Bulk Operation

SYSTEMDS-1284 Linearize search space according to size of intermediates

SYSTEMDS-1284 New rewrite framework for code generation plans

SYSTEMDS-1284 Vector primitive for counting nnz

SYSTEMDS-1284 Improve cost-based optimizer for distributed operations

SYSTEMDS-1284 Support for complex sparse-unsafe binary expressions in codegen outer templates

SYSTEMDS-1284 Avoid dense conversion of sparse side inputs in row templates

SYSTEMDS-1284 Configurable codegen candidate selection policy

SYSTEMDS-1284 Column indexing on main input in row-wise ops

SYSTEMDS-1284 Generalize row-wise cbind handling

SYSTEMDS-1284 Extended plan enumeration statistics

SYSTEMDS-1284 More aggressive dynamic recompilation w/ codegen

SYSTEMDS-1284 Allow external codegen compiler configuration

SYSTEMDS-1284 Extend outer template support for matrix-scalar ops on sparse driver

SYSTEMDS-1284 Improve sparse-safe declaration of fused operators

SYSTEMDS-1284 Automatically determine max memory/compute bandwidth

SYSTEMDS-1284 Rework codegen candidate exporation algorithm

SYSTEMDS-1284 Column aggregation in cellwise templates

SYSTEMDS-1284 Column-range indexing in rowwise templates

SYSTEMDS-1284 Simplify algorithm dml scripts

SYSTEMDS-1284 Support rowwise cumsum operations

SYSTEMDS-1284 Rework codegen cost-based plan selector (opt V2)

SYSTEMDS-1284 Support for sparse-unsafe sparse vector primitives

SYSTEMDS-1284 Reduce garbage collection overhead of the codegen compiler

SYSTEMDS-1284 Full aggregation support in rowwise templates

SYSTEMDS-1284 Common subexpression elimination for codegen plans

SYSTEMDS-1284 Determine minimal number of vector intermediates

SYSTEMDS-1284 Generalized row-wise template (scalar-vector, vector indexing)

SYSTEMDS-1284 Worst-case size estimates for codegen fused operators

SYSTEMDS-1284 Improvements to row templates to capture missed opportunities

SYSTEMDS-1284 Codegen support for various patterns in nn dml library

SYSTEMDS-1284 Improve handling of sparse outputs and sideway inputs

SYSTEMDS-1284 Improve efficiency sparse-unsafe cellwise operations

SYSTEMDS-1284 Include multi-aggregate plans into cost estimation

SYSTEMDS-1284 Support codegen for matrix-matrix multiplications

SYSTEMDS-1284 Multi-aggregates w/ dot products as aggregation roots

SYSTEMDS-1284 Extend cost model for distributed operations and broadcasts

SYSTEMDS-1284 Generalize cell template for sideways row vectors inputs

SYSTEMDS-1284 Create additional common unary and binary row vector operations

SYSTEMDS-1284 Performance spark rowwise codegen instructions

SYSTEMDS-1284 Extended compiler: leave hop dags unchanged during initial compilation

SYSTEMDS-1284 Support log and exp row vector operations

SYSTEMDS-1284 Additional meta operator template: SpoofRowwise

SYSTEMDS-1284 Support cross-partition multi-aggregates with partial shared reads

SYSTEMDS-1284 Avoid dense allocation of empty sideway inputs

SYSTEMDS-1284 Support rowMins/rowMaxs in row aggregate template

SYSTEMDS-1284 Support 'replace' in row aggregate and cell templates

SYSTEMDS-1284 Multi-threaded compilation of fused operators

SYSTEMDS-1284 Handling of plan selection constraints (e.g., memory/blocksize)

SYSTEMDS-1284 Vector primitives for row comparisons

SYSTEMDS-1284 Fuse row aggregate w/ colvector output into cell template

SYSTEMDS-1284 Support min/max/sumsq in cell templates w/ aggregation

SYSTEMDS-1284 Cost model for candidate selection

SYSTEMDS-1284 Sparsity-exploiting cellwise template

SYSTEMDS-1284 Support for different plan selection policies

SYSTEMDS-1284 Support right indexing in cellwise and rowaggregate templates

SYSTEMDS-1284 Simplify cplan construction algorithm

SYSTEMDS-1284 Generalization cellwise template (add mm as aggregation root)

SYSTEMDS-1284 Generalization cellwise template (from sideways vectors to matrices)

SYSTEMDS-1284 Hardening sparse-safe check cellwise operations

SYSTEMDS-1284 Rework codegen algorithm testcases (use current algorithms)

SYSTEMDS-1284 Memory management temporary vector intermediates

SYSTEMDS-1284 Codegen for existing cellwise fused operators

SYSTEMDS-1284 Enable spoof instructions in parfor remote spark

SYSTEMDS-1284 Removal unnecessary int-double casts

SYSTEMDS-1284 Perftest benchmark codegen

SYSTEMDS-1284 Extended explain tool (for generated java/byte code)

SYSTEMDS-1284 Extended stats tool (code generation statistics)

SYSTEMDS-1284 In-memory source code compilation

SYSTEMDS-1284 Disable counter-productive existing fused operators

SYSTEMDS-1284 Extended compiler: materialization decisions

SYSTEMDS-1284 Support spark codegen instructions w/ multiple RDD inputs

SYSTEMDS-1284 Size bounding plan cache

SYSTEMDS-1284 Hardening transfer of generated operators

SYSTEMDS-1284 Support compressed matrix blocks

SYSTEMDS-1284 Additional meta operator template: MultiAggregate

SYSTEMDS-1284 Runtime integration codegen

SYSTEMDS-1284 Compiler integration codegen

SYSTEMDS-1284 Basic code generator