| // Copyright (c) Facebook, Inc. and its affiliates. |
| // All rights reserved. |
| // |
| // Copyright 2019 Google LLC |
| // |
| // This source code is licensed under the BSD-style license found in the |
| // LICENSE file in the root directory of this source tree. |
| // |
| // Auto-generated file. Do not edit! |
| // Specification: test/q8-igemm.yaml |
| // Generator: tools/generate-gemm-test.py |
| |
| |
| #include <gtest/gtest.h> |
| |
| #include <xnnpack/common.h> |
| #include <xnnpack/isa-checks.h> |
| |
| #include <xnnpack/gemm.h> |
| #include <xnnpack/igemm.h> |
| #include <xnnpack/ppmm.h> |
| #include "gemm-microkernel-tester.h" |
| |
| |
| #if XNN_ARCH_ARM || XNN_ARCH_ARM64 |
| TEST(Q8_IGEMM_4X8__NEON, k_eq_8) { |
| TEST_REQUIRES_ARM_NEON; |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(8) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, strided_cn) { |
| TEST_REQUIRES_ARM_NEON; |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(8) |
| .cn_stride(11) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, k_eq_8_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t m = 1; m <= 4; m++) { |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(8) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, k_eq_8_subtile_m) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t m = 1; m <= 4; m++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(8) |
| .k(8) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, k_eq_8_subtile_n) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(n) |
| .k(8) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, k_lt_8) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k < 8; k++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, k_lt_8_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k < 8; k++) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, k_gt_8) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 9; k < 16; k++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, k_gt_8_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 9; k < 16; k++) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, k_div_8) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 16; k <= 80; k += 8) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, k_div_8_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 16; k <= 80; k += 8) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, n_gt_8) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 9; n < 16; n++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, n_gt_8_strided_cn) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 9; n < 16; n++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .cn_stride(11) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, n_gt_8_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 9; n < 16; n++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, n_div_8) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 16; n <= 24; n += 8) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, n_div_8_strided_cn) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 16; n <= 24; n += 8) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(n) |
| .k(k) |
| .cn_stride(11) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, n_div_8_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 16; n <= 24; n += 8) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, small_kernel) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .ks(3) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, small_kernel_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .ks(3) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, n_gt_8_small_kernel) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 9; n < 16; n++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .ks(3) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, n_div_8_small_kernel) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 16; n <= 24; n += 8) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .ks(3) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, strided_cm_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .cm_stride(11) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, a_offset) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .ks(3) |
| .a_offset(163) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, zero) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t mz = 0; mz < 4; mz++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .ks(3) |
| .a_offset(163) |
| .zero_index(mz) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, qmin) { |
| TEST_REQUIRES_ARM_NEON; |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(8) |
| .qmin(128) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, qmax) { |
| TEST_REQUIRES_ARM_NEON; |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(8) |
| .qmax(128) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, strided_cm) { |
| TEST_REQUIRES_ARM_NEON; |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(8) |
| .cm_stride(11) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, no_a_zero_point) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .a_zero_point(0) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, no_b_zero_point) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .b_zero_point(0) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X8__NEON, no_zero_point) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(4) |
| .n(8) |
| .k(k) |
| .a_zero_point(0) |
| .b_zero_point(0) |
| .Test(xnn_q8_igemm_ukernel_4x8__neon); |
| } |
| } |
| #endif // XNN_ARCH_ARM || XNN_ARCH_ARM64 |
| |
| |
| #if XNN_ARCH_ARM || XNN_ARCH_ARM64 |
| TEST(Q8_IGEMM_8X8__NEON, k_eq_8) { |
| TEST_REQUIRES_ARM_NEON; |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(8) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, strided_cn) { |
| TEST_REQUIRES_ARM_NEON; |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(8) |
| .cn_stride(11) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, k_eq_8_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t m = 1; m <= 8; m++) { |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(8) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, k_eq_8_subtile_m) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t m = 1; m <= 8; m++) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(8) |
| .k(8) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, k_eq_8_subtile_n) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(n) |
| .k(8) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, k_lt_8) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k < 8; k++) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, k_lt_8_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k < 8; k++) { |
| for (uint32_t m = 1; m <= 8; m++) { |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, k_gt_8) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 9; k < 16; k++) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, k_gt_8_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 9; k < 16; k++) { |
| for (uint32_t m = 1; m <= 8; m++) { |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, k_div_8) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 16; k <= 80; k += 8) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, k_div_8_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 16; k <= 80; k += 8) { |
| for (uint32_t m = 1; m <= 8; m++) { |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, n_gt_8) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 9; n < 16; n++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, n_gt_8_strided_cn) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 9; n < 16; n++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .cn_stride(11) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, n_gt_8_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 9; n < 16; n++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| for (uint32_t m = 1; m <= 8; m++) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, n_div_8) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 16; n <= 24; n += 8) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, n_div_8_strided_cn) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 16; n <= 24; n += 8) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(n) |
| .k(k) |
| .cn_stride(11) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, n_div_8_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 16; n <= 24; n += 8) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| for (uint32_t m = 1; m <= 8; m++) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, small_kernel) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .ks(3) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, small_kernel_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| for (uint32_t m = 1; m <= 8; m++) { |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .ks(3) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, n_gt_8_small_kernel) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 9; n < 16; n++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .ks(3) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, n_div_8_small_kernel) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t n = 16; n <= 24; n += 8) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .ks(3) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, strided_cm_subtile) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| for (uint32_t m = 1; m <= 8; m++) { |
| for (uint32_t n = 1; n <= 8; n++) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .cm_stride(11) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, a_offset) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .ks(3) |
| .a_offset(331) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, zero) { |
| TEST_REQUIRES_ARM_NEON; |
| for (uint32_t mz = 0; mz < 8; mz++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .ks(3) |
| .a_offset(331) |
| .zero_index(mz) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, qmin) { |
| TEST_REQUIRES_ARM_NEON; |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(8) |
| .qmin(128) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, qmax) { |
| TEST_REQUIRES_ARM_NEON; |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(8) |
| .qmax(128) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, strided_cm) { |
| TEST_REQUIRES_ARM_NEON; |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(8) |
| .cm_stride(11) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, no_a_zero_point) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .a_zero_point(0) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, no_b_zero_point) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .b_zero_point(0) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| |
| TEST(Q8_IGEMM_8X8__NEON, no_zero_point) { |
| TEST_REQUIRES_ARM_NEON; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(8) |
| .nr(8) |
| .kr(1) |
| .sr(1) |
| .m(8) |
| .n(8) |
| .k(k) |
| .a_zero_point(0) |
| .b_zero_point(0) |
| .Test(xnn_q8_igemm_ukernel_8x8__neon); |
| } |
| } |
| #endif // XNN_ARCH_ARM || XNN_ARCH_ARM64 |
| |
| |
| #if XNN_ARCH_X86 || XNN_ARCH_X86_64 |
| TEST(Q8_IGEMM_4X4C2__SSE2, k_eq_8) { |
| TEST_REQUIRES_X86_SSE2; |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(8) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, strided_cn) { |
| TEST_REQUIRES_X86_SSE2; |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(8) |
| .cn_stride(7) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, k_eq_8_subtile) { |
| TEST_REQUIRES_X86_SSE2; |
| for (uint32_t m = 1; m <= 4; m++) { |
| for (uint32_t n = 1; n <= 4; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(8) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, k_eq_8_subtile_m) { |
| TEST_REQUIRES_X86_SSE2; |
| for (uint32_t m = 1; m <= 4; m++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(m) |
| .n(4) |
| .k(8) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, k_eq_8_subtile_n) { |
| TEST_REQUIRES_X86_SSE2; |
| for (uint32_t n = 1; n <= 4; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(n) |
| .k(8) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, k_lt_8) { |
| TEST_REQUIRES_X86_SSE2; |
| for (size_t k = 1; k < 8; k++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, k_lt_8_subtile) { |
| TEST_REQUIRES_X86_SSE2; |
| for (size_t k = 1; k < 8; k++) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| for (uint32_t n = 1; n <= 4; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, k_gt_8) { |
| TEST_REQUIRES_X86_SSE2; |
| for (size_t k = 9; k < 16; k++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, k_gt_8_subtile) { |
| TEST_REQUIRES_X86_SSE2; |
| for (size_t k = 9; k < 16; k++) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| for (uint32_t n = 1; n <= 4; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, k_div_8) { |
| TEST_REQUIRES_X86_SSE2; |
| for (size_t k = 16; k <= 80; k += 8) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, k_div_8_subtile) { |
| TEST_REQUIRES_X86_SSE2; |
| for (size_t k = 16; k <= 80; k += 8) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| for (uint32_t n = 1; n <= 4; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, n_gt_4) { |
| TEST_REQUIRES_X86_SSE2; |
| for (uint32_t n = 5; n < 8; n++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, n_gt_4_strided_cn) { |
| TEST_REQUIRES_X86_SSE2; |
| for (uint32_t n = 5; n < 8; n++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .cn_stride(7) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, n_gt_4_subtile) { |
| TEST_REQUIRES_X86_SSE2; |
| for (uint32_t n = 5; n < 8; n++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, n_div_4) { |
| TEST_REQUIRES_X86_SSE2; |
| for (uint32_t n = 8; n <= 12; n += 4) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, n_div_4_strided_cn) { |
| TEST_REQUIRES_X86_SSE2; |
| for (uint32_t n = 8; n <= 12; n += 4) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(n) |
| .k(k) |
| .cn_stride(7) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, n_div_4_subtile) { |
| TEST_REQUIRES_X86_SSE2; |
| for (uint32_t n = 8; n <= 12; n += 4) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, small_kernel) { |
| TEST_REQUIRES_X86_SSE2; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .ks(3) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, small_kernel_subtile) { |
| TEST_REQUIRES_X86_SSE2; |
| for (size_t k = 1; k <= 40; k += 9) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| for (uint32_t n = 1; n <= 4; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .ks(3) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, n_gt_4_small_kernel) { |
| TEST_REQUIRES_X86_SSE2; |
| for (uint32_t n = 5; n < 8; n++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .ks(3) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, n_div_4_small_kernel) { |
| TEST_REQUIRES_X86_SSE2; |
| for (uint32_t n = 8; n <= 12; n += 4) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .ks(3) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, strided_cm_subtile) { |
| TEST_REQUIRES_X86_SSE2; |
| for (size_t k = 1; k <= 40; k += 9) { |
| for (uint32_t m = 1; m <= 4; m++) { |
| for (uint32_t n = 1; n <= 4; n++) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .cm_stride(7) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, a_offset) { |
| TEST_REQUIRES_X86_SSE2; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .ks(3) |
| .a_offset(163) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, zero) { |
| TEST_REQUIRES_X86_SSE2; |
| for (uint32_t mz = 0; mz < 4; mz++) { |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .ks(3) |
| .a_offset(163) |
| .zero_index(mz) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, qmin) { |
| TEST_REQUIRES_X86_SSE2; |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(8) |
| .qmin(128) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, qmax) { |
| TEST_REQUIRES_X86_SSE2; |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(8) |
| .qmax(128) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, strided_cm) { |
| TEST_REQUIRES_X86_SSE2; |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(8) |
| .cm_stride(7) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, no_a_zero_point) { |
| TEST_REQUIRES_X86_SSE2; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .a_zero_point(0) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, no_b_zero_point) { |
| TEST_REQUIRES_X86_SSE2; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .b_zero_point(0) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| |
| TEST(Q8_IGEMM_4X4C2__SSE2, no_zero_point) { |
| TEST_REQUIRES_X86_SSE2; |
| for (size_t k = 1; k <= 40; k += 9) { |
| GemmMicrokernelTester() |
| .mr(4) |
| .nr(4) |
| .kr(2) |
| .sr(1) |
| .m(4) |
| .n(4) |
| .k(k) |
| .a_zero_point(0) |
| .b_zero_point(0) |
| .Test(xnn_q8_igemm_ukernel_4x4c2__sse2); |
| } |
| } |
| #endif // XNN_ARCH_X86 || XNN_ARCH_X86_64 |
| |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, k_eq_1) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(1) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, strided_cn) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(1) |
| .cn_stride(5) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, k_eq_1_subtile) { |
| for (uint32_t m = 1; m <= 2; m++) { |
| for (uint32_t n = 1; n <= 2; n++) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(1) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, k_eq_1_subtile_m) { |
| for (uint32_t m = 1; m <= 2; m++) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(2) |
| .k(1) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, k_eq_1_subtile_n) { |
| for (uint32_t n = 1; n <= 2; n++) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(n) |
| .k(1) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, k_gt_1) { |
| for (size_t k = 2; k < 10; k++) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, k_gt_1_subtile) { |
| for (size_t k = 2; k < 10; k++) { |
| for (uint32_t m = 1; m <= 2; m++) { |
| for (uint32_t n = 1; n <= 2; n++) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, n_gt_2) { |
| for (uint32_t n = 3; n < 4; n++) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, n_gt_2_strided_cn) { |
| for (uint32_t n = 3; n < 4; n++) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(k) |
| .cn_stride(5) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, n_gt_2_subtile) { |
| for (uint32_t n = 3; n < 4; n++) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| for (uint32_t m = 1; m <= 2; m++) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, n_div_2) { |
| for (uint32_t n = 4; n <= 6; n += 2) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(k) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, n_div_2_strided_cn) { |
| for (uint32_t n = 4; n <= 6; n += 2) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(n) |
| .k(k) |
| .cn_stride(5) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, n_div_2_subtile) { |
| for (uint32_t n = 4; n <= 6; n += 2) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| for (uint32_t m = 1; m <= 2; m++) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, small_kernel) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(k) |
| .ks(3) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, small_kernel_subtile) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| for (uint32_t m = 1; m <= 2; m++) { |
| for (uint32_t n = 1; n <= 2; n++) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .ks(3) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, n_gt_2_small_kernel) { |
| for (uint32_t n = 3; n < 4; n++) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(k) |
| .ks(3) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, n_div_2_small_kernel) { |
| for (uint32_t n = 4; n <= 6; n += 2) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(k) |
| .ks(3) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, strided_cm_subtile) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| for (uint32_t m = 1; m <= 2; m++) { |
| for (uint32_t n = 1; n <= 2; n++) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(m) |
| .n(n) |
| .k(k) |
| .cm_stride(5) |
| .iterations(1) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, a_offset) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(k) |
| .ks(3) |
| .a_offset(13) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, zero) { |
| for (uint32_t mz = 0; mz < 2; mz++) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(k) |
| .ks(3) |
| .a_offset(13) |
| .zero_index(mz) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, qmin) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(1) |
| .qmin(128) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, qmax) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(1) |
| .qmax(128) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, strided_cm) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(1) |
| .cm_stride(5) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, no_a_zero_point) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(k) |
| .a_zero_point(0) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, no_b_zero_point) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(k) |
| .b_zero_point(0) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |
| |
| TEST(Q8_IGEMM_2X2__SCALAR, no_zero_point) { |
| for (size_t k = 1; k <= 5; k += 2) { |
| GemmMicrokernelTester() |
| .mr(2) |
| .nr(2) |
| .kr(1) |
| .sr(1) |
| .m(2) |
| .n(2) |
| .k(k) |
| .a_zero_point(0) |
| .b_zero_point(0) |
| .Test(xnn_q8_igemm_ukernel_2x2__scalar, GemmMicrokernelTester::Variant::Scalar); |
| } |
| } |