mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-10-19 02:52:53 +02:00
262321d1ff
This patch lets the llvm tools handle the new HVX target features that
are added by frontend (clang). The target-features are of the form
"hvx-length64b" for 64 Byte HVX mode, "hvx-length128b" for 128 Byte mode HVX.
"hvx-double" is an alias to "hvx-length128b" and is soon will be deprecated.
The hvx version target feature is upgated form "+hvx" to "+hvxv{version_number}.
Eg: "+hvxv62"
For the correct HVX code generation, the user must use the following
target features.
For 64B mode: "+hvxv62" "+hvx-length64b"
For 128B mode: "+hvxv62" "+hvx-length128b"
Clang picks a default length if none is specified. If for some reason,
no hvx-length is specified to llvm, the compilation will bail out.
There is a corresponding clang patch.
Differential Revision: https://reviews.llvm.org/D38851
llvm-svn: 316101
15 lines
679 B
ArmAsm
15 lines
679 B
ArmAsm
#RUN: llvm-mc -triple=hexagon -mcpu=hexagonv60 -filetype=obj -mhvx %s | \
|
|
#RUN: llvm-objdump -triple=hexagon -mcpu=hexagonv60 -mhvx -d - | \
|
|
#RUN: FileCheck %s
|
|
|
|
V31.b = vlut32(V29.b, V15.b, R1)
|
|
# CHECK: 1b79fd3f { v31.b = vlut32(v29.b,v15.b,r1) }
|
|
V31.b |= vlut32(V29.b, V15.b, R2)
|
|
# CHECK: 1b7afdbf { v31.b |= vlut32(v29.b,v15.b,r2) }
|
|
V31:30.h = vlut16(V29.b, V15.h, R3)
|
|
# CHECK: 1b7bfdde { v31:30.h = vlut16(v29.b,v15.h,r3) }
|
|
v31:30.h |= vlut16(v2.b, v9.h, r4)
|
|
# CHECK: 1b4ce2fe { v31:30.h |= vlut16(v2.b,v9.h,r4) }
|
|
v31.w = vinsert(r4)
|
|
# CHECK: 19a4e03f { v31.w = vinsert(r4) }
|