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
197 lines
9.4 KiB
Plaintext
197 lines
9.4 KiB
Plaintext
// RUN: fir-opt -split-input-file -verify-diagnostics %s
|
|
|
|
func.func @_QPsub1() {
|
|
%0 = fir.alloca !fir.box<!fir.heap<!fir.array<?xf32>>> {bindc_name = "a", uniq_name = "_QFsub1Ea"}
|
|
%s = fir.alloca i64
|
|
%pinned = fir.alloca i1
|
|
%4:2 = hlfir.declare %0 {data_attr = #cuf.cuda<device>, fortran_attrs = #fir.var_attrs<allocatable>, uniq_name = "_QFsub1Ea"} : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>, !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>)
|
|
%11 = fir.convert %4#1 : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> !fir.ref<!fir.box<none>>
|
|
// expected-error@+1{{'cuf.allocate' op pinned and stream cannot appears at the same time}}
|
|
%13 = cuf.allocate %11 : !fir.ref<!fir.box<none>> stream(%s : !fir.ref<i64>) pinned(%pinned : !fir.ref<i1>) {data_attr = #cuf.cuda<device>} -> i32
|
|
return
|
|
}
|
|
|
|
// -----
|
|
|
|
func.func @_QPsub1() {
|
|
%1 = fir.alloca i32
|
|
// expected-error@+1{{'cuf.allocate' op expect box to be a reference to a class or box type value}}
|
|
%2 = cuf.allocate %1 : !fir.ref<i32> {data_attr = #cuf.cuda<device>} -> i32
|
|
return
|
|
}
|
|
|
|
// -----
|
|
|
|
func.func @_QPsub1() {
|
|
%0 = fir.alloca !fir.box<!fir.heap<!fir.array<?xf32>>> {bindc_name = "a", uniq_name = "_QFsub1Ea"}
|
|
%4:2 = hlfir.declare %0 {data_attr = #cuf.cuda<device>, fortran_attrs = #fir.var_attrs<allocatable>, uniq_name = "_QFsub1Ea"} : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>, !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>)
|
|
%c100 = arith.constant 100 : index
|
|
%7 = fir.alloca !fir.char<1,100> {bindc_name = "msg", uniq_name = "_QFsub1Emsg"}
|
|
%8:2 = hlfir.declare %7 typeparams %c100 {uniq_name = "_QFsub1Emsg"} : (!fir.ref<!fir.char<1,100>>, index) -> (!fir.ref<!fir.char<1,100>>, !fir.ref<!fir.char<1,100>>)
|
|
%9 = fir.embox %8#1 : (!fir.ref<!fir.char<1,100>>) -> !fir.box<!fir.char<1,100>>
|
|
%11 = fir.convert %4#1 : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> !fir.ref<!fir.box<none>>
|
|
%16 = fir.convert %9 : (!fir.box<!fir.char<1,100>>) -> !fir.box<none>
|
|
// expected-error@+1{{'cuf.allocate' op expect stat attribute when errmsg is provided}}
|
|
%13 = cuf.allocate %11 : !fir.ref<!fir.box<none>> errmsg(%16 : !fir.box<none>) {data_attr = #cuf.cuda<device>} -> i32
|
|
return
|
|
}
|
|
|
|
// -----
|
|
|
|
func.func @_QPsub1() {
|
|
%0 = fir.alloca !fir.box<!fir.heap<!fir.array<?xf32>>> {bindc_name = "a", uniq_name = "_QFsub1Ea"}
|
|
%4:2 = hlfir.declare %0 {data_attr = #cuf.cuda<device>, fortran_attrs = #fir.var_attrs<allocatable>, uniq_name = "_QFsub1Ea"} : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>, !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>)
|
|
%1 = fir.alloca i32
|
|
%11 = fir.convert %4#1 : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> !fir.ref<!fir.box<none>>
|
|
// expected-error@+1{{'cuf.allocate' op expect errmsg to be a reference to/or a box type value}}
|
|
%13 = cuf.allocate %11 : !fir.ref<!fir.box<none>> errmsg(%1 : !fir.ref<i32>) {data_attr = #cuf.cuda<device>, hasStat} -> i32
|
|
return
|
|
}
|
|
|
|
// -----
|
|
|
|
func.func @_QPsub1() {
|
|
%1 = fir.alloca i32
|
|
// expected-error@+1{{'cuf.deallocate' op expect box to be a reference to class or box type value}}
|
|
%2 = cuf.deallocate %1 : !fir.ref<i32> {data_attr = #cuf.cuda<device>} -> i32
|
|
return
|
|
}
|
|
|
|
// -----
|
|
|
|
func.func @_QPsub1() {
|
|
%0 = fir.alloca !fir.box<!fir.heap<!fir.array<?xf32>>> {bindc_name = "a", uniq_name = "_QFsub1Ea"}
|
|
%4:2 = hlfir.declare %0 {data_attr = #cuf.cuda<device>, fortran_attrs = #fir.var_attrs<allocatable>, uniq_name = "_QFsub1Ea"} : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>, !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>)
|
|
%1 = fir.alloca i32
|
|
%11 = fir.convert %4#1 : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> !fir.ref<!fir.box<none>>
|
|
// expected-error@+1{{'cuf.deallocate' op expect errmsg to be a reference to/or a box type value}}
|
|
%13 = cuf.deallocate %11 : !fir.ref<!fir.box<none>> errmsg(%1 : !fir.ref<i32>) {data_attr = #cuf.cuda<device>, hasStat} -> i32
|
|
return
|
|
}
|
|
|
|
// -----
|
|
|
|
func.func @_QPsub1() {
|
|
%0 = fir.alloca !fir.box<!fir.heap<!fir.array<?xf32>>> {bindc_name = "a", uniq_name = "_QFsub1Ea"}
|
|
%4:2 = hlfir.declare %0 {data_attr = #cuf.cuda<device>, fortran_attrs = #fir.var_attrs<allocatable>, uniq_name = "_QFsub1Ea"} : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>, !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>)
|
|
%c100 = arith.constant 100 : index
|
|
%7 = fir.alloca !fir.char<1,100> {bindc_name = "msg", uniq_name = "_QFsub1Emsg"}
|
|
%8:2 = hlfir.declare %7 typeparams %c100 {uniq_name = "_QFsub1Emsg"} : (!fir.ref<!fir.char<1,100>>, index) -> (!fir.ref<!fir.char<1,100>>, !fir.ref<!fir.char<1,100>>)
|
|
%9 = fir.embox %8#1 : (!fir.ref<!fir.char<1,100>>) -> !fir.box<!fir.char<1,100>>
|
|
%11 = fir.convert %4#1 : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> !fir.ref<!fir.box<none>>
|
|
%16 = fir.convert %9 : (!fir.box<!fir.char<1,100>>) -> !fir.box<none>
|
|
// expected-error@+1{{'cuf.deallocate' op expect stat attribute when errmsg is provided}}
|
|
%13 = cuf.deallocate %11 : !fir.ref<!fir.box<none>> errmsg(%16 : !fir.box<none>) {data_attr = #cuf.cuda<device>} -> i32
|
|
return
|
|
}
|
|
|
|
// -----
|
|
|
|
func.func @_QPsub1() {
|
|
%0 = cuf.alloc f32 {bindc_name = "r", data_attr = #cuf.cuda<device>, uniq_name = "_QFsub1Er"} -> !fir.ref<f32>
|
|
// expected-error@+1{{'cuf.free' op expect device, managed, pinned or unified cuda attribute}}
|
|
cuf.free %0 : !fir.ref<f32> {data_attr = #cuf.cuda<constant>}
|
|
return
|
|
}
|
|
|
|
// -----
|
|
|
|
func.func @_QPsub1(%arg0: !fir.ref<!fir.array<?xf32>> {cuf.data_attr = #cuf.cuda<device>, fir.bindc_name = "adev"}, %arg1: !fir.ref<!fir.array<?xf32>> {fir.bindc_name = "ahost"}, %arg2: !fir.ref<i32> {fir.bindc_name = "n"}, %arg3: !fir.ref<i32> {fir.bindc_name = "m"}) {
|
|
%0 = fir.dummy_scope : !fir.dscope
|
|
%1:2 = hlfir.declare %arg2 dummy_scope %0 {uniq_name = "_QFsub1En"} : (!fir.ref<i32>, !fir.dscope) -> (!fir.ref<i32>, !fir.ref<i32>)
|
|
%2:2 = hlfir.declare %arg3 dummy_scope %0 {uniq_name = "_QFsub1Em"} : (!fir.ref<i32>, !fir.dscope) -> (!fir.ref<i32>, !fir.ref<i32>)
|
|
%3 = fir.load %1#0 : !fir.ref<i32>
|
|
%4 = fir.load %2#0 : !fir.ref<i32>
|
|
%5 = arith.muli %3, %4 : i32
|
|
%6 = fir.convert %5 : (i32) -> i64
|
|
%7 = fir.convert %6 : (i64) -> index
|
|
%c0 = arith.constant 0 : index
|
|
%8 = arith.cmpi sgt, %7, %c0 : index
|
|
%9 = arith.select %8, %7, %c0 : index
|
|
%10 = fir.shape %9 : (index) -> !fir.shape<1>
|
|
%11:2 = hlfir.declare %arg0(%10) dummy_scope %0 {data_attr = #cuf.cuda<device>, uniq_name = "_QFsub1Eadev"} : (!fir.ref<!fir.array<?xf32>>, !fir.shape<1>, !fir.dscope) -> (!fir.box<!fir.array<?xf32>>, !fir.ref<!fir.array<?xf32>>)
|
|
%12 = fir.load %1#0 : !fir.ref<i32>
|
|
%13 = fir.load %2#0 : !fir.ref<i32>
|
|
%14 = arith.muli %12, %13 : i32
|
|
%15 = fir.convert %14 : (i32) -> i64
|
|
%16 = fir.convert %15 : (i64) -> index
|
|
%c0_0 = arith.constant 0 : index
|
|
%17 = arith.cmpi sgt, %16, %c0_0 : index
|
|
%18 = arith.select %17, %16, %c0_0 : index
|
|
%19 = fir.shape %18 : (index) -> !fir.shape<1>
|
|
%20:2 = hlfir.declare %arg1(%19) dummy_scope %0 {uniq_name = "_QFsub1Eahost"} : (!fir.ref<!fir.array<?xf32>>, !fir.shape<1>, !fir.dscope) -> (!fir.box<!fir.array<?xf32>>, !fir.ref<!fir.array<?xf32>>)
|
|
// expected-error@+1{{'cuf.data_transfer' op shape can only be specified on data transfer with references}}
|
|
cuf.data_transfer %20#0 to %11#0, %19 : !fir.shape<1> {transfer_kind = #cuf.cuda_transfer<host_device>} : !fir.box<!fir.array<?xf32>>, !fir.box<!fir.array<?xf32>>
|
|
return
|
|
}
|
|
|
|
// -----
|
|
|
|
module attributes {gpu.container_module} {
|
|
gpu.module @cuda_device_mod {
|
|
gpu.func @_QPsub_device1() {
|
|
gpu.return
|
|
}
|
|
}
|
|
llvm.func internal @__cudaFortranConstructor() {
|
|
%0 = cuf.register_module @cuda_device_mod -> !llvm.ptr
|
|
// expected-error@+1{{'cuf.register_kernel' op only kernel gpu.func can be registered}}
|
|
cuf.register_kernel @cuda_device_mod::@_QPsub_device1(%0 : !llvm.ptr)
|
|
llvm.return
|
|
}
|
|
}
|
|
|
|
// -----
|
|
|
|
module attributes {gpu.container_module} {
|
|
gpu.module @cuda_device_mod {
|
|
gpu.func @_QPsub_device1() {
|
|
gpu.return
|
|
}
|
|
}
|
|
llvm.func internal @__cudaFortranConstructor() {
|
|
%0 = cuf.register_module @cuda_device_mod -> !llvm.ptr
|
|
// expected-error@+1{{'cuf.register_kernel' op device function not found}}
|
|
cuf.register_kernel @cuda_device_mod::@_QPsub_device2(%0 : !llvm.ptr)
|
|
llvm.return
|
|
}
|
|
}
|
|
|
|
// -----
|
|
|
|
module attributes {gpu.container_module} {
|
|
llvm.func internal @__cudaFortranConstructor() {
|
|
%0 = cuf.register_module @cuda_device_mod -> !llvm.ptr
|
|
// expected-error@+1{{'cuf.register_kernel' op gpu module not found}}
|
|
cuf.register_kernel @cuda_device_mod::@_QPsub_device1(%0 : !llvm.ptr)
|
|
llvm.return
|
|
}
|
|
}
|
|
|
|
// -----
|
|
|
|
module attributes {gpu.container_module} {
|
|
llvm.func internal @__cudaFortranConstructor() {
|
|
%0 = cuf.register_module @cuda_device_mod -> !llvm.ptr
|
|
// expected-error@+1{{'cuf.register_kernel' op expect a module and a kernel name}}
|
|
cuf.register_kernel @_QPsub_device1(%0 : !llvm.ptr)
|
|
llvm.return
|
|
}
|
|
}
|
|
|
|
// -----
|
|
|
|
module attributes {gpu.container_module} {
|
|
gpu.module @cuda_device_mod {
|
|
llvm.func @_QPsub_device1() {
|
|
llvm.return
|
|
}
|
|
}
|
|
llvm.func internal @__cudaFortranConstructor() {
|
|
%0 = cuf.register_module @cuda_device_mod -> !llvm.ptr
|
|
// expected-error@+1{{'cuf.register_kernel' op only gpu.kernel llvm.func can be registered}}
|
|
cuf.register_kernel @cuda_device_mod::@_QPsub_device1(%0 : !llvm.ptr)
|
|
llvm.return
|
|
}
|
|
}
|