--- include: rocblas_common.yaml include: known_bugs.yaml Definitions: - &quick_matrix_size_range - { N: -1, K: 1, lda: 1, ldb: 1, ldc: 1 } # bad n - { N: 2, K: -1, lda: 2, ldb: 2, ldc: 1 } # bad k - { N: 0, K: 3, lda: 3, ldb: 3, ldc: 3 } # n==0 - { N: 3, K: 0, lda: 3, ldb: 3, ldc: 3 } # k==0 - { N: 3, K: 1, lda: 1, ldb: 3, ldc: 3 } # bad lda if not transpose - { N: 1, K: 3, lda: 1, ldb: 3, ldc: 3 } # bad lda if transpose - { N: 3, K: 1, lda: 3, ldb: 1, ldc: 3 } # bad ldb if not transpose - { N: 1, K: 3, lda: 3, ldb: 1, ldc: 3 } # bad ldb if transpose - { N: 3, K: 3, lda: 3, ldb: 3, ldc: 1 } # bad ldc - &tiny_matrix_size - { N: 199, lda: 199, K: 32, ldc: 199 } - &medium_matrix_size_range - { N: 199, K: 33, lda: 199, ldb: 199, ldc: 199 } - { N: 88, K: 200, lda: 200, ldb: 200, ldc: 88 } - { N: 64, K: 70, lda: 128, ldb: 128, ldc: 64 } - { N: 199, K: 290, lda: 290, ldb: 290, ldc: 200 } - &large_matrix_size_range - { N: 2011, K: 253, lda: 2011, ldb: 2011, ldc: 2048 } - { N: 1024, K: 1200, lda: 1200, ldb: 1200, ldc: 1024 } - &restrict_size_range_32_8 - { N: 32, K: 8, lda: 32, ldb: 32, ldc: 32 } - &restrict_size_range_16_16 - { N: 16, K: 16, lda: 16, ldb: 16, ldc: 16 } - &restrict_size_range_16_17 - { N: 16, K: 17, lda: 17, ldb: 17, ldc: 17 } - &general_size_range_17 - { N: 17, K: 17, lda: 17, ldb: 17, ldc: 17 } - &alpha_beta_range - { alpha: 1.5, alphai: 1.5, beta: 0.0, betai: 0.0 } - { alpha: -2.0, alphai: 1.0, beta: -1.0, betai: 0.5 } - { alpha: 0.0, alphai: 0.0, beta: 1.0, betai: 0.0 } # quick success - { alpha: 0.0, alphai: 0.0, beta: 2.0, betai: 0.5 } # scale step only - &alpha_beta - { alpha: -2.0, alphai: 1.5, beta: -0.5, betai: 0.5 } - &alpha_beta_restrict_32_8 - { alpha: 1.0, alphai: 0.0, beta: 1.0, betai: 0.0 } - { alpha: 1.0, alphai: 0.0, beta: -1.0, betai: 0.0 } - { alpha: 1.0, alphai: 0.0, beta: 0.0, betai: 0.0 } - { alpha: -1.0, alphai: 0.0, beta: 0.0, betai: 0.0 } - { alpha: 2.0, alphai: 0.0, beta: 0.0, betai: 0.0 } - { alpha: 2.0, alphai: 0.0, beta: 3.0, betai: 0.0 } - &alpha_beta_restrict_16 - { alpha: 1.0, alphai: 0.0, beta: 1.0, betai: 0.0 } - { alpha: 1.0, alphai: 0.0, beta: 0.0, betai: 0.0 } Tests: - name: syrkx_bad category: quick function: - syrkx_bad_arg - syrkx_batched_bad_arg - syrkx_strided_batched_bad_arg precision: *single_double_precisions_complex_real api: [ C, FORTRAN, FORTRAN_64 ] - name: syrkx_size_t category: stress function: syrkx precision: *single_precision uplo: [ U, L ] # op(A) and op(B) are NxK matrix_size: - { transA: N, N: 3, K: 46975, lda: 46975, ldb: 3, ldc: 3 } - { transA: N, N: 3, K: 46975, lda: 3, ldb: 46975, ldc: 3 } - { transA: T, N: 46975, K: 4, lda: 4, ldb: 4, ldc: 46975} alpha_beta: - { alpha: 0.5, beta: 0.5 } batch_count: [ 1 ] pointer_mode_host: false gpu_arch: ['90a','942','950'] os_flags: LINUX - name: syrkx_NaN category: pre_checkin function: syrkx precision: *single_double_precisions_complex uplo: [ U ] transA: [ N ] matrix_size: *tiny_matrix_size alpha: [ 2.0, .NaN ] # NaN is converted to 0.0 in test code beta: [ 1.0, .NaN ] - name: syrkx_restricted_kernel_32_8_real category: quick function: syrkx precision: *single_double_precisions uplo: [ U, L ] transA: [ N, C ] matrix_size: *restrict_size_range_32_8 alpha_beta: *alpha_beta_restrict_32_8 batch_count: [ 1 ] - name: syrkx_restricted_kernel_32_8_complex category: quick function: syrkx precision: *single_double_precisions_complex uplo: [ U, L ] transA: [ N, T ] matrix_size: *restrict_size_range_32_8 alpha_beta: *alpha_beta_restrict_32_8 batch_count: [ 1 ] - name: syrkx_restricted_kernel_16_16 category: quick function: syrkx precision: *single_double_precisions_complex_real uplo: [ U, L ] transA: [ N, T ] matrix_size: *restrict_size_range_16_16 alpha_beta: *alpha_beta_restrict_16 batch_count: [ 1 ] - name: syrkx_restricted_kernel_16 category: quick function: syrkx precision: *single_double_precisions_complex_real uplo: [ U, L ] transA: [ N, T ] matrix_size: *restrict_size_range_16_17 alpha_beta: *alpha_beta_restrict_16 batch_count: [ 1 ] - name: syrkx_general_kernel category: quick function: syrkx precision: *single_double_precisions_complex_real uplo: [ U, L ] transA: [ N, T ] matrix_size: *general_size_range_17 alpha_beta: *alpha_beta_restrict_16 batch_count: [ 1 ] - name: syrkx_quick category: quick function: syrkx precision: *single_precision uplo: [ U ] transA: [ N, T ] matrix_size: *quick_matrix_size_range alpha: [ 0, 1 ] beta: [ 0, 1 ] - name: syrkx_medium category: pre_checkin function: syrkx precision: *single_double_precisions_complex_real uplo: [ U, L ] transA: [ N, T ] matrix_size: *medium_matrix_size_range alpha_beta: *alpha_beta_range api: [ C, FORTRAN ] - name: syrkx_large category: nightly function: syrkx precision: *single_double_precisions_complex_real uplo: [ U ] transA: [ N, T ] matrix_size: *large_matrix_size_range alpha_beta: *alpha_beta pointer_mode_device: false # batched - name: syrkx_batched_quick category: quick function: syrkx_batched precision: *single_precision uplo: [ U ] transA: [ N, T ] matrix_size: *quick_matrix_size_range alpha: [ 0, 1 ] beta: [ 0, 1 ] batch_count: [ 0, 1 ] - name: syrkx_batched_NaN category: pre_checkin function: syrkx_batched precision: *single_precision uplo: [ L ] transA: [ T ] matrix_size: *tiny_matrix_size alpha: [ 2.0, .NaN ] # NaN is converted to 0.0 in test code beta: [ 1.0, .NaN ] batch_count: [ 2 ] - name: syrkx_batched_medium category: pre_checkin function: syrkx_batched precision: *single_double_precisions_complex_real uplo: [ U, L ] transA: [ N, T ] matrix_size: *medium_matrix_size_range alpha_beta: *alpha_beta_range batch_count: [ 1, 2 ] api: [ C, FORTRAN ] - name: syrkx_batched_large category: nightly function: syrkx_batched precision: *single_double_precisions_complex_real uplo: [ U ] transA: [ N, T ] matrix_size: *large_matrix_size_range alpha_beta: *alpha_beta batch_count: [ 2 ] pointer_mode_host: false # strided batched - name: syrkx_strided_batched_quick category: quick function: syrkx_strided_batched precision: *single_precision uplo: [ U ] transA: [ N, T ] matrix_size: *quick_matrix_size_range alpha: [ 0, 1 ] beta: [ 0, 1 ] batch_count: [ 0, 1 ] - name: syrkx_strided_batched_NaN category: pre_checkin function: syrkx_strided_batched precision: *double_precision uplo: [ U ] transA: [ T ] matrix_size: *tiny_matrix_size alpha: [ 2.0, .NaN ] # NaN is converted to 0.0 in test code beta: [ 1.0, .NaN ] batch_count: [ 3 ] - name: syrkx_strided_batched_medium category: pre_checkin function: syrkx_strided_batched precision: *single_double_precisions_complex_real uplo: [ U, L ] transA: [ N, T ] matrix_size: *medium_matrix_size_range alpha_beta: *alpha_beta_range batch_count: [ 2 ] api: [ C, FORTRAN ] - name: syrkx_strided_batched_large category: nightly function: syrkx_strided_batched precision: *single_double_precisions_complex_real uplo: [ L ] transA: [ N, T ] matrix_size: *large_matrix_size_range alpha_beta: *alpha_beta batch_count: [ 2 ] pointer_mode_host: false - name: syrkx_large_batch category: pre_checkin uplo: [ U, L ] arguments: - { N: 2, K: 2, lda: 2, ldb: 2, ldc: 2, batch_count: *c_grid_yz_require_passes, transA: N} alpha_beta: - { alpha: 1.0, beta: 1.0 } function: - syrkx_batched: *single_precision - syrkx_strided_batched: *single_precision stride_scale: 1 - name: syrkx_ILP64 category: stress uplo: [ U, L ] arguments: #- { N: 1, K: 2147483649, lda: 1, ldb: 1, ldc: 2147483649, batch_count: 1, transA: N } #TODO: Commenting out big tests due to alarm time out #- { N: 1, K: 2147483649, lda: 2147483649, ldb: 2147483649, ldc: 1, batch_count: 1, transA: T } #TODO: Commenting out big tests due to alarm time out - { N: 35, K: 2, lda: 2147483649, ldb: 36, ldc: 36, batch_count: 1, transA: N } - { N: 2, K: 35, lda: 36, ldb: 2147483649, ldc: 2147483649, batch_count: 1, transA: T } - { N: 2, K: 2, lda: 2, ldb: 2, ldc: 2, batch_count: *c_grid_yz_require_passes, transA: T} - { N: 2011, K: 253, lda: 2011, ldb: 2011, ldc: 2048, batch_count: 3, transA: N } - { N: 1024, K: 1200, lda: 1200, ldb: 1200, ldc: 1024, batch_count: 2, transA: T } api: [ C_64 ] alpha_beta: - { alpha: 1.0, beta: 0.0 } os_flags: [ LINUX ] function: - syrkx_strided_batched: *single_precision stride_scale: 1 gpu_arch: '9??' - name: syrkx_graph_test category: pre_checkin function: - syrkx - syrkx_batched - syrkx_strided_batched precision: *single_double_precisions_complex_real uplo: [ L ] transA: [ N, T ] matrix_size: - { N: 199, K: 290, lda: 290, ldb: 290, ldc: 200 } batch_count: [ 2 ] alpha_beta: *alpha_beta_range graph_test: true - name: syrkx_repeatability_check category: stress function: - syrkx - syrkx_batched - syrkx_strided_batched precision: *single_double_precisions_complex_real uplo: [ L ] transA: [ N, T ] matrix_size: - { N: 199, K: 290, lda: 290, ldb: 290, ldc: 200 } batch_count: [ 2 ] stride_scale: [ 1 ] initialization: hpl iters : 5 pointer_mode_host: false atomics_mode: 0 alpha_beta: *alpha_beta_range devices: 4 repeatability_check: true ...