diff options
| author | Cliff Click <cliffc@acm.org> | 2026-10-18 22:09:58 -0700 |
|---|---|---|
| committer | Cliff Click <cliffc@acm.org> | 2026-10-18 22:09:58 -0700 |
| commit | 300d2a9c3672acacc8bdd8d5370a424651a00571 (patch) | |
| tree | f33d895f6ec4e08692efa4c090c681257553097d /chapter19/README.md | |
| parent | 0be52f8dd82f70d4c557fe916d47dc425401f848 (diff) | |
| download | Simple-300d2a9c3672acacc8bdd8d5370a424651a00571.tar.gz Simple-300d2a9c3672acacc8bdd8d5370a424651a00571.zip | |
Cleanup & backportclinit
The phase rework, made new phases explicit and add them to the CodeGen driver. Fix the flag-spilling on X86 and ARM. Fix the GCM conservative anti-deps and backport.
Diffstat (limited to 'chapter19/README.md')
| -rw-r--r-- | chapter19/README.md | 10 |
1 files changed, 10 insertions, 0 deletions
diff --git a/chapter19/README.md b/chapter19/README.md index d8f7c9e1..c7bd1b9f 100644 --- a/chapter19/README.md +++ b/chapter19/README.md @@ -43,6 +43,16 @@ the porting code is looked up and used then. This allows the Java implementation at least the ability to add CPU ports lazily. +The driver from Chapter 18 accepts a CPU and calling convention when machine +selection is needed: + +```java +code.driver(CodeGen.Phase.LocalSched, "x86_64_v2", "SystemV"); +``` + +This runs through loop discovery, instruction selection, and both schedulers. +Omitting the CPU keeps the ideal graph for evaluation. + ## Machine Specific Nodes We add a notion of "machine specific" Nodes, which implement the `MachNode` |