derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
79 lines
3.3 KiB
Plaintext
79 lines
3.3 KiB
Plaintext
# RUN: %stinkytofu-opt --arch gfx1250 %s --SetMatrixReusePass --print-output
|
|
#
|
|
# SetMatrixReusePass: set mod.mfma reuseA/reuseB from consecutive matrix insn operands;
|
|
# clear stale flags; skip non-scale f8f6f4 and MX f4 (rocisa gating).
|
|
|
|
#--- Same B (src1), different A: reuseB on first WMMA only
|
|
# CHECK-LABEL: @same_b_operand
|
|
# CHECK: v_wmma_f32_16x16x32_bf16
|
|
# CHECK-SAME: mod.mfma = { reuseA = false, reuseB = true }
|
|
# CHECK: v_wmma_f32_16x16x32_bf16
|
|
# CHECK-SAME: mod.mfma = { reuseA = false, reuseB = false }
|
|
|
|
st.func @same_b_operand() {
|
|
^entry:
|
|
a[0:7] = "st.v_wmma_f32_16x16x32_bf16"(v[8:15], v[16:23], a[0:7]) { mod.mfma = { reuseA = true, reuseB = true } }
|
|
a[32:39] = "st.v_wmma_f32_16x16x32_bf16"(v[24:31], v[16:23], a[32:39]) { mod.mfma = { reuseA = true, reuseB = true } }
|
|
}
|
|
|
|
#--- Non-matrix insn between WMMAs does not break reuse chain
|
|
# CHECK-LABEL: @non_matrix_between
|
|
# CHECK: v_wmma_f32_16x16x32_bf16
|
|
# CHECK-SAME: mod.mfma = { reuseA = true, reuseB = true }
|
|
# CHECK: ds_load_b128
|
|
# CHECK: v_wmma_f32_16x16x32_bf16
|
|
|
|
st.func @non_matrix_between() {
|
|
^entry:
|
|
a[0:7] = "st.v_wmma_f32_16x16x32_bf16"(v[8:15], v[16:23], a[0:7]) { mod.mfma = { reuseA = true, reuseB = true } }
|
|
v[64:67] = "st.ds_load_b128"(v[72]) { issueCycles = 1, latencyCycles = 56, mod.ds = { na = 1, offset = 0, gds = false }, mod.memtoken = { tokens = [0] } }
|
|
a[32:39] = "st.v_wmma_f32_16x16x32_bf16"(v[8:15], v[16:23], a[32:39]) { mod.mfma = { reuseA = true, reuseB = true } }
|
|
}
|
|
|
|
#--- Single WMMA: clear stale reuse flags
|
|
# CHECK-LABEL: @clears_stale_on_last
|
|
# CHECK: v_wmma_f32_16x16x32_bf16
|
|
# CHECK-SAME: mod.mfma = { reuseA = false, reuseB = false }
|
|
|
|
st.func @clears_stale_on_last() {
|
|
^entry:
|
|
a[0:7] = "st.v_wmma_f32_16x16x32_bf16"(v[8:15], v[16:23], a[0:7]) { mod.mfma = { reuseA = true, reuseB = true } }
|
|
}
|
|
|
|
#--- Non-scale WMMA f8f6f4: never set reuse
|
|
# CHECK-LABEL: @skip_f8f6f4
|
|
# CHECK: v_wmma_f32_16x16x128_f8f6f4
|
|
# CHECK-SAME: mod.mfma = { reuseA = false, reuseB = false }
|
|
# CHECK: v_wmma_f32_16x16x128_f8f6f4
|
|
# CHECK-SAME: mod.mfma = { reuseA = false, reuseB = false }
|
|
|
|
st.func @skip_f8f6f4() {
|
|
^entry:
|
|
a[0:7] = "st.v_wmma_f32_16x16x128_f8f6f4"(v[8:15], v[16:23], a[0:7]) { mod.mfma = { reuseA = true, reuseB = true } }
|
|
a[32:39] = "st.v_wmma_f32_16x16x128_f8f6f4"(v[24:31], v[16:23], a[32:39]) { mod.mfma = { reuseA = true, reuseB = true } }
|
|
}
|
|
|
|
#--- MX scale WMMA f4: never set reuse
|
|
# CHECK-LABEL: @skip_mx_f4_scale
|
|
# CHECK: v_wmma_scale_f32_32x16x128_f4
|
|
# CHECK-SAME: mod.mfma = { reuseA = false, reuseB = false }
|
|
|
|
st.func @skip_mx_f4_scale() {
|
|
^entry:
|
|
a[0:7] = "st.v_wmma_scale_f32_32x16x128_f4"(v[8:15], v[16:23], a[0:7]) { mod.mfma = { reuseA = true, reuseB = true } }
|
|
a[32:39] = "st.v_wmma_scale_f32_32x16x128_f4"(v[24:31], v[16:23], a[32:39]) { mod.mfma = { reuseA = true, reuseB = true } }
|
|
}
|
|
|
|
#--- MX scale WMMA f8f6f4: reuse allowed when operands match
|
|
# CHECK-LABEL: @allow_mx_f8f6f4_scale
|
|
# CHECK: v_wmma_scale_f32_16x16x128_f8f6f4
|
|
# CHECK-SAME: mod.mfma = { reuseA = false, reuseB = true }
|
|
# CHECK: v_wmma_scale_f32_16x16x128_f8f6f4
|
|
# CHECK-SAME: mod.mfma = { reuseA = false, reuseB = false }
|
|
|
|
st.func @allow_mx_f8f6f4_scale() {
|
|
^entry:
|
|
a[0:7] = "st.v_wmma_scale_f32_16x16x128_f8f6f4"(v[8:15], v[16:23], a[0:7]) { mod.mfma = { reuseA = true, reuseB = true } }
|
|
a[32:39] = "st.v_wmma_scale_f32_16x16x128_f8f6f4"(v[24:31], v[16:23], a[32:39]) { mod.mfma = { reuseA = true, reuseB = true } }
|
|
}
|