[PATCH] D48496: [llvm-mca] Rename Backend to Pipeline. NFC.

[Andrea Di Biagio via Phabricator via llvm-commits]

andreadb accepted this revision.
andreadb added a comment.

Hi Matt,

The patch looks good modulo the minor nits reported below.

Thanks!
-Andrea



================
Comment at: tools/llvm-mca/HWEventListener.h:34
   // handled by generic components as opaque values, but can still be
-  // emitted by subtarget-specific pipeline components (e.g. Scheduler,
+  // emitted by subtarget-specific pipeline components (e.g., Scheduler,
   // DispatchStage, ...) and interpreted by subtarget-specific EventListener
----------------
Please change Scheduler to ExecuteState. Also change components into 'stages'.


================
Comment at: tools/llvm-mca/Instruction.h:49
   // When computing read latencies, this value is matched against the
-  // "ReadAdvance" information. The hardware backend may implement
+  // "ReadAdvance" information. The hardware pipeline may implement
   // dedicated forwarding paths to quickly propagate write results to dependent
----------------
"backend" was correct here. Please revert this change.


================
Comment at: tools/llvm-mca/Instruction.h:267
 
-/// An instruction dispatched to the out-of-order backend.
+/// An instruction dispatched to the out-of-order pipeline.
 ///
----------------
This comment is no longer valid and should be rewritten. We shouldn't just replace backend with pipeline in this sentence.


================
Comment at: tools/llvm-mca/Pipeline.cpp:11
 ///
-/// Implementation of class Backend which emulates an hardware OoO backend.
+/// Implementation of class Pipeline which emulates a hardware OoO pipeline.
 ///
----------------
Same.


================
Comment at: tools/llvm-mca/Pipeline.h:11
 ///
-/// This file implements an OoO backend for the llvm-mca tool.
+/// This file implements an OoO pipeline for the llvm-mca tool.
 ///
----------------
Remove the OoO part. I don't think it makes much sense when applied to pipeline.


================
Comment at: tools/llvm-mca/Pipeline.h:28
 
-/// An out of order backend for a specific subtarget.
+/// An out of order pipeline for a specific subtarget.
 ///
----------------
Same. This sentence should probably be rewritten.


================
Comment at: tools/llvm-mca/Pipeline.h:30
 ///
 /// It emulates an out-of-order execution of instructions. Instructions are
 /// fetched from a MCInst sequence managed by an initial 'Fetch' stage.
----------------
The default pipeline currently emulates an OoO execution.


================
Comment at: tools/llvm-mca/README.txt:10
 consumption.  The tool currently works for processors with an out-of-order
-backend, for which there is a scheduling model available in LLVM.
+pipeline, for which there is a scheduling model available in LLVM.
 
----------------
Please revert this line. "Backend" was correct.


================
Comment at: tools/llvm-mca/README.txt:21
 The presence of long data dependency chains, as well as poor usage of hardware
-resources may lead to bottlenecks in the back-end.  The tool is able to generate
+resources may lead to bottlenecks in the pipeline.  The tool is able to generate
 a detailed report which should help with identifying and analyzing sources of
----------------
Please revert this part too.


================
Comment at: tools/llvm-mca/README.txt:347
 
-This section describes the instruction flow through the out-of-order backend, as
-well as the functional units involved in the process.
+This section describes the instruction flow through the out-of-order pipeline,
+as well as the functional units involved in the process.
----------------
I think this should still say 'out-of-order backend'.


================
Comment at: tools/llvm-mca/RetireControlUnit.h:29
 /// This class tracks which instructions are in-flight (i.e., dispatched but not
-/// retired) in the OoO backend.
+/// retired) in the OoO pipeline.
 //
----------------
s/pipeline/backend


https://reviews.llvm.org/D48496