v0.25.0

xDSL framework

• core: allow result variadic inference @alexarice (#3559)
• core: fix bitwidth 0 value range @superlopuh (#3552)
• core: add functional-type directive @alexarice (#3517)
• core: Fix 'parse_single_type' for operands/results directives @alexarice (#3553)
• core: fix (and test) operands/results directives with no variadics @alexarice (#3551)
• core: remove unnecessary whitespace printing in assembly format @alexarice (#3530)
• core: Fix extractor logic for optional and default-valued properties @alexarice (#3525)

Dialects

• dialects: (builtin) normalize signless values on init @superlopuh (#3554)
• dialects: (builtin) use truthiness in of constant IntegerAttr values @superlopuh (#3556)
• dialects: (bufferization) fix materialize_in_destination @alexarice (#3495)
• dialects: (arith) add addi constant propagation @alexarice (#3563)
• dialects: (csl) handle negative-valued task ID representation @superlopuh (#3558)
• dialects: (arith) add Cmpi canonicalization @alexarice (#3564)
• dialects (arm): add func op @emmau678 (#3546)
• dialects: (arith) add muli canonicalization @alexarice (#3557)
• dialects: (builtin) fix true/false printing @superlopuh (#3555)
• dialects: (builtin) move value range helper definitions to utils @superlopuh (#3550)
• dialects: (builtin) don't use data in DenseArrayBase @superlopuh (#3542)
• dialects: (builtin) verify that the elements of dense satisfy type @superlopuh (#3544)
• dialects: (builtin) fixed FixedBitwidthType size calculation @superlopuh (#3543)
• dialects: (builtin) DenseIntOrFPElementsAttr: make data private and use getters instead @jorendumoulin (#3535)
• dialects: (builtin) remove Index support in DenseArrayBase @superlopuh (#3541)
• dialects: (bufferization) add CloneOp @jorendumoulin (#3560)

Transformations

• transformations: (convert-stencil-to-csl-stencil) Support multiple input buffers @n-io (#3575)
• transforms: use rewriter and listener in convert-stencil-to-csl-stencil @math-fehr (#3538)
• transforms: use Rewriter instead of PatternRewriter in mlir-opt @math-fehr (#3537)

Interpreter

• interpreter: (pdl) add an initial version of EqSat PDLMatcher @jianyicheng (#3574)
• interpreters: (irdl) generate better op name @alexarice (#3529)

🗎 Documentation

• documentation: add experimental features flag guide to README @superlopuh (#3567)

🐛 Bug Fixes

• transformations: (convert-stencil-to-csl-stencil) Support multiple input buffers @n-io (#3575)
• interactive: fix dark mode @alexarice (#3582)
• dialects: (bufferization) fix materialize_in_destination @alexarice (#3495)
• bug: add init.py @alexarice (#3568)
• dialects: (csl) handle negative-valued task ID representation @superlopuh (#3558)
• dialects: (builtin) fix true/false printing @superlopuh (#3555)
• core: fix (and test) operands/results directives with no variadics @alexarice (#3551)
• core: Fix extractor logic for optional and default-valued properties @alexarice (#3525)

Continuous Integration

• CI: add "tests" to extraPaths for Pyright to find conftest locally @superlopuh (#3571)

Installation

• pip prod(deps): bump textual from 0.88.1 to 0.89.1 @dependabot (#3578)
• pip prod(deps): bump jax from 0.4.35 to 0.4.36 @dependabot (#3577)
• pip prod(deps): bump marimo from 0.9.30 to 0.9.31 @dependabot (#3579)
• pip prod(deps): bump ruff from 0.8.1 to 0.8.2 @dependabot (#3580)
• pip prod(deps): bump pyright from 1.1.389 to 1.1.390 @dependabot (#3570)
• pip prod(deps): bump marimo from 0.9.28 to 0.9.30 @dependabot (#3565)
• pip prod(deps): bump marimo from 0.9.27 to 0.9.28 @dependabot (#3561)
• pip prod(deps): bump textual from 0.87.1 to 0.88.1 @dependabot (#3549)
• pip prod(deps): bump ruff from 0.8.0 to 0.8.1 @dependabot (#3548)

Miscellaneous

• interactive: fix dark mode @alexarice (#3582)
• misc: use experimental TypeForm feature of Pyright @superlopuh (#3450)
• minor: (csl-wrapper-hoist-buffers) Always use good name hints @n-io (#3562)
• core: Fix 'parse_single_type' for operands/results directives @alexarice (#3553)
• core: fix (and test) operands/results directives with no variadics @alexarice (#3551)
• misc: remove unused name argument from builtin attribute parsing @superlopuh (#3545)
• core: remove unnecessary whitespace printing in assembly format @alexarice (#3530)
• tool: (irdl-to-pyrdl) Generate better class name @alexarice (#3528)

Interactive

• interactive: fix dark mode @alexarice (#3582)

v0.24.0

xDSL framework

• core: Add results directive @alexarice (#3518)
• core: Add operands directive @alexarice (#3507)
• core: Move assembly format type parsing to directives @alexarice (#3516)
• core: Remove ParsingContext @alexarice (#3524)
• core: Walk regions for rewritings instead of ops @math-fehr (#3506)
• core: introduce TypedAttributeConstraint @alexarice (#3318)
• dialects: (builtin) make DenseIntOrFPElementsAttr a TypedAttribute @alexarice (#3509)
• core: add get_type helper to TypedAttribute @alexarice (#3508)
• core: remove generic and generic check from TypedAttribute @alexarice (#3504)
• core: refactor assembly format directives @alexarice (#3501)
• core: make BaseAttr inferrable @alexarice (#3491)
• core: make TypedAttribute printing generic @alexarice (#3490)
• core: simplify TypedAttribute @alexarice (#3487)
• core: new inference system for constraints @alexarice (#3456)
• core: (irdl) ParamAttrConstraint can infer recursively @superlopuh (#3477)
• core: use & for AllOf constraint @alexarice (#3441)
• rewriting: add attr_constr_rewrite_pattern @superlopuh (#3439)
• misc: raise error if variadic results types aren't referenced in assembly format @superlopuh (#3416)
• core: Parse FloatAttr from hexadecimal bit representation @n-io (#3378)
• core: Print FloatAttr special values @n-io (#3377)
• core: Print fp literals losslessly @n-io (#3381)
• core: (HasParent) allow no parent for testing @alexarice (#3323)
• core: always print unregistered ops @alexarice (#3303)
• core: relax some types in range variable solving @alexarice (#3317)
• core: default props should always be added @alexarice (#3308)
• core: Add default valued attributes @alexarice (#3305)
• core: Implement RangeVarConstraint @alexarice (#3261)

Dialects

• dialects (arm): add return instruction (bx lr) @emmau678 (#3521)
• dialects (arm): add mul instruction with 2 source registers @emmau678 (#3515)
• dialects (arm): add mov op @emmau678 (#3476)
• dialects: (stream) remove stream dialect and move types to memref_stream and snitch @superlopuh (#3510)
• dialects: (builtin) make DenseIntOrFPElementsAttr a TypedAttribute @alexarice (#3509)
• dialects: (linalg) linalg.fill attribute positioning @qaco (#3496)
• dialects: (stream) remove abstract stream read and write operations @superlopuh (#3486)
• dialects: (stream) use assembly format for stream read and write ops @superlopuh (#3484)
• dialects: (stream) simplify constr helper on stream attributes @superlopuh (#3494)
• dialects: (stream) simplify constr helper on stream attributes @superlopuh (#3473)
• dialects: (builtin) make FloatAttr a TypedAttribute @alexarice (#3488)
• dialects: (tensor) add insertop and extractop @francescodaghero (#3480)
• dialects: (bufferization) Simplify tensor/memref constraint @alexarice (#3475)
• dialects (arm): add first operation @emmau678 (#3462)
• dialects: (stream) remove unused stream.read and stream.write ops @superlopuh (#3470)
• dialects (arm): initialise ARM dialect @emmau678 (#3455)
• dialects: (bufferization) add custom syntax to alloc_tensor @superlopuh (#3451)
• dialects: (accfg) Update ACCFG dialect to match version in snax-mlir @AntonLydike (#3429)
• dialects: (stream) add constr to StreamType base class @superlopuh (#3440)
• dialects: move FastMathFlag and FastMathAttrBase to xdsl.dialects.utils @superlopuh (#3412)
• dialects: (riscv) split out the base classes for custom format and assembly printing @superlopuh (#3436)
• dialects: (eqsat) add min_cost_index optional attribute to eclass @superlopuh (#3433)
• dialects: (eqsat) verify that operands of eclass ops are not produced by eclass @superlopuh (#3432)
• dialects: (llvm) added disjoint flag @lfrenot (#3428)
• dialects: (scf) add assembly format to scf.condition @superlopuh (#3413)
• dialects (scf): Canonicalization for const rehoisting @mamanain (#3421)
• dialects: (llvm) added overflow flags to trunc @lfrenot (#3422)
• dialects: (llvm) added the 'nneg' keyword to zext @lfrenot (#3420)
• dialects: (llvm) changed the exact attribute to a UnitAttr @lfrenot (#3408)
• dialects: (llvm) added the 'exact' keyword to the corresponding operations @lfrenot (#3390)
• dialects: (stim) Add parser for .stim formatted strings @kimxworrall (#3122)
• dialects: llvm fixed parsing of nested structs @watermelonwolverine (#3369)
• dialects: (llvm) fixed generic printing of llvm.icmp @lfrenot (#3375)
• dialects: (llvm) Registering llvm.icmp @lfrenot (#3356)
• dialects: (arith) select canonicalization patterns @alexarice (#3368)
• dialects: (pdl) Filecheck test documenting how to match an MLIR operation without return type @qaco (#3358)
• dialects: (llvm) Added pretty printing to llvm.mlir.constant @lfrenot (#3352)
• dialects: (varith) fix switch parsing @alexarice (#3351)
• dialects: (varith) add switch @alexarice (#3342)
• dialects: remove some stencil dead code. @PapyChacal (#3055)
• dialects: (llvm) Add the integer converstion operations @lfrenot (#3337)
• Arith to riscv lowering fastmath cmpf @knickish (#3277)
• dialects: (csl-stencil) Add coefficients to apply op @n-io (#3320)
• Update mlir to d401987fe349a87c53fe25829215b080b70c0c1a @alexarice (#3310)
• dialects: (arith) fix attribute dictionary printing and parsing @superlopuh (#3327)
• dialects: (csl-stencil) Separate varargs for apply regions @n-io (#3316)
• transformations: (cf) switch canonicalization @alexarice (#3291)
• dialects: (bufferization) add materialize_in_destination @mamanain (#3301)
• transformations: (cf) cf.cond_br truth propagation @alexarice (#3285)
• transformations: (cf) cf.cond_br identical successors @alexarice (#3284)
• dialects: (riscv) add fastmath flag to RdRsRs Float Float Int operations @knickish (#3276)
• dialects: (llvm) Add support for overflow flags @lfrenot (#3293)
• dialects: (builtin) shape for DenseIntORFPElementsAttr is always known @jorendumoulin (#3534)
• dialects: (llvm) avoid double whitespace after exact @tobiasgrosser (#3415)
• dialects: (ptr) Introducing ptr dialect @mamanain (#3380)
• dialect: (tensor) add CollapseShapeOp @mamanain (#3302)
• feat: Added verify to the integer conversion operations to follow llvm semantics @lfrenot (#3341)

Transformations

• bug: (lower-csl-stencil) Zero-out accumulator for full reduction access @n-io (#3520)
• transformations: port MLIR's empty-tensor-to-alloc-tensor pass @superlopuh (#3453)
• transformations: add eqsat-extract @superlopuh (#3434)
• transformations: (convert-riscv-to-llvm) @jorendumoulin (#2468)
• transforms: add restrict flag to StencilShapeMinimize pass @emmau678 (#3411)
• transformations: add eqsat-add-costs @cowardsa (#3315)
• transformations: (pdl) Different type matching between polymorphic and fixed case @qaco (#3405)
• transformations: (scf) loop range folding @mamanain (#3417)
• dialects (scf): Canonicalization for const rehoisting @mamanain (#3421)
• transformations: New test-add-timers-to-top-level-funcs pass @n-io (#3407)
• transformations: Materialise stores before bufferization @n-io (#3403)
• transformations: ptr -> riscv conversion @mamanain (#3393)
• transformations: (pdl) Apply PDL patterns from source @qaco (#3391)
• transformations: Support declarations in func persist arg names @n-io (#3397)
• transformations: Persist func arg names as arg_attr @n-io (#3395)
• transformations: check invalid cases in memref -> ptr conversion @mamanain (#3392)
• transformations: Convert memref to ptrdialect @mamanain (#3383)
• transformations: (csl-stencil-bufferize) Reselect linalg outs @n-io (#3373)
• dialects: (arith) select canonicalization patterns @alexarice (#3368)
• transformations: (canonicalize) Arith const reassociation @n-io (#3364)
• transformations: Canonicalizations for arith on float constants @n-io (#3362)
• transformations: New varith-fuse-repeated-operands pass @n-io (#3357)
• fix: (csl) Use u32 as the counter in async loops @dk949 (#3353)
• transformations: (linalg-to-csl) Lower generic to fmac(h|s) @n-io (#3345)
• transformations: New linalg-fuse-multiply-add pass @n-io (#3347)
• transformations: (csl-stencil-bufferize) Inject accumulator in all csl-stencil linalg ops @n-io (#3343)
• transformations: (lower-csl-wrapper) Add params_as_consts flag @n-io (#3324)
• transformations: (memref-to-dsd) Support memref.load ops @n-io (#3338)
• transformations: Split varith into neighbour and own data across csl_stencil regions @n-io (#3307)
• transformations: (arith-to-varith) Support more cases @n-io (#3330)
• transforms: Add convert-varith-to-arith pass @emmau678 (#3309)
• transformations: (eqsat) add pass to convert non-eclass functions to eclass @jianyicheng (#3189)
• transformation: (convert-varith-to-arith) Minor fix to filecheck @n-io (#3329)
• transformations: Support devito timers in the csl pipeline @n-io (#3312)
• transformations: (cf) switch canonicalization @alexarice (#3291)
• transformations: (lower-csl-stencil) Add iter args to first region @n-io (#3304)
• transformations: (cf) cf.cond_br truth propagation @alexarice (#3285)
• transformations: (cf) cf.cond_br identical successors @alexarice (#3284)

Backend

• transformations: (convert-riscv-to-llvm) @jorendumoulin (#2468)
• transformations: ptr -> riscv conversion @mamanain (#3393)
• backend: (riscv) remove optimization level from test-lower-linalg-to-snitch @superlopuh (#3385)
• fix: (csl) Print float zero as 0.0, not 0 @dk949 (#3339)
• dialects: (csl) Add ptrcast op @n-io (#3313)

Interpreter

• interpreter: (pdl) add the implementation of type interpretation @qaco (#3498)
• interpreter: move ptr.py to utils folder @superlopuh (#3471)
• interpreter: (pdl) add test for attribute matching @superlopuh (#3467)
• interpreter: (pdl) add test for operation matching @superlopuh (#3466)
• interpreter: (pdl) add test for result matching @superlopuh (#3465)
• interpreter: (pdl) fix match_operand matchi...

v0.23.0

xDSL framework

• transforms: (canonicalize) Add full dce @alexarice (#3267)
• core: Fix parsing for hex denseAttrs @vinayakdsci (#3239)
• core: Block predecessors @alexarice (#3235)

Dialects

• dialects: (llvm) Add support for overflow flags @lfrenot (#3269)
• dialects: (cf) cond_br folding @alexarice (#3283)
• dialects: (arith) split generic binary op definition into specific ones @superlopuh (#3274)
• dialects: (llvm) ArithmeticBinOpBase is always on integers @superlopuh (#3270)
• dialects: (mod_arith) create mod_arith dialect and add initial operation @emmau678 (#3218)
• dialects: (stencil) add StencilType constraints @superlopuh (#3254)
• dialects: (builtin) add AnyMemRefTypeConstr and AnyTensorTypeConstr @superlopuh (#3246)
• dialects: (varith) Add varith (variadic arithmetic) dialect @AntonLydike (#3241)
• dialects: (cf) br folding @alexarice (#3240)
• dialects: (dmp) Emit multiple swaps if halo is larger than core size @AntonLydike (#3238)
• tool: tblgen-to-py script @alexarice (#3210)
• dialects: (stim) Add first annotation op @kimxworrall (#3115)
• dialects: (cf) branch canonicalization. @alexarice (#3234)
• dialects: (cf) Assert canonicalization pattern @alexarice (#3228)
• dialects: (riscv) canonicalization pattern added (x|0=x) @lfrenot (#3230)
• dialects: (stim) add qubit attribute and qubit coordinate attribute @kimxworrall (#3114)
• dialects: (riscv) Add rewrite pattern to optimize bitwise xor by zero @emmau678 (#3197)
• dialects: (tosa) turn attributes into properties @jorendumoulin (#3188)
• dialects: (csl) add csl.activate @dk949 (#3191)
• fix: various minor fixes for the csl variables @dk949 (#3184)
• dialects: (scf) small refactor for index_switch @alexarice (#3178)
• dialects: (eqsat) add dialect and create new eclass op @superlopuh (#3175)
• dialects: (cf) add switch @alexarice (#3156)
• dialects: (scf) add index_switch @alexarice (#3157)
• dialects: llvm.InLineAsmOp has_side_effects, is_align_stack are properties @JosseVanDelm (#3147)
• dialects: (scf) make iteration variable type explicit @alexarice (#3138)
• dialects: (cf) Custom Parsing and assert change @alexarice (#3132)
• dialects: (csl) Add direction type @n-io (#3136)
• dialects: (csl) Adds variables to the dialect @dk949 (#3183)
• dialects: (builtin) add i8 default types @jorendumoulin (#3164)
• dialects: (scf) fix scf.if printing bug @alexarice (#3160)
• dialects (tosa): initialize dialect and add Clamp and Rescale operation @jorendumoulin (#3158)

Transformations

• dialects: (cf) cond_br folding @alexarice (#3283)
• transformations: (lower-csl-stencil) Optimise full-stencil access @n-io (#3271)
• transforms: (cf) fix passthrough operands @alexarice (#3282)
• transformations: (memref-to-dsd) Support pre-existing GetMemDsd ops @n-io (#3279)
• transforms: (canonicalize) Add full dce @alexarice (#3267)
• fix: add extension to filename when calling set_tile_code @dk949 (#3268)
• transforms: (dce) region-level dce @alexarice (#3253)
• transformations: Add convert-arith-to-varith pass @AntonLydike (#3242)
• transforms: (canonicalize) improve dead code elimination @alexarice (#3252)
• dialects: (cf) br folding @alexarice (#3240)
• transformations: (lower-csl-stencil) Promote args before outlining @n-io (#3237)
• transformations: (memref-to-dsd) Handle csl variables @n-io (#3236)
• dialects: (cf) branch canonicalization. @alexarice (#3234)
• dialects: (cf) Assert canonicalization pattern @alexarice (#3228)
• transformations: New stencil-shape-minimize pass @n-io (#3229)
• transformations: New csl-stencil-materialize-stores pass @n-io (#3222)
• transformations: (lower-csl-stencil) Send only core data @n-io (#3223)
• transformations: (licm) add can_be_hoisted helper function @RRavikiran66 (#3078)
• fix: (csl) adjusted width and height of the PE @dk949 (#3209)
• transformations: Add CSL dsd canonicalisation @n-io (#3208)
• transformations: (scf-to-cf) add lowering for index_switch @alexarice (#3179)
• transformations: (lower-csl-stencil) Store results to apply.dest @n-io (#3203)
• transformations: (csl-stencil-to-csl-wrapper) Add unblock_cmd_stream call @n-io (#3198)
• transformations: (csl-stencil) Add pass to handle async ops and enclosing cf @n-io (#3192)
• transformation: Scf to cf lowering @alexarice (#3153)
• transformations: Add new csl-wrapper-hoist-buffers pass @n-io (#3165)
• fix: Add the missing call to @set_rectangle @dk949 (#3149)
• transformations: (csl) add csl.rpc op @dk949 (#3143)
• transformations: Lower csl_wrapper.import to csl.import @dk949 (#3137)
• transformations: Add pass to lower csl-stencil dialect @n-io (#3134)
• transformations: from csl_wrapper to csl modules @dk949 (#3107)

Backend

• backend: (csl) Array literals @n-io (#3292)
• backend: (csl) Support zero-stride dsd printing @n-io (#3280)
• backend: (csl) allow get_dir to be inlined @dk949 (#3202)
• backend: (csl) added arith.select to csl backend @dk949 (#3199)
• backend: (csl) Don't use CSL keywords as var names @dk949 (#3152)
• fix: Minor corrections to the CSL backend @dk949 (#3142)
• backend: (csl) Add comparison operators to CSL @dk949 (#3139)

Interpreter

• interpreter: use interpreter default bitwidth in run main @superlopuh (#3034)

API

• api: add replace_by_if helper to control ssa value replacement @superlopuh (#3196)

🐛 Bug Fixes

• transforms: (cf) fix passthrough operands @alexarice (#3282)
• bug: Tighten type constraints on DenseArrayBase.create_dense_int_or_index @vinayakdsci (#3258)
• transformations: (lower-csl-stencil) Promote args before outlining @n-io (#3237)
• bug: Print csl namehint fix @n-io (#3207)
• bug: (csl-stencil-handle-async-flow) Increment counter correctly @n-io (#3201)

Testing

• transforms: (canonicalize) improve dead code elimination @alexarice (#3252)
• dialects: (cf) br folding @alexarice (#3240)
• core: Block predecessors @alexarice (#3235)
• dialects: (cf) branch canonicalization. @alexarice (#3234)
• dialects: (cf) Assert canonicalization pattern @alexarice (#3228)
• transformations: (scf-to-cf) add lowering for index_switch @alexarice (#3179)
• dialects: (cf) add switch @alexarice (#3156)
• dialects: (scf) add index_switch @alexarice (#3157)
• transformation: Scf to cf lowering @alexarice (#3153)
• dialects: (scf) make iteration variable type explicit @alexarice (#3138)
• dialects: (cf) Custom Parsing and assert change @alexarice (#3132)

Continuous Integration

• Revert "pip prod(deps): bump jax from 0.4.31 to 0.4.32 (#3161)" @superlopuh (#3166)

Installation

• pip prod(deps): bump marimo from 0.9.4 to 0.9.6 @dependabot (#3288)
• pip prod(deps): bump textual from 0.82.0 to 0.83.0 @dependabot (#3289)
• pip prod(deps): bump pyright from 1.1.345 to 1.1.384 @dependabot (#3278)
• pip prod(deps): bump marimo from 0.9.2 to 0.9.4 @dependabot (#3266)
• pip prod(deps): bump pre-commit from 4.0.0 to 4.0.1 @dependabot (#3265)
• pip prod(deps): bump marimo from 0.9.1 to 0.9.2 @dependabot (#3257)
• pip prod(deps): bump ruff from 0.6.8 to 0.6.9 @dependabot (#3256)
• pip prod(deps): bump marimo from 0.9.0 to 0.9.1 @dependabot (#3249)
• pip prod(deps): bump jax from 0.4.33 to 0.4.34 @dependabot (#3250)
• pip prod(deps): bump numpy from 2.1.1 to 2.1.2 @dependabot (#3248)
• pip prod(deps): bump pre-commit from 3.8.0 to 4.0.0 @dependabot (#3251)
• pip prod(deps): bump marimo from 0.8.22 to 0.9.0 @dependabot (#3243)
• pip prod(deps): bump textual from 0.81.0 to 0.82.0 @dependabot (#3244)
• pip prod(deps): bump onnx from 1.16.2 to 1.17.0 @dependabot (#3232)
• pip prod(deps): bump marimo from 0.8.20 to 0.8.22 @dependabot (#3226)
• pip prod(deps): bump ruff from 0.6.5 to 0.6.8 @dependabot (#3219)
• pip prod(deps): bump marimo from 0.8.17 to 0.8.20 @dependabot (#3215)
• pip prod(deps): bump textual from 0.80.0 to 0.81.0 @dependabot (#3217)
• pip prod(deps): bump textual from 0.79.1 to 0.80.0 @dependabot (#3212)
• pip prod(deps): bump marimo from 0.8.15 to 0.8.17 @dependabot (#3193)
• pip prod(deps): bump wgpu from 0.18.0 to 0.18.1 @dependabot (#3187)
• pip prod(deps): bump jax from 0.4.31 to 0.4.33 @dependabot (#3177)
• pip prod(deps): bump wgpu from 0.17.3 to 0.18.0 @dependabot (#3176)
• pip prod(deps): bump marimo from 0.8.14 to 0.8.15 @dependabot (#3168)
• pip prod(deps): bump ruff from 0.6.4 to 0.6.5 @dependabot (#3167)
• pip prod(deps): bump wgpu from 0.16.0 to 0.17.3 @dependabot (#3163)
• pip prod(deps): bump jax from 0.4.31 to 0.4.32 @dependabot (#3161)
• pip prod(deps): bump filecheck from 1.0.0 to 1.0.1 @dependabot (#3150)
• pip prod(deps): bump marimo from 0.8.13 to 0.8.14 @dependabot (#3159)
• pip prod(deps): bump marimo from 0.8.11 to 0.8.13 @dependabot (#3151)
• pip prod(deps): bump ruff from 0.6.3 to 0.6.4 @dependabot (#3148)
• pip prod(deps): bump marimo from 0.8.7 to 0.8.11 @dependabot (#3146)
• pip prod(deps): bump numpy from 2.1.0 to 2.1.1 @dependabot (#3140)

Miscellaneous

• backend: (csl) Array literals @n-io (#3292)
• misc: remove unnecessary pyright ignore comments [NFC] @superlopuh (#3286)
• misc: use VarConstraint instead of Annotated[ConstraintVar] in operation definitions @superlopuh (#3264)
• backend: (csl) Support zero-stride dsd printing @n-io (#3280)
• transformations: (memref-to-dsd) Support pre-existing GetMemDsd ops @n-io (#3279)
• misc: add more .constr helpers and constraints @superlopuh (#3273)
• fix: add extension to filename when calling set_tile_code @dk949 (#3268)
• misc: add pyright ignores for places where it's not going to be happy @superlopuh (#3263)
• misc: miscellaneous linting wrt Pylance @superlopuh (#3262)
• misc: minor linting @superlopuh (#3259)
• transformations: (memref-to-dsd) Handle csl variables @n-io (#3236)
• misc: re-run notebooks with newer version of marimo @superlopuh (#3227)
• transformations: (lower-csl-stencil) Send only core data @n-io (#...

v0.22.0

xDSL framework

• core: Assembly format shouldn't require $*SegmentSizes @alexarice (#3123)
• core: Add successors to assembly format @alexarice (#3124)
• core: (irdl) Add support for regions in optional groups to declarative assembly format @kimxworrall (#3106)
• core: Implement speculability trait and helper. @PapyChacal (#3102)
• frontend: (stablehlo) add ir generation from jaxpr @shreya-um (#3038)
• core: Implement IRDL SameSize options @PapyChacal (#3067)
• rewriting: Make modify_block_argument_type modify_value_type @PapyChacal (#3045)
• transformations: Implement shape-inference pass @PapyChacal (#3047)
• core: Improved treatment of whitespace in assembly format @alexarice (#3044)
• core: CanonicaliZation naming consistency @PapyChacal (#3040)
• core: helpers for IRDL construct type access. @PapyChacal (#3025)
• core: Implement simple helpers for construct type access. @PapyChacal (#3024)
• core: implement EffectInstance @PapyChacal (#2987)
• (core) Support UnitAttr in assembly format @n-io (#2995)
• core: (irdl) Add regions to declarative assembly format @kimxworrall (#3065)
• core: Add BitEnumAttribute @zero9178 (#2617)
• core: (affine) override python floordiv symbol @jorendumoulin (#2942)

Dialects

• dialects: (csl) added csl.addressof_fn @dk949 (#3135)
• dialects: Add NoMemoryEffect trait in various LLVM and builtin ops @JosseVanDelm (#3127)
• dialects: (riscv) canonize x ^ x = 0 @mamanain (#3121)
• dialects: (stim) initialise dialect @kimxworrall (#3042)
• dialects: (stablehlo) add stablehlo.dot @superlopuh (#3090)
• dialects: (stablehlo) add stablehlo.after_all @efferifick (#3104)
• dialects: (stablehlo) add stablehlo.case @efferifick (#3095)
• dialects: (stablehlo) add stablehlo.bitcast_convert @efferifick (#3100)
• dialects: (transform) Fix amount of returned loops in TileOp @JonasCrols (#3080)
• dialects: (stablehlo) add stablehlo.token @efferifick (#3093)
• dialects: (stablehlo) add stablehlo.precision @superlopuh (#3089)
• dialects: (stablehlo) Add support for returnOp @efferifick (#3085)
• dialects: (transform) Add structured match operation @JonasCrols (#3079)
• dialects: (stablehlo) Add support for andOp @efferifick (#3081)
• dialects: stencil: Allow loading and storing from a same field. @PapyChacal (#3071)
• dialects: Update dialects with SameSize IRDL options. @PapyChacal (#3070)
• dialects: Update tensor side-effect traits @n-io (#3069)
• transformations: Enable dmp.swap stencil bufferization. @PapyChacal (#3066)
• dialects: Update dmp.swap @PapyChacal (#3056)
• transformations: Decouple distribute-stencil from shape-inference @PapyChacal (#3054)
• dialects: dmp: make strategies attributes and carry them in IR @PapyChacal (#3050)
• dialects: (Transform) Add new selecting operations @JonasCrols (#3015)
• dialect: (irdl) Add region and attributes operations @alexarice (#3049)
• dialects: (stablehlo) add transpose @superlopuh (#3031)
• dialects: (irdl) Update for mlir compatibility @alexarice (#3002)
• dialects: (stablehlo) add elementwise add, sub, mul @superlopuh (#3030)
• dialects: (stablehlo) Add initial contribution for stablehlo dialect @efferifick (#3020)
• dialects: (csl_stencil) Enable bufferized csl_stencil.apply @n-io (#2994)
• dialects: (riscv_snitch) add vfmax.s @superlopuh (#2988)
• dialects: Add stencil features meant for bufferization @PapyChacal (#2985)
• dialects: Enable bufferized stencil.apply @PapyChacal (#2982)
• dialects: (tensor) Incorrect default 'dynamic index' on insert_slice op @n-io (#2974)
• dialects: (arith, core) Constant op to accept tensor and memref @n-io (#2969)
• (bug): linalg named ops to include hidden regions in generic print @n-io (#2957)
• (dialects): improving stencil attr syntax mlir compatibility @n-io (#2947)
• dialects: (scf) Add parse/print to scf.if @alexarice (#3126)
• dialects: (vector) add pure trait to vector.broadcast and vector.fma @knickish (#3094)
• dialects: (bufferization) Add to_memref op @n-io (#2996)
• dialects: (linalg) Add Quantized Matmul @JonasCrols (#2999)
• dialects: (stencil) Relax stencil.return for tensors @n-io (#2989)
• dialects: (transform) Add some transform dialect operations @JonasCrols (#2973)
• dialects: (tensor) add tensor.dim and tensor.cast op @jorendumoulin (#2968)
• dialects: (linalg) add correct generic printing for fill, and matmul @jorendumoulin (#2971)
• transforms: (memref_stream) add missing doc and library call in generic constructors @jorendumoulin (#2941)

Transformations

• transformations: (stencil-tensorize-z-dim) Support loops @n-io (#3118)
• transforms: Fix constant pinning breaking for external functions @AntonLydike (#3131)
• bug: (csl-stencil-to-csl-wrapper) Reverse generated operands @n-io (#3112)
• transformations: (csl-stencil-to-csl-wrapper) Generate symbol export and no memref.globals @n-io (#3101)
• transformations: Translate memref to dsd @n-io (#3092)
• transformations: Support safe inplace stencil bufferization. @PapyChacal (#3077)
• transformations: Implement control-flow-hoist @PapyChacal (#3103)
• transformations: (csl-stencil-to-csl-wrapper) Support reading properties from bufferized csl-stencil @n-io (#3083)
• transformations: (csl-stencil-bufferize) Fold csl_stencil.access that have no effect @n-io (#3084)
• transformations: stencil-bufferize side-effect analysis fix @PapyChacal (#3076)
• dialects: stencil: Allow loading and storing from a same field. @PapyChacal (#3071)
• transformations: Enable bufferized dmp.swap lowering. @PapyChacal (#3073)
• transformations: Forgotten xdsl/transforms/shape_inference_patterns/init.py @PapyChacal (#3072)
• transformations: (stencil-to-csl-stencil) Support also bufferized dmp.swap @n-io (#3068)
• transformations: Enable dmp.swap stencil bufferization. @PapyChacal (#3066)
• transformations: Tensorize new result of dmp.swap @n-io (#3063)
• dialects: Update dmp.swap @PapyChacal (#3056)
• transformations: Decouple distribute-stencil from shape-inference @PapyChacal (#3054)
• dialects: dmp: make strategies attributes and carry them in IR @PapyChacal (#3050)
• transformations: (csl-stencil-to-csl-wrapper) Translate func args to exportable memref.global @n-io (#3051)
• transformations: update stencil lowering to properly accomodate bufferized stencils. @PapyChacal (#3016)
• transformations: Implement shape-inference pass @PapyChacal (#3047)
• transformations: stencil shape inference polish @PapyChacal (#3046)
• transformations: stencil bufferization polish @PapyChacal (#3001)
• transformations: Add linalg-to-csl pass @n-io (#3028)
• transformations: (csl-stencil-bufferize) Inject iter_arg into linalg compute @n-io (#3033)
• transformations: (csl_stencil) Add bufferization pass @n-io (#3004)
• transformations: First implementation of stencil-bufferize @PapyChacal (#2983)
• transformations: Enable bufferized stencil lowering. @PapyChacal (#2984)
• transformations: (stencil-tensorize-z-dim) Scaling z-shifts to avoid negative offsets @n-io (#2975)
• transform: (stencil-tensorize-z-dimension) Tensorize arith.constant directly @n-io (#2970)
• transformations: lift arith to linalg named ops @n-io (#2963)
• (transforms): dont have tensor inside memref @n-io (#2962)
• (transform): csl_wrapper program module init @n-io (#2891)
• (transform): csl_stencil canonicalization pass @n-io (#2814)
• transformations: (memref_stream) add memref_stream.generic legalization pass @nazavode (#2929)

Backend

• fix: Param names were not correctly referenced in the printer @dk949 (#3108)
• backend: (riscv) print list of registers to regalloc stats @superlopuh (#3111)
• backend: (JAX) add some syntax sugar for typed compiled functions @superlopuh (#3058)
• backend: (JAX) add export to JAX executable @superlopuh (#3057)
• backend: (riscv) fix register allocation stats @superlopuh (#2961)
• backend: (riscv) add option to insert regalloc stats @superlopuh (#2958)
• backend: (snitch) add an optimization-level option to linalg-snitch pass @superlopuh (#2956)
• backend: (snitch) unify all the test passes as a single end-to-end pass @superlopuh (#2955)
• dialects: (riscv_snitch) lower arith.addf on vector<4xf16> to vfadd.h @nazavode (#2934)

Interpreter

• interpreter: fix int64 format in ptr dtypes @superlopuh (#3035)

API

• api: deprecate print_string_raw in favor of indent=0 @superlopuh (#2951)
• api: remove parameters field from Trait @superlopuh (#2940)

🗎 Documentation

• docs: (marimo) use carousel in linalg snitch notebook @superlopuh (#2933)

🐛 Bug Fixes

• core: Assembly format shouldn't require $*SegmentSizes @alexarice (#3123)
• tests: Increase timeout on lit tests @alexarice (#3125)
• bug: remove duplicate isattr definition @superlopuh (#3088)
• transformations: stencil-bufferize side-effect analysis fix @PapyChacal (#3076)
• dependencies: fix numpy version for subprojects that depend on it @superlopuh (#3062)
• interpreter: fix int64 format in ptr dtypes @superlopuh (#3035)
• bug: (snitch_stream) fix stride pattern repeat param parsing @superlopuh (#2986)
• transformations: (stencil-tensorize-z-dim) Scaling z-shifts to avoid negative offsets @n-io (#2975)
• dialects: (arith, core) Constant op to accept tensor and memref @n-io (#2969)
• backend: (riscv) fix register allocation stats @superlopuh (#2961)
• (bug): linalg named ops to include hidden regions in generic print @n-io (#2957)
• (bug): fix csl_stencil canonicalise @n-io (#2949)

Testing

• tests: Increase timeout on lit tests @alexarice (#3125)
• dialect: (irdl) Add region and attributes operations @alexarice (#3049)
• core: Improved treatment of whitespace in assembly format @alexarice (#3044)
• dialects: (irdl) Update for mlir compatibility @alexarice (#3002)
• misc: Add Dialect.split_name method @alexarice (#3003)
• testing: add relu f32 to bottom-up tests @SuperLo...

v0.21.1

xDSL framework

• core: clone block argument name hints @superlopuh (#2910)
• core: Implement getAffine for MemrefLayout attributes @jorendumoulin (#2922)
• core: Add FixedbitWidthType Interface @jorendumoulin (#2904)
• core: fix default builtin type definitions @jorendumoulin (#2907)

Dialects

• dialects: (riscv_snitch) add vfmac.s and vfsum.s @superlopuh (#2932)
• dialects: (snitch_stream) add repetition to stride pattern @superlopuh (#2927)
• dialects: (riscv_snitch) add fastmath flags in vector ops @superlopuh (#2913)
• dialects: (Qref) fix bug in cnot initialiser @alexarice (#2899)
• dialects: (memref_stream) add doc and library call support @jorendumoulin (#2928)

Transformations

• (transform, minor): stencil-to-csl-stencil retain interesting compute structure property @n-io (#2926)
• transformations: use block argument types when streamifying @superlopuh (#2920)
• transformations: (snitch) add comments when lowering snitch config ops @superlopuh (#2925)
• transformations: handle packed simd vectors in stream op lowering @superlopuh (#2915)
• transformations: add convert-arith-to-riscv-snitch pass @superlopuh (#2914)
• (transform): csl_stencil erase merged prefetch op @n-io (#2908)

Backend

• backend: add RegisterConstraints and use in riscv backend @superlopuh (#2930)
• backend: (riscv) add allocate_same method to register allocator @superlopuh (#2924)
• backend: (riscv) add a reserve_registers helper to register queue @superlopuh (#2923)

Interpreter

• interpreter: support callable operation impls @superlopuh (#2921)

🗎 Documentation

• docs: (marimo) add interpreter trace to linalg snitch notebook @superlopuh (#2931)
• docs: (marimo) move non-onnx part of notebook to main marimo folder @superlopuh (#2903)

🐛 Bug Fixes

• bug: wrong dataclass imported in stencil_tensorize_z_dimension @superlopuh (#2906)
• bug: ship cmath.irdl file in pypi @superlopuh (#2905)
• dialects: (Qref) fix bug in cnot initialiser @alexarice (#2899)

Testing

• transformations: don't check element type of memref_stream operands @superlopuh (#2916)
• testing: add ssum testcase to bottom-up tests @superlopuh (#2912)

Continuous Integration

• ci: use uppercase I64, I32 etc as constraints in operation definitions @superlopuh (#2909)

Installation

• pip prod(deps): bump marimo from 0.7.8 to 0.7.9 @dependabot (#2918)
• pip prod(deps): update pytest requirement from <8.3 to <8.4 @dependabot (#2917)
• pip prod(deps): bump ruff from 0.5.3 to 0.5.4 @dependabot (#2919)
• pip prod(deps): bump textual from 0.72.0 to 0.73.0 @dependabot (#2902)
• pip prod(deps): bump marimo from 0.7.7 to 0.7.8 @dependabot (#2900)
• pip prod(deps): bump ruff from 0.5.2 to 0.5.3 @dependabot (#2901)
• pip prod(deps): bump marimo from 0.7.5 to 0.7.7 @dependabot (#2896)
• pip prod(deps): bump pytest-asyncio from 0.23.7 to 0.23.8 @dependabot (#2898)

Miscellaneous

• (transform, minor): stencil-to-csl-stencil retain interesting compute structure property @n-io (#2926)
• (transform): csl_stencil erase merged prefetch op @n-io (#2908)
• dialects: (Qref) fix bug in cnot initialiser @alexarice (#2899)

v0.21.0

xDSL framework

• ci: ignore more pyright errors in the frontend @superlopuh (#2895)
• core: (rewriting) Fix recursive type conversion in block arguments for nested regions @webmiche (#2871)
• core: Implement EffectKind @PapyChacal (#2793)
• core: Add option to clone name hints @superlopuh (#2822)
• core: Make print_string aware of indentation @Moxinilian (#2809)
• core: store blocks as doubly-linked list on Region [2/2] @superlopuh (#2795)
• core: add RegionBlocks to abstract away the block storage [1/2] @superlopuh (#2794)
• core: Implement block dominance @PapyChacal (#2790)
• core: replace BlockReverseOps with a Reversible conformance @superlopuh (#2792)
• misc: Small refactor to PatternRewriter.replace_op @alexarice (#2789)
• core: [NFC] Split irdl.py @PapyChacal (#2705)
• core: make VarOperand, VarOpResult, and VarRegion tuple instead of list @superlopuh (#2767)
• core: Move MLContext out of ir, move get_all_dialects to dialects. @PapyChacal (#2749)
• core: Implement MemrefLayoutAttr base class and use in parser. @PapyChacal (#2718)
• core: (rewriter) Correctly pass safe_erase in replace_op @AntonLydike (#2707)
• core: Allow to parse string enums @math-fehr (#2696)
• core: Implement and use ConstraintContext @PapyChacal (#2700)
• core: allow to define a RangeConstraint on a region's entry arguments. @PapyChacal (#2699)

Dialects

• dialects: add parametrised rz gate @alexarice (#2821)
• (dialect): minor changes to csl and csl_wrapper @n-io (#2888)
• dialects: (riscv_snitch) remove always ZERO rd in snitch config ops @superlopuh (#2886)
• (dialects): adding csl_wrapper dialect @n-io (#2867)
• dialects: (riscv_snitch) Add f32 mul, add, pack from Snitch packed SIMD extension @nazavode (#2872)
• dialects: (linalg) Fix matmul custom syntax and custom init @qaco (#2852)
• dialects: (memref_stream) remove unused operands in canonicalization @superlopuh (#2851)
• dialects: (affine) Add custom syntax to affine.apply @superlopuh (#2848)
• dialects: (transform) Define non-parametrized transform types @qaco (#2860)
• dialects: (snitch_stream) add custom syntax to streaming region @superlopuh (#2859)
• dialects: (memref) add custom syntax for memref_subview @superlopuh (#2840)
• dialects: (riscv) remove noisy () -> in get_(float_)register and li @superlopuh (#2837)
• dialects: (memref) Make memref.extract_strided_metadata have NoMemoryEffect @AntonLydike (#2844)
• dialects: (memref) Make memref.extract_strided_metadata have NoMemoryEffect @AntonLydike (#2842)
• dialects: (memref_stream) add verification for block arguments @superlopuh (#2838)
• dialects: (memref_stream) print memref_stream.generic with newlines @superlopuh (#2815)
• dialects: (memref_stream) add interleaved iterator @superlopuh (#2813)
• dialects: (memref_stream) make bounds in dialect index typed not IntAttr @superlopuh (#2810)
• dialects: (riscv) RISCVOp -> RISCVAsmOperation subclass of IRDLOperation @superlopuh (#2805)
• dialects: (wasm) Add wat encoding @superlopuh (#2801)
• (dialects): csl_stencil apply op @n-io (#2781)
• dialects: (riscv) Make riscv.get_register pure @AntonLydike (#2797)
• dialects: (wasm) Add binary encoding interface @superlopuh (#2788)
• dialects: (riscv) Add HasInsntrait and implement trait for some operations @AntonLydike (#2784)
• dialects: Implement MemoryEffect on RISCV op, just reuse main CSE. @PapyChacal (#2685)
• dialects: (riscv, x86) do not print empty <> for unallocated registers @superlopuh (#2786)
• dialects: (wasm) Add wasm dialect @superlopuh (#2780)
• dialects: (riscv) Fix I, S and shift operation immediate bounds @superlopuh (#2785)
• dialects: (memref_stream) Add memref_stream.fill @superlopuh (#2772)
• dialects: (memref_stream) Add an inits field to memref_stream.generic [1/3] @superlopuh (#2763)
• dialects: (csl) Add csl.constants op and printing @AntonLydike (#2782)
• (minor): allow csl_stencil.access to operate on own data @n-io (#2777)
• dialects: (memref) canonicalize removes unused memref allocs @superlopuh (#2774)
• dialects: Add qref dialect @alexarice (#2769)
• (dialect+transform): stencil to new csl_stencil dialect and transform @n-io (#2766)
• dialects: (builtin) Fix printing and parsing of UnitAttr @superlopuh (#2744)
• dialects: Add qssa dialect @alexarice (#2746)
• dialects: Add memref.extract_strided_metadata op @JosseVanDelm (#2748)
• dialects: (csl) Improve some constructor inits @AntonLydike (#2745)
• dialects: (csl) adding constructors for CSL ops @n-io (#2742)
• dialects: (memref_stream) Add some more verification for generic op @superlopuh (#2721)
• dialect: (csl) Add signedness cast and concat struct ops @AntonLydike (#2714)
• dialects: (csl) Small incremental fixes to dialect structure (stacked PR) @AntonLydike (#2711)
• dialects: (csl) set_tile_code fix traits @dk949 (#2712)
• dialects: Add NoMemoryEffects to memref and snitch_runtime dialect ops @AntonLydike (#2706)
• dialects: Define memref.atomic rmw @PapyChacal (#2702)
• dialects: (csl) print dsd ops @n-io (#2694)
• dialects: (arith) Add verifier to 'arith.index_cast' op and fix returns in inits @AntonLydike (#2682)
• dialects: (csl) builtin math lib for DSDs @n-io (#2686)

Transformations

• (transform): csl-stencil to csl-wrapper pass @n-io (#2878)
• transformations: Add memref-stream-tile-outer-loops @superlopuh (#2869)
• transformations: handle strided layout attr in memref_stream lowering @superlopuh (#2866)
• transformations: Handle interleaving in convert-memref-stream-to-loops @superlopuh (#2831)
• transformations: convert-memref-stream-to-snitch-stream fix name @superlopuh (#2865)
• transformations: Handle affine.apply in lower-affine @superlopuh (#2849)
• transformations: Add memref.subview lowering to convert-memref-to-riscv @superlopuh (#2847)
• transformations: Add dealloc lowering to convert-memref-to-riscv @superlopuh (#2856)
• transformations: Handle more division sizes in memref-stream-interleave @superlopuh (#2845)
• transformations: Handle more data types in memref to riscv lowering @superlopuh (#2846)
• transformations: Fix subview hack in stencil conversion @PapyChacal (#2843)
• (transform): convert to csl_stencil.apply @n-io (#2803)
• transformations: add memref-stream-interleave @superlopuh (#2823)
• dialects: Implement MemoryEffect on RISCV op, just reuse main CSE. @PapyChacal (#2685)
• transformations: Add memref-stream-(infer|generalize)-fill passes @superlopuh (#2776)
• transformations: Only stream used inputs in memref-streamify @superlopuh (#2775)
• transformations: Add memref-stream-fold-fill @superlopuh (#2773)
• transformations: Fix bug in memref-stream-unnest-out-parameters @superlopuh (#2771)
• transformations: memref_streamify handle constant inits [3/3] @superlopuh (#2765)
• transformations: Support constant inits in memref_stream.generic lowering [2/3] @superlopuh (#2764)
• transforms: (convert-snrt-to-riscv) Convert to arith where possible @AntonLydike (#2783)
• transformations: add qref/qssa conversions @alexarice (#2770)
• (minor): allow csl_stencil.access to operate on own data @n-io (#2777)
• (dialect+transform): stencil to new csl_stencil dialect and transform @n-io (#2766)
• transformations: fix yielding of values in memref_stream.generic lowering @superlopuh (#2760)
• transformations: do not insert affine.apply ops when streaming @superlopuh (#2758)
• transforms: stencil-tensorize-z-dim accept stencil.field as function arg @n-io (#2756)
• transforms: (experimental) minor updates to stencil-tensorize-z-dim @n-io (#2741)
• transformations: Add support for strided MemRefs in 'load' and 'store' @zero9178 (#2854)

Backend

• backend: (riscv) preserve name hints when casting operands @superlopuh (#2858)
• backend: (riscv) don't print external function declaration in assembly @superlopuh (#2855)
• backend: reserve registers not preserved across function calls in RISC-V @superlopuh (#2853)
• backend: (riscv) move register-allocation tests in their own directory @tobiasgrosser (#2811)
• dialects: (riscv) Add HasInsntrait and implement trait for some operations @AntonLydike (#2784)
• dialects: (csl) Add csl.constants op and printing @AntonLydike (#2782)
• backend: (riscv) riscv-prologue-epilogue-insertion @bpervan (#2778)
• backend: (riscv) Add prologue epilogue insertion @zero9178 (#2752)
• backend: (csl) Slightly rework printer to promote certain values to inline expressions @AntonLydike (#2728)
• backend: (csl) Rework printing logic for printing individual modules @AntonLydike (#2710)
• backend: (csl) conditional statements @n-io (#2697)
• dialects: (csl) print dsd ops @n-io (#2694)

Interpreter

• interpreter: (memref_stream) add interpreter support for imperfect nesting @superlopuh (#2743)

API

• api: use new base helper in more operation definitions @superlopuh (#2893)
• api: move PunctuationSpelling to module-level @superlopuh (#2894)
• api: use constraints to carry type information in pyrdl @superlopuh (#2890)
• api: add generic parameter to attribute constraint @superlopuh (#2889)
• api: add printing and parsing of dynamic lists as in MLIR @superlopuh (#2839)
• misc: remove deprecated methods that have been deprecated for a while @superlopuh (#2729)

🗎 Documentation

• docs: (marimo) prettify onnx notebook @superlopuh (#2880)
• docs: (toy) remove all custom instructions from Toy @superlopuh (#2863)
• docs: add onnx demo notebook @superlopuh (#2796)
• documentation: allow marimo notebooks as documentation @superlopuh (#2787)

🐛 Bug Fixes

• (minor): print properties on csl_stencil.apply @n-io (#2868)
• bug: (snitch_stream) simplify crashes when all bounds are 1 @superlopuh (#2818)
• backend: (riscv) riscv-prologue-epilogue-insertion @bpervan (#2778)
• transformations: fix yielding of values in memref_stream.generic lowering @superlopuh (#2760)

Testing

• testing: fix irdl-to-pyrdl filecheck @superL...

v0.20.0

xDSL framework

• core: unify (Pattern)Rewriter insertion and inlining APIs around InsertPoint @superlopuh (#2660)
• core: (irdl) Allow ClassVar with all uppercase name @AntonLydike (#2687)
• core: implement RangeConstraint and use it without functional change @PapyChacal (#2675)
• core: add inline_block_at_location to PatternRewriter @superlopuh (#2664)
• core: make AttrConstraint frozen @superlopuh (#2659)
• core: preserve name hints when replacing ops with multiple results @superlopuh (#2655)
• rewriting: Add convert-ml-program-to-memref @superlopuh (#2580)
• core: Implement a simplified MemoryEffect interface. @PapyChacal (#2640)
• core: Make printer scoped. @PapyChacal (#2614)
• core: Implement and test eq and hash on OpOperands. @PapyChacal (#2639)
• core: Expect Data's carried type to be Hashable. @PapyChacal (#2630)
• core: Refactor irdl_op_definition @math-fehr (#2575)
• core: fix line numbers in split input mode. @PapyChacal (#2563)
• core: Block inlining API improvements @PapyChacal (#2572)
• core: Strip multi-digit suffixes in name_hint, not only single-digit. @PapyChacal (#2551)
• core: Fix xDSL printing multiple values with the same name. @PapyChacal (#2545)
• core: add argument and result types getters to CallableOpInterface @superlopuh (#2535)
• core: Allowing overload of ClassVar from superclass @n-io (#2690)

Dialects

• dialects: (memref_stream) add a verifier for order of iterator types in memref_stream.generic @superlopuh (#2680)
• dialects: (riscv_func) mark riscv_func.func as isolated from above @superlopuh (#2669)
• dialects: (memref_stream) use memref_stream iterator type, not linalg @superlopuh (#2665)
• dialects: (riscv) mark assembly region as isolated from above @superlopuh (#2668)
• dialects: [arith] Add extended multiplication operations @math-fehr (#2670)
• dialects: (onnx) Add onnx.Sigmoid @alecerio (#2479)
• dialects: stencil: more canonicalization @PapyChacal (#2645)
• dialects: (x86) - AVX registers + some instructions (bonus PR!) @KGrykiel (#2586)
• dialects: Update arith, math, scf and stencil side-effect traits. @PapyChacal (#2641)
• dialects: stencil: Add canonicalization patterns and related traits. @PapyChacal (#2623)
• dialects: (csl) Added operations to manage symbols @dk949 (#2616)
• dialects: arith.constant is Pure (no side effects) @PapyChacal (#2632)
• dialects: (csl) Adding Constructors for comptime_struct and color, as well as some layout functions. @dk949 (#2612)
• dialects: (csl) Added call op and tasks @dk949 (#2603)
• dialects: (hw) Add hw.module.extern @Moxinilian (#2597)
• core: Add irdl.base @math-fehr (#2581)
• dialects: (csl) Added basic types @dk949 (#2594)
• dialects: (csl) Add csl dialect ops @AntonLydike (#2591)
• dialects: (x86) PR20 - remaining x86Lite instructions + cleanup (season finale) @KGrykiel (#2573)
• dialects: (x86) PR19 - move assembly helpers to separate class @KGrykiel (#2570)
• dialects: (x86) PR18 - m_pop + added rsp dataflow to push @KGrykiel (#2560)
• dialects: (memref_stream) add StridePattern @superlopuh (#2548)
• dialects: (x86) PR17 - m_idiv and single operand imul @KGrykiel (#2558)
• dialects: (x86) PR16 - Conditional Jumps @KGrykiel (#2552)
• dialects: (onnx) add onnx.Squeeze @alecerio (#2489)
• dialects: (x86) PR15 - S Operations @KGrykiel (#2539)
• dialects: Implement helpers on stencil attributes and simplify shape inference. @PapyChacal (#2517)
• dialects: (x86) PR14 - Labels and Directives @KGrykiel (#2538)
• dialects: (x86) - mem access helper @KGrykiel (#2537)
• dialects: (linalg) add linalg.sub @superlopuh (#2533)
• dialects: (x86) PR13 - M Operations @KGrykiel (#2531)
• dialects: (x86) PR12 - IDivOP (register version) @KGrykiel (#2530)
• dialects: (x86) PR11 - RMI Operations @KGrykiel (#2529)
• dialects: (x86) PR10 - RRI Operations @KGrykiel (#2524)
• dialects: Implement llvm.mlir.zero @PapyChacal (#2527)
• Dialects: (csl) Simplify DSD types in IR @n-io (#2683)
• Dialects: (csl) adding dsd ops @n-io (#2674)
• dialects: (csl) Data structure descriptors @n-io (#2648)
• api: Fixed builtin.DenseArrayBase not supporting IndexType @watermelonwolverine (#2633)
• dialects: (csl) add modules @dk949 (#2602)
• dialects: (hw) Add HWModuleLike trait @Moxinilian (#2577)

Transformations

• transformations: Handle imperfect loop nest when lowering memref_stream.generic to loops @superlopuh (#2678)
• transformations: add memref-stream-unnest-out-parameters @superlopuh (#2677)
• transformations: linalg.generic to loops factor out some helpers @superlopuh (#2662)
• transformations: linalg.generic to loops uses InsertPoint @superlopuh (#2661)
• transformations: scf-for-loop-folding handle iter args @superlopuh (#2657)
• transforms: Add function-constant-pinning pass to specialize functions @AntonLydike (#2389)
• transformations: fold scf loops instead of riscv_scf @superlopuh (#2656)
• transformations: Stencil shape inference fixes @PapyChacal (#2595)
• transformations: CSE: Refactor for reuse-friendliness @PapyChacal (#2642)
• transformations: Implement generic CSE. @PapyChacal (#2631)
• dialects: arith.constant is Pure (no side effects) @PapyChacal (#2632)
• transformations: fix memref_streamify non idempotency and wrong maps for outs @superlopuh (#2589)
• transformations: (riscv) handle f64 lowering in int to float conversion @superlopuh (#2561)
• transformations: (convert-snrt-to-riscv) Add lowerings for even more info ops @AntonLydike (#2428)
• transformations: implement stencil-tensorize-z-dimension @n-io (#2516)
• transforms: (convert-snrt-to-riscv) Add smaller number of ops @AntonLydike (#2559)
• transformations: Fix mlir opt printing to stderr @AntonLydike (#2553)
• transformations: stencil conversion fixes and improvements @PapyChacal (#2526)
• transformation: stencil shape inference fixes @PapyChacal (#2522)
• transformations: handle scalar arguments to generic ops when lowering @superlopuh (#2543)
• transformations: Implement and test memref-to-gpu. @PapyChacal (#2542)
• transformations: (riscv_scf) add a pass to fuse perfectly nested loops @superlopuh (#2540)
• transformations: add malloc to riscv function call lowering @superlopuh (#2532)
• dialects: Implement helpers on stencil attributes and simplify shape inference. @PapyChacal (#2517)
• transformations: add onnx.Sub to linalg lowering @superlopuh (#2534)
• transformations: Switch MPI lowering to use llvm.mlir.zero @PapyChacal (#2528)

Backend

• backend: (csl) Added printing for param, comptime_struct and some builtins @dk949 (#2684)
• backend: (csl) Maths and memory operation printing @dk949 (#2676)
• backend: (csl) printing for functions, tasks and layout related things @dk949 (#2647)
• backend: (riscv) reduce register pressure for stream configuration @superlopuh (#2621)
• backend: (riscv) Don't make unreserved registers available @superlopuh (#2620)
• backend: (riscv) Allocate ub and step before loop, and lb after @superlopuh (#2619)
• backend: (csl) Add printing some scf and memref features @dk949 (#2610)
• backend: (csl) Add printing for CSL dialect ops @AntonLydike (#2593)
• backend: (riscv) flatten loops that don't use induction variables @superlopuh (#2585)
• backend: (riscv) use fcvt.d.w for integer-valued float constants @superlopuh (#2583)
• backend: (snitch) insert fmv ops when necessary when lowering streams @superlopuh (#2574)
• backend: (riscv) move results of lowered riscv_scf.for loops @superlopuh (#2568)

Interpreter

• misc: Implement and test dynamic IRDL dialect registering. @PapyChacal (#2654)
• interpreter: use function type when parsing xdsl-run arguments @superlopuh (#2536)
• interpreter: (riscv) add fmv.d implementation @superlopuh (#2571)

🗎 Documentation

• docs: Change link to pr title formatting @AntonLydike (#2637)
• documentation: superscript i @superlopuh (#2622)

🐛 Bug Fixes

• backend: (riscv) Don't make unreserved registers available @superlopuh (#2620)
• backend: (riscv) Allocate ub and step before loop, and lb after @superlopuh (#2619)
• core: Strip multi-digit suffixes in name_hint, not only single-digit. @PapyChacal (#2551)

Testing

• testing: unflatten the loops with iteration arguments in bottom-up tests @superlopuh (#2658)
• testing: fix affine maps in matmul test @superlopuh (#2666)
• tests: Fix broken filecheck @AntonLydike (#2667)
• testing: represent kernels with no reduction in linalg.generic on memrefs @superlopuh (#2590)
• testing: make loop bounds match parallel iterators in bottom-up test @superlopuh (#2588)
• testing: (snitch) bottom-up tests have no riscv @superlopuh (#2579)
• testing: raise riscv_scf to scf in bottom-up tests @superlopuh (#2578)
• testing: no more inout parameters in bottom-up test @superlopuh (#2565)
• testing: remove dense test kernel from bottom-up tests @superlopuh (#2564)
• testing: use arith in bottom-up tests @superlopuh (#2562)
• testing: update matmul bottom-up test @superlopuh (#2547)

Installation

• pip prod(deps): bump ruff from 0.4.7 to 0.4.8 @dependabot (#2691)
• pip prod(deps): bump ruff from 0.4.6 to 0.4.7 @dependabot (#2681)
• pip prod(deps): update typing-extensions requirement from <4.12,>=4.7 to >=4.7,<4.13 @dependabot (#2644)
• pip prod(deps): bump ruff from 0.4.5 to 0.4.6 @dependabot (#2653)
• pip prod(deps): bump textual from 0.63.1 to 0.63.2 @dependabot (#2635)
• pip prod(deps): bump onnx from 1.16 to 1.16.1 @dependabot (#2636)
• Fix typing-extensions version @math-fehr (#2638)
• pip prod(deps): bump ruff from 0.4.4 to 0.4.5 @dependabot (#2627)
• pip prod(deps): bump textual from 0.62.0 to 0.63.1 @dependabot (#2625)
• pip prod(deps): bump pytest-asyncio from 0.23.6 to 0.23.7 @dependabot (#2607)
• pip prod(deps): bump textual from 0.61.1 to 0.62.0 @dependabot (#2613)
• pip prod(deps): bump textual from 0.60.1 to 0.61.1 @dependabot (#2608)
• pip prod(deps): bump ...

v0.19.0

Changes

xDSL framework

• core: Implement HasAncestor. @PapyChacal (#2514)
• core: Factor out visibility keyword parsing @Moxinilian (#2507)
• core: Handling true and false literals in DenseBaseAttr @mesham (#2492)
• core: Implement TypedAttribute, leverage in declarative assembly format. @PapyChacal (#2493)
• core: Remove redundant output from IRDLOperation's verification error. @PapyChacal (#2481)
• core: Implement unqualified attributes in declarative syntax. @PapyChacal (#2480)
• core: Implement MessageConstraint @PapyChacal (#2474)
• core: Fix OptionallyParsableDirective. @PapyChacal (#2475)

Dialects

• dialects: (x86) PR9 - MI Operations @KGrykiel (#2519)
• dialects: (x86): RI Operations adjustment @KGrykiel (#2518)
• dialects: (x86) PR8 - MR Operations @KGrykiel (#2510)
• dialects: (hw) Fix printing of module visibility @Moxinilian (#2512)
• dialects: Factor out sym_visibility parsing in all dialects @Moxinilian (#2513)
• dialects: (x86) PR7 - RI Operations @KGrykiel (#2502)
• dialects: Implement stencil custom syntax @PapyChacal (#2501)
• dialects: Simplify stencil's offset mapping implementation @PapyChacal (#2509)
• dialects: stencil IRDL update. @PapyChacal (#2500)
• dialects: (hw) Add visibility keyword to hw.module @Moxinilian (#2508)
• dialects: (HLFIR) Addition of the HLFIR dialect @mesham (#2491)
• dialects: (x86) PR6 - RM Operations @KGrykiel (#2490)
• dialects: (hw) Revert HWModuleOp's usage of InnerRefNamespaceTrait @Moxinilian (#2498)
• dialects: (hw) Add hw.instance operation @Moxinilian (#2486)
• dialects: (x86) - Reformats and changes in naming convention @KGrykiel (#2487)
• dialects: (fir) Update to the MLIR hash version of FIR that we use @mesham (#2476)
• dialects: (onnx) add onnx.Transpose @alecerio (#2477)
• dialects: (x86) PR5_3 - single operand instructions - source and destination @KGrykiel (#2485)
• dialects: (x86) PR5_2 - single operand instructions - one destination @KGrykiel (#2455)
• transformations: lower stencil.combine, refactor stencil.buffer @PapyChacal (#2457)
• dialects: (llvm) Fix elem_type bug in GEPOp builder @Moxinilian (#2472)
• dialects: (onnx) integrate onnx.Transpose in front-end @alecerio (#2488)
• frontend: (onnx) integration matmul op in onnx frontend @alecerio (#2469)

Transformations

• transformations: (onnx) fix lower onnx.Relu lowering @kayode-gif (#2435)
• transformations: lower stencil.combine, refactor stencil.buffer @PapyChacal (#2457)

🗎 Documentation

• misc: add total and last week download counts @superlopuh (#2525)

🐛 Bug Fixes

• dialects: (hw) Fix printing of module visibility @Moxinilian (#2512)
• dialects: (hw) Revert HWModuleOp's usage of InnerRefNamespaceTrait @Moxinilian (#2498)
• core: Fix OptionallyParsableDirective. @PapyChacal (#2475)
• interactive: don't include the child in the root to child list @superlopuh (#2282)
• dialects: (llvm) Fix elem_type bug in GEPOp builder @Moxinilian (#2472)

Testing

• testing: fix pyright failures when onnx is not installed @superlopuh (#2471)

Continuous Integration

• Revert "ci: use a separate workflow for codecov upload" @superlopuh (#2503)
• ci: use a separate workflow for codecov upload @superlopuh (#2497)

Installation

• dependencies: (onnx) use versioned onnx instead of weekly @superlopuh (#2515)
• pip prod(deps): bump textual from 0.58.0 to 0.58.1 @dependabot (#2520)
• pip prod(deps): update pytest requirement from <8.2 to <8.3 @dependabot (#2511)
• pip prod(deps): bump textual from 0.57.1 to 0.58.0 @dependabot (#2505)
• pip prod(deps): bump ruff from 0.3.7 to 0.4.2 @dependabot (#2504)
• pip prod(deps): bump black[jupyter] from 24.4.1 to 24.4.2 @dependabot (#2506)
• pip prod(deps): bump black[jupyter] from 24.4.0 to 24.4.1 @dependabot (#2496)
• pip prod(deps): bump textual from 0.56.4 to 0.57.1 @dependabot (#2484)

Miscellaneous

• core: Implement HasAncestor. @PapyChacal (#2514)
• core: Remove redundant output from IRDLOperation's verification error. @PapyChacal (#2481)
• core: Implement MessageConstraint @PapyChacal (#2474)
• core: Fix OptionallyParsableDirective. @PapyChacal (#2475)

Interactive

• interactive: don't include the child in the root to child list @superlopuh (#2282)
• interactive : remove bug that happens when argument screen is cleared and entered @dshaaban01 (#2364)

v0.18

xDSL framework

• frontend (onnx): implementation graph visiting @alecerio (#2464)
• frontend: (onnx) visit_node implementation @alecerio (#2441)
• core: Fix optional operand declarative printing. @PapyChacal (#2448)
• core: More informative rewrite errors. @PapyChacal (#2445)
• frontend: [onnx] add test_visit_value_info_multiple_time @alecerio (#2437)
• core: Add optional attributes to declarative assembly format @lucjaulmes (#2407)
• frontend: (onnx) Unify type builders again @alecerio (#2418)
• frontend: [onnx] unify type builders @alecerio (#2334)
• core: Make location parsing public @Moxinilian (#2371)
• core: Fix printing of non-identifier symbols @Moxinilian (#2369)
• core: Make print_ssa_value return chosen name @Moxinilian (#2370)
• rewriting: Fix TypeConversionPattern recursive support for DictionaryAttr @Moxinilian (#2374)
• core: Fix bug in pattern rewriter when ErasedSSAValue is present @math-fehr (#2365)
• core: rewrite_op returns whether an action was performed @superlopuh (#2331)
• rewriting: Extend TypeConversionPattern annotation @Moxinilian (#2332)
• core: add insert_op_at_location to PatternRewriter @superlopuh (#2308)
• core: (pattern_rewriter) Revert accidental change @AntonLydike (#2318)
• core: use InsertPoint for block inlining @superlopuh (#2311)
• core: declarative format variable fix @PapyChacal (#2313)
• core: Make SSAValue hashable @AntonLydike (#2302)
• core: move InsertPoint to Rewriter [NFC] @superlopuh (#2310)
• core: make ModulePass frozen @superlopuh (#2254)
• core: Refix opaque syntax attribute spacing @PapyChacal (#2256)
• core: Spacing fix in EnumAttribute and OpaqueSyntaxAttribute @PapyChacal (#2248)
• core: Implement first cases of assembly format optional groups @PapyChacal (#2239)
• core: Assembly format spacing fix. @PapyChacal (#2238)
• Implement and test optional operand and result variables. @PapyChacal (#2107)
• core: Add checks on chaining variadics in declarative assembly format. @PapyChacal (#2106)
• core: add Hasher and HashableModule @superlopuh (#2210)
• core: Allow to get the expected base of an IRDL constraint @math-fehr (#2175)
• core: Make IRDL operations and attributes final @math-fehr (#2159)
• core: Allow unregistered citizens in a registered dialects @PapyChacal (#2161)
• frontend: (onnx) add get_shape and get_tensor_type @superlopuh (#2134)
• core: add custom attr and prop name support to assembly format @superlopuh (#2123)
• frontend: (onnx) re-arrange elem type @alecerio (#2132)
• core: Implement Printer property retrocompatibility. @PapyChacal (#2082)
• core: add replacement arg_values to inline_block_before @superlopuh (#2061)

Dialects

• dialects: (onnx) implementation onnx.MatMul @alecerio (#2467)
• dialects: (x86) PR5_1 - single operand instructions - one source @KGrykiel (#2398)
• dialects: (seq) Add ConstClockOp @Moxinilian (#2461)
• dialects: (comb) Accept non-integer types in mux @Moxinilian (#2459)
• dialects: (comb) Fix bugs in variadic operations @Moxinilian (#2458)
• dialects: (comb) Fix lowBit case in extract op @Moxinilian (#2460)
• dialects: Implement stencil.combine @PapyChacal (#2454)
• dialects: (seq) Finish seq.compreg optionals @lucjaulmes (#2450)
• dialects: Implement stencil.dyn_access, test it (and stencil.index) @PapyChacal (#2380)
• dialects: Add accfg dialect operations @JosseVanDelm (#2430)
• dialects: (linalg) add linalg.broadcast @kayode-gif (#2433)
• dialects: (seq) Add builder and verifier for CompRegOp @Moxinilian (#2429)
• dialects: (builtin) Fix bound computation crash for i0 @Moxinilian (#2403)
• dialects: (seq) Add CompRegOp @Moxinilian (#2404)
• dialects: (riscv_snitch) Add Xdma extension operations @AntonLydike (#2410)
• dialects: (hw) Add hw.module and hw.output operations @Moxinilian (#2366)
• dialects: (linalg) add linalg.conv_2d_nchw_fchw @kayode-gif (#2400)
• dialects: (linalg) add attributes to linalg.pooling_nchw_max and base class for pooling ops @kayode-gif (#2392)
• dialects: (x86) PR4 - two-register-operand instructions @KGrykiel (#2379)
• dialects: Fix affine.ParallelOp boundsGroup checks. @PapyChacal (#2235)
• dialects: (linalg) add linalg.pooling_nchw_max @kayode-gif (#2388)
• dialects: (x86) PR3 - Assembly emission @KGrykiel (#2375)
• dialects: (x86) PR2 - Operations: add @KGrykiel (#2351)
• dialects: (comb) Fix replicate op @Moxinilian (#2372)
• dialects: (linalg) allow for empty results in linalg.fill custom format @superlopuh (#2363)
• dialects: (comb) Fix icmp parser/printer discrepancy and 2-state semantics modelling @Moxinilian (#2335)
• dialects: (x86) PR1 - Registers @KGrykiel (#2333)
• dialects: (gpu) Adding gpu.wait op @tavakkoliamirmohammad (#2298)
• dialects: (comb) Fix extract op @Moxinilian (#2327)
• dialects: (comb) Fix concat op @Moxinilian (#2326)
• dialects: (air) Add air dialect @gabrielrodcanal (#2234)
• bug: (onnx) fix onnx.Gemm @kayode-gif (#2323)
• dialects: (riscv) add index to registers @zero9178 (#2321)
• dialects: (snitch_stream) make snitch_stream parameters go outwards-in @superlopuh (#2294)
• dialects: (llvm) Fix constructor for llvm.InlineAsmOp @AntonLydike (#2312)
• dialects (hw): Fix Formatting @webmiche (#2317)
• dialects: (hw) Add InnerSymAttr attributes to HW dialect. @lucjaulmes (#2251)
• dialects: Implement tensor.insert_slice and tensor.extract_slice @PapyChacal (#2305)
• dialects: Implement bufferization.alloc_tensor and bufferization.to_tensor @PapyChacal (#2306)
• dialects: linalg minor fixes @PapyChacal (#2307)
• dialects: (linalg) add linalg.matmul @kayode-gif (#2296)
• (dialects): add linalg.tranpose @kayode-gif (#2280)
• dialects: (snitch_stream) simplify bounds of size 1 in stride pattern @superlopuh (#2276)
• dialects: (linalg) add linalg.mul @kayode-gif (#2278)
• dialects: reformat onnx.Constant to align with MNIST models one @kayode-gif (#2261)
• dialects: (riscv) mul is pure @superlopuh (#2257)
• dialects: make linalg.IteratorType an EnumAttribute @PapyChacal (#2249)
• dialects: (snitch_stream) add offsets to StridePattern and use in interpreter @superlopuh (#2252)
• dialects: (memref_stream) allow scalar argument to memref_stream generic @superlopuh (#2250)
• dialects: (memref) add custom format to memref.dealloc @superlopuh (#2247)
• onnx.reshape should check for product not length @kayode-gif (#2241)
• (dialects): add tensor.reshape @kayode-gif (#2240)
• dialects: (ml_program) add ml_program.global and ml_program.global_load_const @kayode-gif (#2209)
• dialects: Fix inline_asm unitattrs to not be properties @JosseVanDelm (#2223)
• dialects: Add LLVM InlineAsm op @JosseVanDelm (#2203)
• dialects: (hw) Add InnerSymbolTable and InnerRefNamespace to HW dialect @lucjaulmes (#1780)
• dialects: Update gpu.launch_func @PapyChacal (#2163)
• dialects: add linalg.fill @kayode-gif (#2189)
• dialects: (memref) add custom syntax to memref.alloc @superlopuh (#2191)
• dialects: (memref_stream) add memref_stream.generic @superlopuh (#2148)
• dialects: (memref_stream) add streaming region @superlopuh (#2146)
• dialects: (tensor) add tensor.Empty @kayode-gif (#2133)
• dialects: (snitch_stream) replace stride pattern op with attribute @superlopuh (#2143)
• dialects: (snitch_stream) add stride pattern attribute @superlopuh (#2141)
• dialects: (LLVM) Fix use of attributes dict when dialect specifies prop_def @ed741 (#2144)
• dialects: (aie) Add aie dialect @gabrielrodcanal (#1763)
• dialects: (snitch_stream) reverse bounds and strides in snitch_stream @superlopuh (#2139)
• dialects: (snitch_stream) remove StridedReadOp and LowerStridedWriteOp @superlopuh (#2127)
• dialects: (snitch_stream) remove stride pattern operand from strided read and write @superlopuh (#2126)
• dialects: (memref_stream) add memref_stream dialect with read and write ops @superlopuh (#2111)
• dialects: (memref) add custom syntax to memref.get_global @superlopuh (#2124)
• dialects: (linalg) add linalg.add @kayode-gif (#2129)
• dialects / bug : (linalg) linalg.generic constructor does not have a result types argument @kayode-gif (#2125)
• dialects: (stream) add abstract superclass for stream read and write @superlopuh (#2110)
• dialects: (riscv) canonicalization should move constants not replace with li @superlopuh (#2104)
• dialects: (arith) adding an arith canonicalisation pattern @dshaaban01 (#2094)
• dialects: (riscv) don't canonicalize getting the zero register to li 0 @superlopuh (#2101)
• dialects: (riscv) fix signedness in riscv @superlopuh (#2100)
• dialects: (builtin) fix ranges for integers @superlopuh (#2099)
• dialects: (riscv_scf) make riscv_scf.yield implicit @superlopuh (#2092)
• dialects: (riscv) Canonicalize the loading 0 to li zero, 0 @superlopuh (#2077)
• dialects (riscv_cf): Allow float register types for block argument types @compor (#2083)
• dialects (riscv): Add canonicalization pattern to fuse fmul.d and fadd.d to fmadd.d @compor (#2079)
• dialects: (onnx) add onnx.EntryPoint @kayode-gif (#2078)
• dialects: (onnx) add onnx.MaxPoolSingleOut @kayode-gif (#2072)
• dialects (riscv): Fix missing attribute name from custom format @compor (#2075)
• dialects: (hw) Add InnerRefNamespaceTrait to HWModuleOp @lucjaulmes (#2406)
• dialects: (pdl) Fix printer and parser for PDL range type @Moxinilian (#2325)
• dialects: (linalg) add custom printer and parser support for doc and library call @jorendumoulin (#2090)

Transformations

• dialects: Implement stencil.combine @PapyChacal (#2454)
• dialects: Implement stencil.dyn_access, test it (and stencil.index) @PapyChacal (#2380)
• transformations: implement stencil.store_result lowering. @PapyChacal (#2447)
• transformations: Improve and test stencil.index's lowering. @PapyChacal (#2444)
• core: More informative rewrite errors. @PapyChacal (#2445)
• transformations: (onnx) lower onnx.Conv to linalg @kayode-gif (#2396)
• transformations...

v0.17

Changes

xDSL framework

• core: Implement BytesAttr @PapyChacal (#2032)
• core: Implement basic ConstraintVar resolution for assembly format @PapyChacal (#1996)
• core: Add listener in PatternRewriter @math-fehr (#1905)
• core: Fix and test empty variadic directives in declarative assembly format. @PapyChacal (#2069)
• core: make ParametrizedAttribute parameters a tuple @superlopuh (#2054)
• core: properties in declarative format @math-fehr (#2031)
• core: (affine) add compress_dims @superlopuh (#2025)
• core: (affine) add used_dims helper to AffineExpr @superlopuh (#2030)
• core: Add attribute variables to the declarative assembly format. @math-fehr (#2028)
• dialects: Define nontemporal properties on memref accesses. @PapyChacal (#2024)
• dialects: (builtin) add NoneType @kayode-gif (#2016)
• core: Support variadic operands and results in declarative assembly format. @PapyChacal (#2011)
• core: Add support for nested inferrence in IRDL @math-fehr (#2005)
• core: Make assembly format use attr-dict for properties @PapyChacal (#2013)
• core: Simplify the pattern rewriter @math-fehr (#1910)
• core: use worklist in pattern matching algorithm @math-fehr (#1908)
• core: Support hex denseAttr parsing for integers, f32, f64 @JosseVanDelm (#1974)
• core: Move to assertions for modification permission checks @tarinduj (#1984)
• core: Add parsing of aliases @math-fehr (#1972)
• core: Add support for lazily loading dialects @math-fehr (#1917)
• core: (printer) use the same code to print op attrs, properties, and DictionaryAttr @superlopuh (#1954)
• core: (parser) add UnresolvedArgument with no type @superlopuh (#1952)
• core: Add region inlining helpers to rewriter (before/after/at_start/at_end) @superlopuh (#1943)
• core: add arg_types param to Block.split_before @superlopuh (#1944)
• core: add helper for function type printing @superlopuh (#1926)
• core: add method for splitting block before a given operation @superlopuh (#1936)
• core: Register passes lazily @math-fehr (#1911)
• core: Add options to walk to control traversal order @math-fehr (#1904)
• core: Add and \n directives to assembly format @math-fehr (#1915)
• core: Add a worklist for pattern rewriting @math-fehr (#1903)
• core: Add callbacks to the Builder @math-fehr (#1892)
• core: Add methods to create blocks with the builder @math-fehr (#1891)
• core: Add constructors to Builder and InsertPoint @math-fehr (#1890)
• core: Expose InsertPoint class (NFC) @math-fehr (#1888)

Dialects

• dialects: Add memref.expand_shape and memref.collapse_shape ops @JosseVanDelm (#2036)
• dialects: (onnx) add onnx.Constant @kayode-gif (#2037)
• dialects: (riscv) remove SImm and replace with annotated IntegerAttr @superlopuh (#2055)
• dialects: (riscv_scv) Add reverse order for (rof) op @AntonLydike (#2048)
• dialects: (onnx) add onnx.Conv @kayode-gif (#1961)
• dialects: Define nontemporal properties on memref accesses. @PapyChacal (#2024)
• dialects: Switch memref accesses to assembly format @PapyChacal (#2023)
• dialects: (builtin) add NoneType @kayode-gif (#2016)
• dialects: Import MemRefType from builtin. @PapyChacal (#2022)
• dialects: (onnx) add onnx.Abs @kayode-gif (#2012)
• dialects: (riscv_debug) add RISCV_Debug dialect with printf op @superlopuh (#1990)
• dialects: (onnx) add onnx.Reshape @kayode-gif (#1951)
• dialects: (snitch_stream) replace snitch_stream generic with streaming region @superlopuh (#1986)
• dialects: Add result arguments to func.func @JosseVanDelm (#1987)
• dialects: (onnx) factor out broadcastability helper in onnx dialect @kayode-gif (#1968)
• dialects: Add constant unitattr to global @JosseVanDelm (#1983)
• dialects: Implement affine.parallel @PapyChacal (#1979)
• dialects: Refresh Affine, add interoperability test. @PapyChacal (#1981)
• dialects: (riscv_snitch) add get_stream, read, write ops @superlopuh (#1970)
• dialects: (onnx) sort onnx ops alphabetically @kayode-gif (#1969)
• dialects: (onnx) lint onnx.py @superlopuh (#1966)
• dialects: (onnx) add onnx.Gemm @kayode-gif (#1945)
• dialects: (func) add support for arg_attrs in func.func @superlopuh (#1955)
• dialects: (riscv_cf) add riscv_cf dialect @superlopuh (#1946)
• dialects: (riscv_scf) add custom syntax to riscv_scf.while @superlopuh (#1934)
• dialects: (onnx) add onnx.Relu @kayode-gif (#1939)
• dialects: (scf) add custom syntax for scf while @superlopuh (#1927)
• dialects: (riscv_scf) add custom syntax for riscv_scf.condition @superlopuh (#1925)
• dialects: (scf) add custom syntax for scf.condition @superlopuh (#1924)
• dialects: (riscv) add fmv.d op @superlopuh (#1931)
• dialects: (arith) make binary ops pure @superlopuh (#1930)
• dialects (onnx): Initial onnx dialect support @brnorris03 (#1914)
• dialects (pdl): Add better verifiers to pdl.pattern @math-fehr (#1913)
• dialects: (llvm) Make llvm.alloca's alignment property optional @PapyChacal (#1887)
• dialects: Update pdl dialect to leverage properties @martin-luecke (#1989)
• dialects: Add optional alignment property to memref.global @JosseVanDelm (#1976)
• dialects: Move MemRefType et al. to builtin. @PapyChacal (#1997)

Transformations

• transforms: (riscv) Add pattern to convert move of immediate value to immediate instantiation @AntonLydike (#2057)
• Add pass to set fastmath flags of FP binary operations from arith @compor (#2018)
• backend: (riscv_cf) add convert-riscv-scf-to-riscv-cf @superlopuh (#1958)
• transforms: Canonicalize without the MLContext @math-fehr (#1912)
• transforms: (riscv_cf) Add canonicalization for branch ops with constant inputs @AntonLydike (#2049)

Backend

• Add fastmath flags to riscv dialect @compor (#2058)
• backend: (riscv) add automatic conversion of riscv_scf.for loop to ri… @superlopuh (#2033)
• backend: (riscv) add a pass to lower print_format to riscv_debug.printf @superlopuh (#1992)
• dialects: (snitch_stream) replace snitch_stream generic with streaming region @superlopuh (#1986)
• backend: (snitch) add iteration arguments to frep @superlopuh (#1982)
• backend: (snitch) read and write to streams in frep body @superlopuh (#1977)
• dialects: (riscv_snitch) add get_stream, read, write ops @superlopuh (#1970)
• backend: (riscv_cf) add convert-riscv-scf-to-riscv-cf @superlopuh (#1958)
• backend: (snitch_stream) add rank to stride pattern type @superlopuh (#1963)
• backend: (riscv) reserve iter arg from allocation inside of riscv_scf body @superlopuh (#1875)
• backend: (riscv) don't dce after lowering arith @superlopuh (#1877)

Interpreter

• interpreter: (affine) add interpreter impl for affine.apply @superlopuh (#2029)
• interpreter: (linalg) add support for reduction iterators for generic @superlopuh (#2020)
• interpreter: (riscv_debug) add implementation for riscv_debug.printf @superlopuh (#1998)
• interpreter: (riscv_snitch) add riscv_snitch functions @superlopuh (#1985)
• interpreter: add symbol parameter to xdsl-run @superlopuh (#1962)
• interpreter: (riscv_cf) add one more interpreter test @superlopuh (#1957)
• interpeter: (riscv_cf) add interpreter functions @superlopuh (#1956)
• interpreter: (linalg) allow scalar params for linalg generic @superlopuh (#1932)
• interpreter: (riscv) add interpreter impls for addi and fcvt.d.w @superlopuh (#1933)
• interpreter: (riscv_scf) add interpreter function for while op @superlopuh (#1923)

🗎 Documentation

• documentation: (Toy) add instruction on how to jit Toy using lli @superlopuh (#2002)
• documentation: (toy) remove unused printf lowering @superlopuh (#1999)
• documentation: (Toy) lower toy.print to print_format and printf @superlopuh (#1993)
• documentation: (toy) use riscv_cf in Toy @superlopuh (#1959)
• docs: Minor stencil docs improvement @georgebisbas (#1949)
• fix attribute docstring @jorendumoulin (#1909)

🐛 Bug Fixes

• core: make ParametrizedAttribute parameters a tuple @superlopuh (#2054)
• bug: fix AffineMap.inverse_permutation @superlopuh (#2019)
• backend: (riscv) reserve iter arg from allocation inside of riscv_scf body @superlopuh (#1875)

Testing

• testing: add tests for some kernels in the snitch backend paper @superlopuh (#2021)
• dialects: Import MemRefType from builtin. @PapyChacal (#2022)
• testing: (riscv) add test for riscemu emulation when lowering to riscv_cf @superlopuh (#1980)
• testing: (snitch) remove block names in test @superlopuh (#1929)
• test: update math tests' constructor @ShaolunWang (#1947)

Continuous Integration

• CI: Fix user name for api key @math-fehr (#1889)
• CI: Use PyPi token to publish package @math-fehr (#1886)

Installation

• pip prod(deps): bump ruff from 0.1.14 to 0.1.15 @dependabot (#2046)
• pip prod(deps): update pytest requirement from <8.0 to <9.0 @dependabot (#2035)
• pip prod(deps): bump pytest-asyncio from 0.23.3 to 0.23.4 @dependabot (#2034)
• pip prod(deps): bump ruff from 0.1.13 to 0.1.14 @dependabot (#1995)
• pip prod(deps): bump ruff from 0.1.12 to 0.1.13 @dependabot (#1975)
• pip prod(deps): bump ruff from 0.1.11 to 0.1.12 @dependabot (#1973)
• pip prod(deps): bump pyright from 1.1.344 to 1.1.345 @dependabot (#1953)
• pip dev(deps-dev): bump textual from 0.47.0 to 0.47.1 @dependabot (#1948)
• pip dev(deps-dev): bump textual from 0.46.0 to 0.47.0 @dependabot (#1941)
• pip prod(deps): bump textual-dev from 1.3.0 to 1.4.0 @dependabot (#1940)
• pip prod(deps): bump ruff from 0.1.9 to 0.1.11 @dependabot (#1937)
• pip prod(deps): bump pytest-asyncio from 0.23.2 to 0.23.3 @dependabot (#1935)
• pip prod(deps): bump pyright from 1.1.343 to 1.1.344 @dependabot (#1928)
• pip prod(deps): bump pyright from 1.1.342 to 1.1.343 @dependabot (#1916)
• pip prod(deps): bump ruff from 0.1.8 to 0.1.9 @dependabot (#1906)
• pip dev(deps-dev): bump wgpu from 0.13.1 to 0.13.2 @dependabot (#1907)
• pip prod(deps): bump pyright from 1.1.341 to 1.1.342 @dependabot (#1902)
• pip pr...