1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-01-31 20:51:52 +01:00
llvm-mirror/test/CodeGen/AMDGPU/at-least-one-def-value-assert.mir
Stanislav Mekhanoshin 5742e35593 [AMDGPU] Define 16 bit VGPR subregs
We have loads preserving low and high 16 bits of their
destinations. However, we always use a whole 32 bit register
for these. The same happens with 16 bit stores, we have to
use full 32 bit register so if high bits are clobbered the
register needs to be copied. One example of such code is
added to the load-hi16.ll.

The proper solution to the problem is to define 16 bit subregs
and use them in the operations which do not read another half
of a VGPR or preserve it if the VGPR is written.

This patch simply defines subregisters and register classes.
At the moment there should be no difference in code generation.
A lot more work is needed to actually use these new register
classes. Therefore, there are no new tests at this time.

Register weight calculation has changed with new subregs so
appropriate changes were made to keep all calculations just
as they are now, especially calculations of register pressure.

Differential Revision: https://reviews.llvm.org/D74873
2020-03-31 11:49:06 -07:00

29 lines
1.2 KiB
YAML

# RUN: not --crash llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx906 -verify-machineinstrs -run-pass=machine-scheduler -verify-misched -o /dev/null %s 2>&1 | FileCheck %s
# CHECK: *** Bad machine code: No live subrange at use ***
# CHECK-NEXT: - function: at_least_one_value_should_be_defined_by_this_mask
# CHECK-NEXT: - basic block: %bb.0
# CHECK-NEXT: - instruction: 48B dead undef %2.sub0:vreg_128 = COPY %0.sub0:vreg_128
# CHECK-NEXT: - operand 1: %0.sub0:vreg_128
# CHECK-NEXT: - interval: %0 [16r,48r:0) 0@16r L000000000000000C [16r,32r:0) 0@16r weight:0.000000e+00
# This used to assert with: !SR.empty() && "At least one value should be defined by this mask"
# This MIR is invalid and should be caught by the verifier. %0.sub0 is
# used, but not defined. There are also lanes in %0 that are not used
# or defined anywhere. Previously there was an assertion in the
# LiveInterval computation, which was more confusing. The invalid
# LiveRange should be produced and the verifier will catch it.
---
name: at_least_one_value_should_be_defined_by_this_mask
tracksRegLiveness: true
body: |
bb.0:
undef %0.sub1:vreg_128 = V_MOV_B32_e32 0, implicit $exec
%1:vreg_128 = COPY %0
undef %2.sub0:vreg_128 = COPY %0.sub0
...