| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* #44 Start on chapter 7
* #44
Temp hack on printing to not break on cycles.
Parse while statement.
* #44 Remove unnecessary method
* #44 Remove unnecessary phi map
* #44 Phi should have existing nodes from both scopes as input
* #44 We need to only insert phis once for each name that is looked up. To do so,
we need to know when a named is looked up first time inside the loop body.
The phi must have its second input added later on after the loop is fully parsed.
* #44 Add phi also when update() is called.
Only add phi if name was known at loop start
* #44 Only create phi in loop body if var binding in head != body.
During finish() lookup in body directly.
* #44 Update the print output to make it similar to that in Xmilia's fork
* #44 Disable peephole for phis & region involved in while loop
* Fix broken tests
Remove "disablePeephole" set in some tests, and remaining set for other tests, so a bulk test-all fails because disablePeephole is true for all.
Reset disablePeephole in Node.reset
Switch HashSet for BitSet (touches all).
* Restructure Node class
Group printing, group node & edge manipulations, group optimizations.
Also, fix LoopRegion idom, it does not use Region idom but uses the original
* #44 Node printing
* #44 Node printing
* #44 Fix control edges in loop region
Change label of loop region
Fix the output count in dump
* #44 Revise dump output
* #44 When a LoopRegionNode is duplicated, ensure the dupe is also a LoopRegionNode
* #44 When a LoopScopeNode is merged, we merge the names set too
* #44 When a LoopScopeNode is merged, we merge the names set too
* #44 Fix print1() on RegionNode to use label()
* #44 Fix bug caused by incorrect comparison of nodes between body and head of loop; the check must be performed before any phis are created
* Revert "#44 Fix bug caused by incorrect comparison of nodes between body and head of loop; the check must be performed before any phis are created"
This reverts commit 5f729709855046c5c10904f08385fd1aec9a298a.
* #44 Fix bug caused by incorrect comparison of nodes between body and head of loop; the check must be performed before any phis are created
* #44 when we create a phi incrementally for a loop region, all scopes that are cloned from the original must get the same phi update
* #44 test case
* #44 test case
* Squashed commit of the following:
commit 15666e287e6bbc39b3fcc49939165a1e908d43b2
Author: Cliff Click <cliffc@acm.org>
Date: Mon Nov 27 09:45:40 2023 -0800
lazy/eager phis
Eager phis by default, but can flip Parser.LAZY to switch to lazy phis. Several tests golden answers change when flipping this flag due to minor node renumbering.
Phi combines same_inputs and singleUniqueInput.
Fancier basic-block-like printing. Uses isMultiHead & isMultiTail which show up in many Nodes.
Remove blanks in triple-quote strings.
commit 6ea70182a92b7db4fe5a1fdc2c4491842ef77ee5
Author: Cliff Click <cliffc@acm.org>
Date: Sun Nov 26 11:13:22 2023 -0800
Remove dead scopes
Add a debug find
commit a52c06a47bf1e56866d8b4d043290838f053ad76
Author: Cliff Click <cliffc@acm.org>
Date: Sun Nov 26 10:21:48 2023 -0800
No constraints on backedges
nicer loopy graphs
commit b5ab550098355f0b66bf24b40adf2e5ba4d86e4e
Author: Cliff Click <cliffc@acm.org>
Date: Sun Nov 26 10:13:22 2023 -0800
Fix regression
commit 30a7f5c734c686c02d3ca479914c9427b65cc469
Author: Cliff Click <cliffc@acm.org>
Date: Sat Nov 25 21:51:47 2023 -0800
minor cleanup
commit ad742b795ce0c93cfc82b2b11740261cfd923d02
Author: Cliff Click <cliffc@acm.org>
Date: Sat Nov 25 21:47:05 2023 -0800
Fixes lazy phi
Scope now takes a 'loop' option on dupping; the dup'ed Scope is filled with "Lazy Phi" sentinels - markers to indicate a Phi is need if this variable is ever referenced. If its never referenced, at the loop exit the "Lazy Phi" unwinds back to the original value. Lookup and update calls lazily insert concrete Phis over "Lazy Phis" at first touch.
Rename _allScopes to _xScopes its "active" not "all" scopes now.
Remove try/finally block at end, because a compiler bug throws, then attempts a GraphViz on a broken graph, then throws again, hiding the original bug.
Removed 1 extra layer of Scopes.
Scope printing now prints in edge order; added a reversemap for printing.
More touches to ascii printing.
* Remove flag to parseBlock
* Bugfix for lazy on empty loop body
Remove extra phi name-matching code
* Keep head alive until entire end_loop merging is done
* Create chapter 8 as a copy of chapter 7
* #44 Chapter 7 is now eager phis only (#46)
* #44 Chapter 7 is now eager phis only
* MOVE IR Printing to dedicated class to declutter Node. Remove Ary (#49)
* #44 Move IR printing to dedicated class to declutter Node, and remove Ary
* #44 refactor Node.delUse() to use common del()
* #44 Some cleanups
* #44 Some cleanups
* #44 Some cleanups
* #44 Some cleanups
* Revert "#44 Some cleanups"
This reverts commit 2702c338c8bdaa5ef2128000e26f1a8496baf34a.
* #44 Some cleanups
* #44 Some cleanups
* #44 Some cleanups
* #44 Fix typo
* #44 Fix markdown formatting
* #44 Remove line numbers from code snippets
---------
Co-authored-by: Cliff Click <cliffc@acm.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* #37 Add a Utils class to hold some useful functions. (#39)
Move the name '$ctrl' to the ScopeNode
* Fix maven pom
Missing an explicit compiler target, so IntelliJ/Maven defaulted to version 5 ???
Remove explict ver 17 from Makefile.
* Add immediate dominator peephole
* A little more robust & simpler idom search
* #40 Revise chapter 6 writeup to include a description of dominators
* #40 Revise chapter 6 writeup to include a description of dominators
---------
Co-authored-by: Cliff Click <cliffc@acm.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* #37 Create chapter 6 skeleton
* #37 Revise type system - TOP, BOTTOM, CONTROL are now directly implemented in Type, XCONTROL is a new type and that's implemented directly too.
We have a _type field that is like an enum.
* #37 Revise Type.meet() - this is now final.
Subclasses can override xmeet() which is protected but
now handles simple types in Type itself.
* #37 Basic test for peephole of CFG
* #37 Revise compute of if
* #37 Revise compute of if
* #37 WIP peephole of If
* #37 WIP peephole of If
* #37 WIP peephole of If
* #37 WIP peephole of If
* #37 WIP peephole of If
* #37 WIP fix the chars inserted by IntelliJ
* #37 WIP Fix compute type in IfNode
* #37 WIP Null check region inputs
* #37 WIP Fix where we choose great grand parent instead of grand parent
Add new test
* #37 WIP More fixes and tests
* #37 WIP Some minor adjustments
* Cleanup folding dead IF tests
Remove isDeadCtrl & just test for the constant.
Remove DEAD_CTRL, doesn't properly reset between tests.
Move keep/unkeep to Node.
PhiNode has a proper compute() call.
ProjNode ideal is a lot simpler.
Region doesn't ever get a null.
Move the mergeScopes Region peep till after all Phis are added.
Type TOP and XCONTROL count as constants.
TypeTuple gets a toString
* #37 WIP Upate README
* #37 WIP Region node should check if one of its inputs is a region that needs to be peepholed
* #37 WIP If node should check if its ctrl input is a Region that has only 1 live input so can be peepholed
* #37 WIP Further update to README to make the peephole logic clearer. Also discuss the relationship between Region and Phi nodes
* Region removes dead input paths
Remove region-folding in Return, If.
Add a Makefile.
Reformat triple-quote lines to be emacs-friendly.
* StopNode idealize
Return type sharpens to include control input.
* Remove !=, >=, >; use Not and <, <=, ==
Fix bug mixing rotating different ops.
Generalize Add fold constants; apply to Mul, EQ.
Fix Div.copy
* Fix bug rotating e.g. < or <=
Peep is only valid for associative ops
* #37 WIP Doc updates
* Work over the Ch6 Readme
Major issue: Region inputs align 1-to-1 with Phi inputs, but a Region can have zero to many Phis. When we lose a Region input (dead path), we need to lose the same inputs to each Phi - but we don't lose the Phis themselves, unless we're folding away down to the last live input.
* Update chapter06/README.md
Co-authored-by: Yichen Yan <oraluben@outlook.com>
* Update chapter06/README.md
Co-authored-by: Yichen Yan <oraluben@outlook.com>
* #37 Backport type system changes to chapter 2
* #37 Backport type system changes to chapter 3
* #37 Backport type system changes to chapter 4
* #37 Backport type system changes to chapter 5
* #37 Merge chapter 5 updates from main.
Remove language grammar module as we describe the grammar incrementally in each chapter.
* #37 Chapter 6 doc updates - added visuals to show effects of CFG peephole
* #37 Sync some updates to chapter 2
* #37 Sync some updates to chapter 3
* #37 Sync some updates to chapter 4
---------
Co-authored-by: Cliff Click <cliffc@acm.org>
Co-authored-by: Yichen Yan <oraluben@outlook.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* Backport fix for set_def killing new_def (#21)
* Use JDK 21
* Remove duplicate field in pom
* Enable GA test for pull requests
---------
Co-authored-by: XmiliaH <45106915+XmiliaH@users.noreply.github.com>
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Initial work on chapter 4.
Updated language grammar for chapter 4
Enhanced Parser to implement comparison operators
Parser now allows setting 'arg' as input parameter to program
The StartNode now has a tuple as result with two values.
We set up two projection nodes as output of Start - the first is bound to scope as '$ctrl' and second as 'arg'.
Return node now obtains the ctrl input by looking up '$ctrl' in scope; thus it gets the projection node.
Added TypeTuple - represents a tuple of values. This is the type of Start.
Added BoolNode - for comparison operators.
Added MultiNode - interface for nodes that generate tuple results. Start implements this interface.
Added ProjNode - projections to extract values from Start node. Each projection takes a MultiNode and an index value.
Initial writeup of chapter 4 in README.
GraphVisualizer now searches node outputs as well to locate all nodes
* Add TypeControl.
The first member of Start's tuple is of TypeControl.
Correct the projType() method
* Quote port names in edges as unquoted $ctrl is not valid.
Revise how we decide an edge is a control edge so that it works correctly for projections.
* We used to compute initial type in Node but this is problematic for new node types such as
ProjNode and BoolNode as the construction is not yet complete when Node calls compute().
So we have to move compute() call down to the sub type's constructor.
Add some tests for comparison ops.
* Describe the $ctrl name binding and how it represents the Ptri net model.
Add a diagram to show why we need algebraic manipulation of expressions.
* Implement a simplification for expressions such as
(con1 + arg) + con2 -> arg + con3
con1 + (arg + con2) -> arg + con3
* Add diagram to show what the graph looks like post peephole
* Explain projection nodes
* Cleanup Chapter 4, add peepholes.
Add like 6 peepholes. Replace the 2 prior ones with 1-liners.
BoolNode: replace the switch with 6 1-line classes.
Giant comment discussion on idealize() calling convention.
TypeInteger shorten isTop and isBot, and use them instead of calling "equals()".
* 2 more peeps, no tests yet
* Cleanup Chapter 4
`peephole()` kills nodes killed by `idealize()` calls... and works recursively. This requires keeping "mid-peephole" nodes alive. The edge-swap code in Add and Mul now keeps their inputs alive mid-swap.
MultiNode is a class, not an interface.
Move the initial 'compute()' to peephole, instead of requiring each constructor to do it. Start and Constant remember their original types, even if the `_type` field changes.
`kill()` flags nodes as DEAD. Print DEAD nodes with "DEAD". Change recursive `_print()` to a tik-tok style, so each individual print does not need to handle DEAD printing. This printing will soon need a "visit" bit to avoid infinite recursion.
Projs pick up a debugging name.
GraphViz: combine Multis and their Projs. Drop Constant inputs from Start. Fix coloring Control edges red.
* minor WS cleanup
* Cleaned up Node.peephole
Use a smart DCE call after peephole. More comments. Tidy interfaces.
* minor comment cleanup
* But the proj nodes inside a table
* No need to set column span on multi node as we put the proj nodes in a table
* Update the dot diagrams
* Rename method to make it more obvious to the reader
* Start to add a description of peephole
* Rename del() to delUse() and make it an instance method as it reads better that way.
Add new method addUse() again to improve readability.
Adjust removeDeadCode() to make the intention clear.
* Update docs explaining how peephole works
Update the diagrams
* Fix errors
* Add diagram to show what the graph looks like post peephole
* 2 more peeps, no tests yet
* Cleanup Chapter 4
`peephole()` kills nodes killed by `idealize()` calls... and works recursively. This requires keeping "mid-peephole" nodes alive. The edge-swap code in Add and Mul now keeps their inputs alive mid-swap.
MultiNode is a class, not an interface.
Move the initial 'compute()' to peephole, instead of requiring each constructor to do it. Start and Constant remember their original types, even if the `_type` field changes.
`kill()` flags nodes as DEAD. Print DEAD nodes with "DEAD". Change recursive `_print()` to a tik-tok style, so each individual print does not need to handle DEAD printing. This printing will soon need a "visit" bit to avoid infinite recursion.
Projs pick up a debugging name.
GraphViz: combine Multis and their Projs. Drop Constant inputs from Start. Fix coloring Control edges red.
* minor WS cleanup
* Cleaned up Node.peephole
Use a smart DCE call after peephole. More comments. Tidy interfaces.
* minor comment cleanup
* But the proj nodes inside a table
* No need to set column span on multi node as we put the proj nodes in a table
* Update the dot diagrams
* Rename method to make it more obvious to the reader
* Start to add a description of peephole
* Rename del() to delUse() and make it an instance method as it reads better that way.
Add new method addUse() again to improve readability.
Adjust removeDeadCode() to make the intention clear.
* Update docs explaining how peephole works
Update the diagrams
* Minor cleanups
Split isUnused from isDead, with comments.
Rename removeDeadCode back to deadCodeElim after discussion; matches the historical compiler wording.
Fixup large javadoc example with a table.
* match doc and code
* Fix name of method
* Fix name of method
---------
Co-authored-by: dibyendumajumdar <mobile@majumdar.org.uk>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Chapter 3 - add vars and lexical scope.
* Fix typo
* Overhaul graph viz
Makes the scopes pointing into the graph being build more visually appealing.
Constants are a printed with just the constant value.
* Fix control edge color
* Fix missing grouping operator parse
Add a parens test.
Shorten names 'Identifier' to 'Id'
_lexer.match -> match
require() passes a Node thru, so can be flow-coded.
Uniform syntax error messages.
Handy Lexer.toString for debugging.
Drop Optional for null returns; less total code; less objects; less confusion.
camelCase some fcns.
* All parser helpers return Node, not void.
Drop _ret. parseBlock and parseStatement now the same.
* CLosely align chap2 and chap3 parsers
Mostly whitespace aligned; dragged a few functions around to make more sense.
I think the diff from Ch2 and Ch3 is down a few differences about what is a statement, plus all the symbol stuff for scopes and variables.
* Add examples illustrating the symbol table creation.
* Update docs and comments
---------
Co-authored-by: Cliff Click <cliffc@acm.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Start on Chapter 2.
* Start on Chapter 2.
* Refactor
* Refactor
* Initial attempt at peephole opt for add operation
* cleanup unused import
* Start on graph visualizer
* Start on graph visualizer
* Rework chapter 2
Add a short discourse on the Types, and provide a Bottom type.
Make a final toString; require each Node to specify its special printing via label().
Add a recursing printer called print(); toString is used by default in the debugger and so cannot be recursive.
Make the peephole rules based on a proper Type and idealize calls.
Big comments about the Type computations.
Every node picks up a _print, compute and idealize() call.
Rename long arith ops Divide, UnaryMinus, Multiply, Subtract into shorter uniform-length names. Cleans up the parser where he does the precedence op parsing.
Rework Parser:
- Drop Token class, its all Strings except in the one place we know a number is coming
- Embed the Lexer as a field in the Parser, instead of trying to pass it around everywhere in the Parser
- StartNode becomes a read-only global, so we don't have to pass it around everywhere in the Parser AND optimizer.
- Lexer matches on strings and returns strings. Match and consume *at the moment* instead of carrying around a current token/string and forcing the next-token after consuming the current one.
- All parse subexpressions shrink by 1/2 after these changes.
* Initial writeup of Chapter 2
* Initial writeup of Chapter 2
* Initial writeup of Chapter 2
* Initial writeup of Chapter 2
* Remove range
* progress Chapter 2
Drop integer ranges.
Broader lattice pic.
Lots of prose changes in the readme.
Fixed title in gviz dot generation.
Fixed bug with peephole making redundant constants.
* Implement compute for all the arithmetic node types
Add a flag to switch off peephole so we can see the full graph
Add post peephole graph to writeup
* Implement compute for all the arithmetic node types
Add a flag to switch off peephole so we can see the full graph
Add post peephole graph to writeup
* Fix bad path to the svg
* Update pre-peephole diagram
* Add dead code elimination
...and a set_def to proper update graph edges.
* Update README.md
* Cleanup the GraphVisualizer a bit.
Update the graphs.
Update the chapter list as chapter 3 deleted, and chapter 4 is now chapter 3.
---------
Co-authored-by: Cliff Click <cliffc@acm.org>
|
| | |
|
| | |
|
| |
|
|
|
| |
For pattern-match switch statements. Much nicer looking code.
Also IDEAJ gave me huge grief, so I might have hit this with a larger hammer than needed.
|
| |
|
|
|
|
| |
Removed the common chapter.
Reimplemented Lexer.
Revised the list of chapters to something more accurate.
|
| |
|
|
|
|
|
|
| |
Updated README
Added language module with grammar
Split the code into common and chapter01 modules.
Removed a number of utility classes that were imported from AA.
Added github workflow file.
|
| |
|
|
| |
Removed the hash quality checks in Util as these access internals of maps in JCTools
|
|
|
util classes from AA
Lexer code extracted from the Parse class.
This is all draft as likely to change as we start creating the AST parser.
|
