241a56dfad
\[NVPTX\] Add ApplyPriority intrinsics This PR adds applypriority.\* intrinsics with relevant eviction priorities. * The lowering is handled from nvvm to nvptx tablegen directly. * Lit tests are added as part of applypriority.ll * The generated PTX is verified with a 12.3 ptxas executable. * Added docs for these intrinsics in NVPTXUsage.rst. For more information, refer to the PTX ISA `<https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-applypriority>`_. --------- Co-authored-by: abmajumder <abmajumder@nvidia.com>
35 lines
1.5 KiB
LLVM
35 lines
1.5 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
|
|
; RUN: llc < %s -mtriple=nvptx64 -mcpu=sm_80 -mattr=+ptx74| FileCheck --check-prefixes=CHECK-PTX64 %s
|
|
; RUN: %if ptxas-11.4 %{ llc < %s -mtriple=nvptx64 -mcpu=sm_80 -mattr=+ptx74| %ptxas-verify -arch=sm_80 %}
|
|
|
|
target triple = "nvptx64-nvidia-cuda"
|
|
|
|
declare void @llvm.nvvm.applypriority.global.L2.evict.normal(ptr addrspace(1) %global_ptr, i64 %size)
|
|
declare void @llvm.nvvm.applypriority.L2.evict.normal(ptr %ptr, i64 %size)
|
|
|
|
define void @applypriority_global_L2(ptr addrspace(1) %global_ptr, i64 %size) {
|
|
; CHECK-PTX64-LABEL: applypriority_global_L2(
|
|
; CHECK-PTX64: {
|
|
; CHECK-PTX64-NEXT: .reg .b64 %rd<2>;
|
|
; CHECK-PTX64-EMPTY:
|
|
; CHECK-PTX64-NEXT: // %bb.0:
|
|
; CHECK-PTX64-NEXT: ld.param.u64 %rd1, [applypriority_global_L2_param_0];
|
|
; CHECK-PTX64-NEXT: applypriority.global.L2::evict_normal [%rd1], 128;
|
|
; CHECK-PTX64-NEXT: ret;
|
|
tail call void @llvm.nvvm.applypriority.global.L2.evict.normal(ptr addrspace(1) %global_ptr, i64 128)
|
|
ret void
|
|
}
|
|
|
|
define void @applypriority_L2(ptr %ptr, i64 %size) {
|
|
; CHECK-PTX64-LABEL: applypriority_L2(
|
|
; CHECK-PTX64: {
|
|
; CHECK-PTX64-NEXT: .reg .b64 %rd<2>;
|
|
; CHECK-PTX64-EMPTY:
|
|
; CHECK-PTX64-NEXT: // %bb.0:
|
|
; CHECK-PTX64-NEXT: ld.param.u64 %rd1, [applypriority_L2_param_0];
|
|
; CHECK-PTX64-NEXT: applypriority.L2::evict_normal [%rd1], 128;
|
|
; CHECK-PTX64-NEXT: ret;
|
|
tail call void @llvm.nvvm.applypriority.L2.evict.normal(ptr %ptr, i64 128)
|
|
ret void
|
|
}
|