint16_t, double, 128)
+ RUN2 (uint16_t, double, 256)
+ RUN2 (uint16_t, double, 279)
+ RUN2 (uint16_t, double, 555)
+ RUN2 (uint16_t, double, 1024)
+ RUN2 (uint16_t, double, 1389)
+ RUN2 (uint16_t, double, 2048)
+ RUN2 (uint16_t, double, 3989)
+ RUN2 (uint16_t, double, 4096)
+ RUN2 (uint16_t, double, 5975)
+
RUN (int32_t, double, 3)
RUN (int32_t, double, 4)
RUN (int32_t, double, 7)
#include "vfwcvt-itof-template.h"
-/* { dg-final { scan-assembler-times {\tvfwcvt\.f\.x\.v} 3 } } */
+/* Conversions that the vectorizer does via multiple intermediate
+ types end up as signed conversions. */
+/* { dg-final { scan-assembler-times {\tvfwcvt\.f\.x\.v} 9 } } */
/* { dg-final { scan-assembler-times {\tvfwcvt\.f\.xu\.v} 3 } } */
#include "vfwcvt-itof-template.h"
-/* { dg-final { scan-assembler-times {\tvfwcvt\.f\.x\.v} 3 } } */
+/* Conversions that the vectorizer does via multiple intermediate
+ types end up as signed conversions. */
+/* { dg-final { scan-assembler-times {\tvfwcvt\.f\.x\.v} 9 } } */
/* { dg-final { scan-assembler-times {\tvfwcvt\.f\.xu\.v} 3 } } */
}
#define TEST_ALL() \
+ TEST (int8_t, _Float16) \
+ TEST (uint8_t, _Float16) \
+ TEST (int8_t, float) \
+ TEST (uint8_t, float) \
+ TEST (int8_t, double) \
+ TEST (uint8_t, double) \
TEST (int16_t, float) \
TEST (uint16_t, float) \
+ TEST (int16_t, double) \
+ TEST (uint16_t, double) \
TEST (int32_t, double) \
TEST (uint32_t, double) \
- TEST (int8_t, _Float16) \
- TEST (uint8_t, _Float16) \
TEST_ALL ()
RUN (int8_t, _Float16, 20)
RUN (int8_t, _Float16, 27)
- RUN (int8_t, _Float16, 4)
- RUN (int8_t, _Float16, 8)
- RUN (int8_t, _Float16, 11)
- RUN (int8_t, _Float16, 29)
- RUN (int8_t, _Float16, 49)
+ RUN (uint8_t, _Float16, 4)
+ RUN (uint8_t, _Float16, 8)
+ RUN (uint8_t, _Float16, 11)
+ RUN (uint8_t, _Float16, 29)
+ RUN (uint8_t, _Float16, 49)
}