Skip to content

[X86][BF16] Add libcall for FP128 -> BF16 #115825

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Nov 12, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions compiler-rt/lib/builtins/CMakeLists.txt
Original file line number Diff line number Diff line change
Expand Up @@ -194,6 +194,7 @@ set(BF16_SOURCES
truncdfbf2.c
truncxfbf2.c
truncsfbf2.c
trunctfbf2.c
)

set(GENERIC_TF_SOURCES
Expand Down
18 changes: 18 additions & 0 deletions compiler-rt/lib/builtins/trunctfbf2.c
Original file line number Diff line number Diff line change
@@ -0,0 +1,18 @@
//===--------- lib/trunctfbf2.c - quad -> bfloat conversion -------*- C -*-===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
#define QUAD_PRECISION
#include "fp_lib.h"

#if defined(CRT_HAS_TF_MODE) && defined(__x86_64__)
#define SRC_QUAD
#define DST_BFLOAT
#include "fp_trunc_impl.inc"

COMPILER_RT_ABI dst_t __trunctfbf2(src_t a) { return __truncXfYf2__(a); }

#endif
1 change: 1 addition & 0 deletions llvm/include/llvm/IR/RuntimeLibcalls.def
Original file line number Diff line number Diff line change
Expand Up @@ -383,6 +383,7 @@ HANDLE_LIBCALL(FPROUND_PPCF128_F16, "__trunctfhf2")
HANDLE_LIBCALL(FPROUND_F32_BF16, "__truncsfbf2")
HANDLE_LIBCALL(FPROUND_F64_BF16, "__truncdfbf2")
HANDLE_LIBCALL(FPROUND_F80_BF16, "__truncxfbf2")
HANDLE_LIBCALL(FPROUND_F128_BF16, "__trunctfbf2")
HANDLE_LIBCALL(FPROUND_F64_F32, "__truncdfsf2")
HANDLE_LIBCALL(FPROUND_F80_F32, "__truncxfsf2")
HANDLE_LIBCALL(FPROUND_F128_F32, "__trunctfsf2")
Expand Down
2 changes: 2 additions & 0 deletions llvm/lib/CodeGen/TargetLoweringBase.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -171,6 +171,8 @@ RTLIB::Libcall RTLIB::getFPROUND(EVT OpVT, EVT RetVT) {
return FPROUND_F64_BF16;
if (OpVT == MVT::f80)
return FPROUND_F80_BF16;
if (OpVT == MVT::f128)
return FPROUND_F128_BF16;
} else if (RetVT == MVT::f32) {
if (OpVT == MVT::f64)
return FPROUND_F64_F32;
Expand Down
22 changes: 22 additions & 0 deletions llvm/test/CodeGen/X86/bfloat.ll
Original file line number Diff line number Diff line change
Expand Up @@ -1975,3 +1975,25 @@ define bfloat @PR108936(x86_fp80 %0) nounwind {
%2 = fptrunc x86_fp80 %0 to bfloat
ret bfloat %2
}

define bfloat @PR115710(fp128 %0) nounwind {
; X86-LABEL: PR115710:
; X86: # %bb.0:
; X86-NEXT: subl $28, %esp
; X86-NEXT: vmovaps {{[0-9]+}}(%esp), %xmm0
; X86-NEXT: vmovups %xmm0, (%esp)
; X86-NEXT: calll __trunctfbf2
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The result from __trunctfbf2 should be in xmm0 not eax right?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Right. I didn't notice the problem..

; X86-NEXT: # kill: def $ax killed $ax def $eax
; X86-NEXT: vmovw %eax, %xmm0
; X86-NEXT: addl $28, %esp
; X86-NEXT: retl
;
; CHECK-LABEL: PR115710:
; CHECK: # %bb.0:
; CHECK-NEXT: pushq %rax
; CHECK-NEXT: callq __trunctfbf2@PLT
; CHECK-NEXT: popq %rax
; CHECK-NEXT: retq
%2 = fptrunc fp128 %0 to bfloat
ret bfloat %2
}
Loading