Searched refs:spills (Results 1 – 25 of 175) sorted by relevance
1234567
20 ; REMARK: remark: /tmp/kk.c:3:20: 1 spills 1 reloads generated in loop{{$}}22 ; REMARK: remark: /tmp/kk.c:2:20: 1 spills 1 reloads generated in loop{{$}}24 ; REMARK: remark: /tmp/kk.c:1:20: 2 spills 2 reloads generated in loop{{$}}27 ; HOTNESS: remark: /tmp/kk.c:3:20: 1 spills 1 reloads generated in loop (hotness: 300)29 ; HOTNESS: remark: /tmp/kk.c:2:20: 1 spills 1 reloads generated in loop (hotness: 30000)31 ; HOTNESS: remark: /tmp/kk.c:1:20: 2 spills 2 reloads generated in loop (hotness: 300)36 ; THRESHOLD: remark: /tmp/kk.c:2:20: 1 spills 1 reloads generated in loop (hotness: 30000)46 ; YAML: - String: ' spills '59 ; YAML: - String: ' spills '72 ; YAML: - String: ' spills '[all …]
1 ; RUN: llc < %s -mcpu=cortex-a8 -align-neon-spills=0 | FileCheck %s2 ; RUN: llc < %s -mcpu=cortex-a8 -align-neon-spills=1 | FileCheck %s --check-prefix=NEON27 ; Stack pointer must be updated before the spills.32 ; This could legally happen before the spills.58 ; Stack pointer must be updated before the spills.85 ; Stack pointer must be updated before the spills.
2 ; RUN: llc < %s -mtriple=x86_64-linux -sink-insts-to-avoid-spills | FileCheck %s -check-prefix=SINK5 ; spills.
5 ; cause spills!
3 ; Now this test spills one register. But a reload in the loop is cheaper than
8 ; When the offset of VGPR spills into scratch space gets too large, an additional SGPR10 ; mechanism works even when many spills happen.
1 ; Check that we don't insert unnecessary CC spills
5 ; Check that there are no spills.
1 # RUN: llc -mtriple=amdgcn-amd-amdhsa -verify-machineinstrs -run-pass=si-lower-sgpr-spills -o - %s …
2 # RUN: llc -march=amdgcn -mcpu=tahiti -verify-machineinstrs -run-pass=si-lower-sgpr-spills -o - %s …
6 ; This function spills two values: %p and the materialized large constant.
22 # The fast register allocator puts some spills in -- these are no-ops as far23 # as the slot indexes are concerned. It doesn't matter which side of spills
1 …-mcpu=pwr9 -mtriple=powerpc64le-unknown-linux-gnu -ppc-enable-gpr-to-vsr-spills < %s | FileCheck …
11 ; will fail the small-frame-size check and the function has spills).
6 ; Check that this code only spills a single vector.
6 ; spills and fills. This is the case for <2 x double>,
4 ; This test makes sure spills of 64-bit pairs in Thumb mode actually