Segmented load test for load16_index32 * This test currently included a work around for compiler bug * To test bazel run //tests/cocotb:rvv_load_store_test_load16_index32_seg Change-Id: Ie5fb2b0bc63c396c18be6d77f44aeb5a32497442
diff --git a/tests/cocotb/BUILD b/tests/cocotb/BUILD index eb70a5c..07c64ac 100644 --- a/tests/cocotb/BUILD +++ b/tests/cocotb/BUILD
@@ -187,6 +187,7 @@ "load16_index8", "load16_index8_seg", "load16_index16_seg", + "load16_index32_seg", "load16_seg_unit", "load16_stride4_m1", "load16_stride4_m1_partial",
diff --git a/tests/cocotb/rvv/load_store/BUILD b/tests/cocotb/rvv/load_store/BUILD index 9719dcb..c62c274 100644 --- a/tests/cocotb/rvv/load_store/BUILD +++ b/tests/cocotb/rvv/load_store/BUILD
@@ -83,6 +83,9 @@ "load16_index16_seg": { "srcs": ["load16_index16_seg.cc"], }, + "load16_index32_seg": { + "srcs": ["load16_index32_seg.cc"], + }, "load16_seg_unit": { "srcs": ["load16_seg_unit.cc"], }, @@ -170,6 +173,7 @@ ":load16_index8.elf", ":load16_index8_seg.elf", ":load16_index16_seg.elf", + ":load16_index32_seg.elf", ":load16_seg_unit.elf", ":load16_stride4_m1.elf", ":load16_stride4_m1_partial.elf",
diff --git a/tests/cocotb/rvv/load_store/load16_index32_seg.cc b/tests/cocotb/rvv/load_store/load16_index32_seg.cc new file mode 100644 index 0000000..271d541 --- /dev/null +++ b/tests/cocotb/rvv/load_store/load16_index32_seg.cc
@@ -0,0 +1,593 @@ +// Copyright 2025 Google LLC +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +#include <riscv_vector.h> +#include <stdint.h> + +namespace { +constexpr size_t lut_size = 15000; +// Double sized so we can check trailing regions are not read/written. +constexpr size_t buf_size = 128; +} // namespace + +size_t vl __attribute__((section(".data"))) = 8; +// Indices are always unsigned. +uint32_t index_buf[buf_size] __attribute__((section(".data"))); +// These instructions don't differentiate signed/unsigned so we only need to +// test one. The types come from intrinsic level. +uint16_t in_buf[lut_size] __attribute__((section(".data"))); +uint16_t out_buf[buf_size] __attribute__((section(".data"))); + +extern "C" { +// Unordered, segment 2 +__attribute__((used, retain)) void vluxseg2ei32_v_u16mf2x2() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg2ei32_v_u16mf2x2(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x2_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x2_u16mf2(data, 1), + vl); +} + +__attribute__((used, retain)) void vluxseg2ei32_v_u16m1x2() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg2ei32_v_u16m1x2(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x2_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x2_u16m1(data, 1), + vl); +} + +__attribute__((used, retain)) void vluxseg2ei32_v_u16m2x2() { + auto indices = __riscv_vle32_v_u32m4(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg2ei32_v_u16m2x2(in_buf, indices, vl); + __riscv_vse16_v_u16m2(out_buf, __riscv_vget_v_u16m2x2_u16m2(data, 0), vl); + __riscv_vse16_v_u16m2(out_buf + vl, __riscv_vget_v_u16m2x2_u16m2(data, 1), + vl); +} + +__attribute__((used, retain)) void vluxseg2ei32_v_u16m4x2() { + auto indices = __riscv_vle32_v_u32m8(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg2ei32_v_u16m4x2(in_buf, indices, vl); + __riscv_vse16_v_u16m4(out_buf, __riscv_vget_v_u16m4x2_u16m4(data, 0), vl); + __riscv_vse16_v_u16m4(out_buf + vl, __riscv_vget_v_u16m4x2_u16m4(data, 1), + vl); +} + +// Unordered, segment 3 +__attribute__((used, retain)) void vluxseg3ei32_v_u16mf2x3() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg3ei32_v_u16mf2x3(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x3_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x3_u16mf2(data, 1), + vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 2, + __riscv_vget_v_u16mf2x3_u16mf2(data, 2), vl); +} + +__attribute__((used, retain)) void vluxseg3ei32_v_u16m1x3() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg3ei32_v_u16m1x3(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x3_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x3_u16m1(data, 1), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 2, __riscv_vget_v_u16m1x3_u16m1(data, 2), + vl); +} + +__attribute__((used, retain)) void vluxseg3ei32_v_u16m2x3() { + auto indices = __riscv_vle32_v_u32m4(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg3ei32_v_u16m2x3(in_buf, indices, vl); + __riscv_vse16_v_u16m2(out_buf, __riscv_vget_v_u16m2x3_u16m2(data, 0), vl); + __riscv_vse16_v_u16m2(out_buf + vl, __riscv_vget_v_u16m2x3_u16m2(data, 1), + vl); + __riscv_vse16_v_u16m2(out_buf + vl * 2, __riscv_vget_v_u16m2x3_u16m2(data, 2), + vl); +} + +// Unordered, segment 4 +__attribute__((used, retain)) void vluxseg4ei32_v_u16mf2x4() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg4ei32_v_u16mf2x4(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x4_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x4_u16mf2(data, 1), + vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 2, + __riscv_vget_v_u16mf2x4_u16mf2(data, 2), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 3, + __riscv_vget_v_u16mf2x4_u16mf2(data, 3), vl); +} + +__attribute__((used, retain)) void vluxseg4ei32_v_u16m1x4() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg4ei32_v_u16m1x4(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x4_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x4_u16m1(data, 1), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 2, __riscv_vget_v_u16m1x4_u16m1(data, 2), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 3, __riscv_vget_v_u16m1x4_u16m1(data, 3), + vl); +} + +__attribute__((used, retain)) void vluxseg4ei32_v_u16m2x4() { + auto indices = __riscv_vle32_v_u32m4(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg4ei32_v_u16m2x4(in_buf, indices, vl); + __riscv_vse16_v_u16m2(out_buf, __riscv_vget_v_u16m2x4_u16m2(data, 0), vl); + __riscv_vse16_v_u16m2(out_buf + vl, __riscv_vget_v_u16m2x4_u16m2(data, 1), + vl); + __riscv_vse16_v_u16m2(out_buf + vl * 2, __riscv_vget_v_u16m2x4_u16m2(data, 2), + vl); + __riscv_vse16_v_u16m2(out_buf + vl * 3, __riscv_vget_v_u16m2x4_u16m2(data, 3), + vl); +} + +// Unordered, segment 5 +__attribute__((used, retain)) void vluxseg5ei32_v_u16mf2x5() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg5ei32_v_u16mf2x5(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x5_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x5_u16mf2(data, 1), + vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 2, + __riscv_vget_v_u16mf2x5_u16mf2(data, 2), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 3, + __riscv_vget_v_u16mf2x5_u16mf2(data, 3), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 4, + __riscv_vget_v_u16mf2x5_u16mf2(data, 4), vl); +} + +__attribute__((used, retain)) void vluxseg5ei32_v_u16m1x5() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg5ei32_v_u16m1x5(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x5_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x5_u16m1(data, 1), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 2, __riscv_vget_v_u16m1x5_u16m1(data, 2), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 3, __riscv_vget_v_u16m1x5_u16m1(data, 3), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 4, __riscv_vget_v_u16m1x5_u16m1(data, 4), + vl); +} + +// Unordered, segment 6 +__attribute__((used, retain)) void vluxseg6ei32_v_u16mf2x6() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg6ei32_v_u16mf2x6(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x6_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x6_u16mf2(data, 1), + vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 2, + __riscv_vget_v_u16mf2x6_u16mf2(data, 2), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 3, + __riscv_vget_v_u16mf2x6_u16mf2(data, 3), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 4, + __riscv_vget_v_u16mf2x6_u16mf2(data, 4), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 5, + __riscv_vget_v_u16mf2x6_u16mf2(data, 5), vl); +} + +__attribute__((used, retain)) void vluxseg6ei32_v_u16m1x6() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg6ei32_v_u16m1x6(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x6_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x6_u16m1(data, 1), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 2, __riscv_vget_v_u16m1x6_u16m1(data, 2), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 3, __riscv_vget_v_u16m1x6_u16m1(data, 3), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 4, __riscv_vget_v_u16m1x6_u16m1(data, 4), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 5, __riscv_vget_v_u16m1x6_u16m1(data, 5), + vl); +} + +// Unordered, segment 7 +__attribute__((used, retain)) void vluxseg7ei32_v_u16mf2x7() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg7ei32_v_u16mf2x7(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x7_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x7_u16mf2(data, 1), + vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 2, + __riscv_vget_v_u16mf2x7_u16mf2(data, 2), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 3, + __riscv_vget_v_u16mf2x7_u16mf2(data, 3), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 4, + __riscv_vget_v_u16mf2x7_u16mf2(data, 4), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 5, + __riscv_vget_v_u16mf2x7_u16mf2(data, 5), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 6, + __riscv_vget_v_u16mf2x7_u16mf2(data, 6), vl); +} + +__attribute__((used, retain)) void vluxseg7ei32_v_u16m1x7() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg7ei32_v_u16m1x7(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x7_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x7_u16m1(data, 1), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 2, __riscv_vget_v_u16m1x7_u16m1(data, 2), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 3, __riscv_vget_v_u16m1x7_u16m1(data, 3), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 4, __riscv_vget_v_u16m1x7_u16m1(data, 4), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 5, __riscv_vget_v_u16m1x7_u16m1(data, 5), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 6, __riscv_vget_v_u16m1x7_u16m1(data, 6), + vl); +} + +// Unordered, segment 8 +__attribute__((used, retain)) void vluxseg8ei32_v_u16mf2x8() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg8ei32_v_u16mf2x8(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x8_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x8_u16mf2(data, 1), + vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 2, + __riscv_vget_v_u16mf2x8_u16mf2(data, 2), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 3, + __riscv_vget_v_u16mf2x8_u16mf2(data, 3), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 4, + __riscv_vget_v_u16mf2x8_u16mf2(data, 4), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 5, + __riscv_vget_v_u16mf2x8_u16mf2(data, 5), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 6, + __riscv_vget_v_u16mf2x8_u16mf2(data, 6), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 7, + __riscv_vget_v_u16mf2x8_u16mf2(data, 7), vl); +} + +__attribute__((used, retain)) void vluxseg8ei32_v_u16m1x8() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vluxseg8ei32_v_u16m1x8(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x8_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x8_u16m1(data, 1), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 2, __riscv_vget_v_u16m1x8_u16m1(data, 2), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 3, __riscv_vget_v_u16m1x8_u16m1(data, 3), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 4, __riscv_vget_v_u16m1x8_u16m1(data, 4), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 5, __riscv_vget_v_u16m1x8_u16m1(data, 5), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 6, __riscv_vget_v_u16m1x8_u16m1(data, 6), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 7, __riscv_vget_v_u16m1x8_u16m1(data, 7), + vl); +} + +// Ordered, segment 2 +__attribute__((used, retain)) void vloxseg2ei32_v_u16mf2x2() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg2ei32_v_u16mf2x2(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x2_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x2_u16mf2(data, 1), + vl); +} + +__attribute__((used, retain)) void vloxseg2ei32_v_u16m1x2() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg2ei32_v_u16m1x2(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x2_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x2_u16m1(data, 1), + vl); +} + +__attribute__((used, retain)) void vloxseg2ei32_v_u16m2x2() { + auto indices = __riscv_vle32_v_u32m4(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg2ei32_v_u16m2x2(in_buf, indices, vl); + __riscv_vse16_v_u16m2(out_buf, __riscv_vget_v_u16m2x2_u16m2(data, 0), vl); + __riscv_vse16_v_u16m2(out_buf + vl, __riscv_vget_v_u16m2x2_u16m2(data, 1), + vl); +} + +__attribute__((used, retain)) void vloxseg2ei32_v_u16m4x2() { + auto indices = __riscv_vle32_v_u32m8(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg2ei32_v_u16m4x2(in_buf, indices, vl); + __riscv_vse16_v_u16m4(out_buf, __riscv_vget_v_u16m4x2_u16m4(data, 0), vl); + __riscv_vse16_v_u16m4(out_buf + vl, __riscv_vget_v_u16m4x2_u16m4(data, 1), + vl); +} + +// Ordered, segment 3 +__attribute__((used, retain)) void vloxseg3ei32_v_u16mf2x3() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg3ei32_v_u16mf2x3(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x3_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x3_u16mf2(data, 1), + vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 2, + __riscv_vget_v_u16mf2x3_u16mf2(data, 2), vl); +} + +__attribute__((used, retain)) void vloxseg3ei32_v_u16m1x3() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg3ei32_v_u16m1x3(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x3_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x3_u16m1(data, 1), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 2, __riscv_vget_v_u16m1x3_u16m1(data, 2), + vl); +} + +__attribute__((used, retain)) void vloxseg3ei32_v_u16m2x3() { + auto indices = __riscv_vle32_v_u32m4(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg3ei32_v_u16m2x3(in_buf, indices, vl); + __riscv_vse16_v_u16m2(out_buf, __riscv_vget_v_u16m2x3_u16m2(data, 0), vl); + __riscv_vse16_v_u16m2(out_buf + vl, __riscv_vget_v_u16m2x3_u16m2(data, 1), + vl); + __riscv_vse16_v_u16m2(out_buf + vl * 2, __riscv_vget_v_u16m2x3_u16m2(data, 2), + vl); +} + +// Ordered, segment 4 +__attribute__((used, retain)) void vloxseg4ei32_v_u16mf2x4() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg4ei32_v_u16mf2x4(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x4_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x4_u16mf2(data, 1), + vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 2, + __riscv_vget_v_u16mf2x4_u16mf2(data, 2), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 3, + __riscv_vget_v_u16mf2x4_u16mf2(data, 3), vl); +} + +__attribute__((used, retain)) void vloxseg4ei32_v_u16m1x4() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg4ei32_v_u16m1x4(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x4_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x4_u16m1(data, 1), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 2, __riscv_vget_v_u16m1x4_u16m1(data, 2), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 3, __riscv_vget_v_u16m1x4_u16m1(data, 3), + vl); +} + +__attribute__((used, retain)) void vloxseg4ei32_v_u16m2x4() { + auto indices = __riscv_vle32_v_u32m4(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg4ei32_v_u16m2x4(in_buf, indices, vl); + __riscv_vse16_v_u16m2(out_buf, __riscv_vget_v_u16m2x4_u16m2(data, 0), vl); + __riscv_vse16_v_u16m2(out_buf + vl, __riscv_vget_v_u16m2x4_u16m2(data, 1), + vl); + __riscv_vse16_v_u16m2(out_buf + vl * 2, __riscv_vget_v_u16m2x4_u16m2(data, 2), + vl); + __riscv_vse16_v_u16m2(out_buf + vl * 3, __riscv_vget_v_u16m2x4_u16m2(data, 3), + vl); +} + +// Ordered, segment 5 +__attribute__((used, retain)) void vloxseg5ei32_v_u16mf2x5() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg5ei32_v_u16mf2x5(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x5_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x5_u16mf2(data, 1), + vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 2, + __riscv_vget_v_u16mf2x5_u16mf2(data, 2), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 3, + __riscv_vget_v_u16mf2x5_u16mf2(data, 3), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 4, + __riscv_vget_v_u16mf2x5_u16mf2(data, 4), vl); +} + +__attribute__((used, retain)) void vloxseg5ei32_v_u16m1x5() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg5ei32_v_u16m1x5(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x5_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x5_u16m1(data, 1), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 2, __riscv_vget_v_u16m1x5_u16m1(data, 2), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 3, __riscv_vget_v_u16m1x5_u16m1(data, 3), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 4, __riscv_vget_v_u16m1x5_u16m1(data, 4), + vl); +} + +// Ordered, segment 6 +__attribute__((used, retain)) void vloxseg6ei32_v_u16mf2x6() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg6ei32_v_u16mf2x6(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x6_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x6_u16mf2(data, 1), + vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 2, + __riscv_vget_v_u16mf2x6_u16mf2(data, 2), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 3, + __riscv_vget_v_u16mf2x6_u16mf2(data, 3), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 4, + __riscv_vget_v_u16mf2x6_u16mf2(data, 4), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 5, + __riscv_vget_v_u16mf2x6_u16mf2(data, 5), vl); +} + +__attribute__((used, retain)) void vloxseg6ei32_v_u16m1x6() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg6ei32_v_u16m1x6(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x6_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x6_u16m1(data, 1), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 2, __riscv_vget_v_u16m1x6_u16m1(data, 2), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 3, __riscv_vget_v_u16m1x6_u16m1(data, 3), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 4, __riscv_vget_v_u16m1x6_u16m1(data, 4), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 5, __riscv_vget_v_u16m1x6_u16m1(data, 5), + vl); +} + +// Ordered, segment 7 +__attribute__((used, retain)) void vloxseg7ei32_v_u16mf2x7() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg7ei32_v_u16mf2x7(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x7_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x7_u16mf2(data, 1), + vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 2, + __riscv_vget_v_u16mf2x7_u16mf2(data, 2), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 3, + __riscv_vget_v_u16mf2x7_u16mf2(data, 3), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 4, + __riscv_vget_v_u16mf2x7_u16mf2(data, 4), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 5, + __riscv_vget_v_u16mf2x7_u16mf2(data, 5), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 6, + __riscv_vget_v_u16mf2x7_u16mf2(data, 6), vl); +} + +__attribute__((used, retain)) void vloxseg7ei32_v_u16m1x7() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg7ei32_v_u16m1x7(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x7_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x7_u16m1(data, 1), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 2, __riscv_vget_v_u16m1x7_u16m1(data, 2), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 3, __riscv_vget_v_u16m1x7_u16m1(data, 3), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 4, __riscv_vget_v_u16m1x7_u16m1(data, 4), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 5, __riscv_vget_v_u16m1x7_u16m1(data, 5), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 6, __riscv_vget_v_u16m1x7_u16m1(data, 6), + vl); +} + +// Unordered, segment 8 +__attribute__((used, retain)) void vloxseg8ei32_v_u16mf2x8() { + auto indices = __riscv_vle32_v_u32m1(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg8ei32_v_u16mf2x8(in_buf, indices, vl); + __riscv_vse16_v_u16mf2(out_buf, __riscv_vget_v_u16mf2x8_u16mf2(data, 0), vl); + __riscv_vse16_v_u16mf2(out_buf + vl, __riscv_vget_v_u16mf2x8_u16mf2(data, 1), + vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 2, + __riscv_vget_v_u16mf2x8_u16mf2(data, 2), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 3, + __riscv_vget_v_u16mf2x8_u16mf2(data, 3), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 4, + __riscv_vget_v_u16mf2x8_u16mf2(data, 4), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 5, + __riscv_vget_v_u16mf2x8_u16mf2(data, 5), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 6, + __riscv_vget_v_u16mf2x8_u16mf2(data, 6), vl); + __riscv_vse16_v_u16mf2(out_buf + vl * 7, + __riscv_vget_v_u16mf2x8_u16mf2(data, 7), vl); +} + +__attribute__((used, retain)) void vloxseg8ei32_v_u16m1x8() { + auto indices = __riscv_vle32_v_u32m2(index_buf, vl); + // TODO(davidgao): Remove once compiler bug is eliminated + asm volatile("" ::: "vtype"); + auto data = __riscv_vloxseg8ei32_v_u16m1x8(in_buf, indices, vl); + __riscv_vse16_v_u16m1(out_buf, __riscv_vget_v_u16m1x8_u16m1(data, 0), vl); + __riscv_vse16_v_u16m1(out_buf + vl, __riscv_vget_v_u16m1x8_u16m1(data, 1), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 2, __riscv_vget_v_u16m1x8_u16m1(data, 2), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 3, __riscv_vget_v_u16m1x8_u16m1(data, 3), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 4, __riscv_vget_v_u16m1x8_u16m1(data, 4), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 5, __riscv_vget_v_u16m1x8_u16m1(data, 5), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 6, __riscv_vget_v_u16m1x8_u16m1(data, 6), + vl); + __riscv_vse16_v_u16m1(out_buf + vl * 7, __riscv_vget_v_u16m1x8_u16m1(data, 7), + vl); +} +} + +void (*impl)() __attribute__((section(".data"))) = &vluxseg2ei32_v_u16m1x2; + +int main(int argc, char** argv) { + impl(); + return 0; +}
diff --git a/tests/cocotb/rvv_load_store_test.py b/tests/cocotb/rvv_load_store_test.py index 39094b7..00c499a 100644 --- a/tests/cocotb/rvv_load_store_test.py +++ b/tests/cocotb/rvv_load_store_test.py
@@ -1526,6 +1526,114 @@ @cocotb.test() +async def load16_index32_seg(dut): + """Test vl*xseg*ei32_v_u16 usage accessible from intrinsics.""" + def make_test_case(impl: str, vl: int, n_segs: int): + return { + 'impl': impl, + 'vl': vl, + 'segments': n_segs, + 'in_bytes': 30000, + 'out_size': vl * n_segs * 2, + } + + await vector_load_segmented_indexed( + dut = dut, + elf_name = 'load16_index32_seg.elf', + cases = [ + # Unordered, segment 2 + make_test_case('vluxseg2ei32_v_u16mf2x2', vl=4, n_segs=2), + make_test_case('vluxseg2ei32_v_u16mf2x2', vl=3, n_segs=2), + make_test_case('vluxseg2ei32_v_u16m1x2', vl=8, n_segs=2), + make_test_case('vluxseg2ei32_v_u16m1x2', vl=7, n_segs=2), + make_test_case('vluxseg2ei32_v_u16m2x2', vl=16, n_segs=2), + make_test_case('vluxseg2ei32_v_u16m2x2', vl=15, n_segs=2), + # make_test_case('vluxseg2ei32_v_u16m4x2', vl=32, n_segs=2), + # make_test_case('vluxseg2ei32_v_u16m4x2', vl=31, n_segs=2), + # Unordered, segment 3 + make_test_case('vluxseg3ei32_v_u16mf2x3', vl=4, n_segs=3), + make_test_case('vluxseg3ei32_v_u16mf2x3', vl=3, n_segs=3), + make_test_case('vluxseg3ei32_v_u16m1x3', vl=8, n_segs=3), + make_test_case('vluxseg3ei32_v_u16m1x3', vl=7, n_segs=3), + # make_test_case('vluxseg3ei32_v_u16m2x3', vl=16, n_segs=3), + # make_test_case('vluxseg3ei32_v_u16m2x3', vl=15, n_segs=3), + # Unordered, segment 4 + make_test_case('vluxseg4ei32_v_u16mf2x4', vl=4, n_segs=4), + make_test_case('vluxseg4ei32_v_u16mf2x4', vl=3, n_segs=4), + make_test_case('vluxseg4ei32_v_u16m1x4', vl=8, n_segs=4), + make_test_case('vluxseg4ei32_v_u16m1x4', vl=7, n_segs=4), + # make_test_case('vluxseg4ei32_v_u16m2x4', vl=16, n_segs=4), + # make_test_case('vloxseg4ei32_v_u16m2x4', vl=15, n_segs=4), + # Unordered, segment 5 + make_test_case('vluxseg5ei32_v_u16mf2x5', vl=4, n_segs=5), + make_test_case('vluxseg5ei32_v_u16mf2x5', vl=3, n_segs=5), + #make_test_case('vluxseg5ei32_v_u16m1x5', vl=8, n_segs=5), + #make_test_case('vluxseg5ei32_v_u16m1x5', vl=7, n_segs=5), + # Unordered, segment 6 + make_test_case('vluxseg6ei32_v_u16mf2x6', vl=4, n_segs=6), + make_test_case('vluxseg6ei32_v_u16mf2x6', vl=3, n_segs=6), + # make_test_case('vluxseg6ei32_v_u16m1x6', vl=8, n_segs=6), + # make_test_case('vluxseg6ei32_v_u16m1x6', vl=7, n_segs=6), + # Unordered, segment 7 + make_test_case('vluxseg7ei32_v_u16mf2x7', vl=4, n_segs=7), + make_test_case('vluxseg7ei32_v_u16mf2x7', vl=3, n_segs=7), + # make_test_case('vluxseg7ei32_v_u16m1x7', vl=8, n_segs=7), + # make_test_case('vluxseg7ei32_v_u16m1x7', vl=7, n_segs=7), + # Unordered, segment 8 + make_test_case('vluxseg8ei32_v_u16mf2x8', vl=4, n_segs=8), + make_test_case('vluxseg8ei32_v_u16mf2x8', vl=3, n_segs=8), + # make_test_case('vluxseg8ei32_v_u16m1x8', vl=8, n_segs=8), + # make_test_case('vluxseg8ei32_v_u16m1x8', vl=7, n_segs=8), + # Ordered, segment 2 + make_test_case('vloxseg2ei32_v_u16mf2x2', vl=4, n_segs=2), + make_test_case('vloxseg2ei32_v_u16mf2x2', vl=3, n_segs=2), + make_test_case('vloxseg2ei32_v_u16m1x2', vl=8, n_segs=2), + make_test_case('vloxseg2ei32_v_u16m1x2', vl=7, n_segs=2), + make_test_case('vloxseg2ei32_v_u16m2x2', vl=16, n_segs=2), + make_test_case('vloxseg2ei32_v_u16m2x2', vl=15, n_segs=2), + # make_test_case('vloxseg2ei32_v_u16m4x2', vl=32, n_segs=2), + # make_test_case('vloxseg2ei32_v_u16m4x2', vl=31, n_segs=2), + # Ordered, segment 3 + make_test_case('vloxseg3ei32_v_u16mf2x3', vl=4, n_segs=3), + make_test_case('vloxseg3ei32_v_u16mf2x3', vl=3, n_segs=3), + make_test_case('vloxseg3ei32_v_u16m1x3', vl=8, n_segs=3), + make_test_case('vloxseg3ei32_v_u16m1x3', vl=7, n_segs=3), + # make_test_case('vloxseg3ei32_v_u16m2x3', vl=16, n_segs=3), + # make_test_case('vloxseg3ei32_v_u16m2x3', vl=15, n_segs=3), + # Ordered, segment 4 + make_test_case('vloxseg4ei32_v_u16mf2x4', vl=4, n_segs=4), + make_test_case('vloxseg4ei32_v_u16mf2x4', vl=3, n_segs=4), + make_test_case('vloxseg4ei32_v_u16m1x4', vl=8, n_segs=4), + make_test_case('vloxseg4ei32_v_u16m1x4', vl=7, n_segs=4), + # make_test_case('vloxseg4ei32_v_u16m2x4', vl=16, n_segs=4), + # make_test_case('vloxseg4ei32_v_u16m2x4', vl=15, n_segs=4), + # Ordered, segment 5 + make_test_case('vloxseg5ei32_v_u16mf2x5', vl=4, n_segs=5), + make_test_case('vloxseg5ei32_v_u16mf2x5', vl=3, n_segs=5), + #make_test_case('vloxseg5ei32_v_u16m1x5', vl=8, n_segs=5), + #make_test_case('vloxseg5ei32_v_u16m1x5', vl=7, n_segs=5), + # Ordered, segment 6 + make_test_case('vloxseg6ei32_v_u16mf2x6', vl=4, n_segs=6), + make_test_case('vloxseg6ei32_v_u16mf2x6', vl=3, n_segs=6), + # make_test_case('vloxseg6ei32_v_u16m1x6', vl=8, n_segs=6), + # make_test_case('vloxseg6ei32_v_u16m1x6', vl=7, n_segs=6), + # Ordered, segment 7 + make_test_case('vloxseg7ei32_v_u16mf2x7', vl=4, n_segs=7), + make_test_case('vloxseg7ei32_v_u16mf2x7', vl=3, n_segs=7), + # make_test_case('vloxseg7ei32_v_u16m1x7', vl=8, n_segs=7), + # make_test_case('vloxseg7ei32_v_u16m1x7', vl=7, n_segs=7), + # Ordered, segment 8 + make_test_case('vloxseg8ei32_v_u16mf2x8', vl=4, n_segs=8), + make_test_case('vloxseg8ei32_v_u16mf2x8', vl=3, n_segs=8), + #make_test_case('vloxseg8ei32_v_u16m1x8', vl=8, n_segs=8), + #make_test_case('vloxseg8ei32_v_u16m1x8', vl=7, n_segs=8), + ], + dtype = np.uint16, + index_dtype = np.uint32, + ) + + +@cocotb.test() async def load16_seg_unit(dut): """Test vlseg*e16 usage accessible from intrinsics.""" def make_test_case(impl: str, vl: int, n_segs: int):