| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2016-09-12 | Rework map functions as class methods | jackkoenig | |
| Changed code from match statements in Mappers.scala to methods on the various IR classes. This allows custom IR nodes to implement the mapper functions and thus work (ie. not match error) when map is called on them. This also should have a marginal performance increase because of use of virtual function calls rather than match statements. | |||
| 2016-09-12 | Fixed bug where nodes of Invalids where created | azidar | |
| 2016-09-12 | Bug fix -- remove all empty expressions after ReplSeqMem passes (#294) | Angie Wang | |
| * Bug fix -- remove all empty expressions after ReplSeqMem passes * Added test to make sure ReplSeqMem can handle BundleType SMem (EmptyExpression leakage) | |||
| 2016-09-08 | Revert Expand Whens to process a set of connection lhs | jackkoenig | |
| 2016-09-08 | memoize nodes in ExpandWhens | Donggyu Kim | |
| 2016-09-08 | remove Utils.{AND, OR, NOT, EQV} | Donggyu Kim | |
| hidden const props not desirable | |||
| 2016-09-08 | clean up ExpandWhens | Donggyu Kim | |
| 2016-09-08 | refactor RemoveCHIRRTL | Donggyu Kim | |
| 2016-09-08 | refactor resolves | Donggyu Kim | |
| 2016-09-08 | refactor InferTypes | Donggyu Kim | |
| 2016-09-08 | split Passes.scala into multiple files(InferTypes.scala, Resolves.scala, ↵ | Donggyu Kim | |
| RemoveCHIRRTL.scala) | |||
| 2016-09-08 | refactor InferWidths | Donggyu Kim | |
| 2016-09-07 | put InferWidths in a seperate file and fix spaces | Donggyu Kim | |
| 2016-09-07 | refactor checks | Donggyu Kim | |
| 2016-09-07 | clean up PrimOps | Donggyu Kim | |
| 2016-09-07 | clean up miscs | Donggyu Kim | |
| 2016-09-07 | clean up WIR.scala | Donggyu Kim | |
| 2016-09-07 | Merge branch 'master' into cleanup_passes | Adam Izraelevitz | |
| 2016-09-07 | clean up Emitter.scala (#275) | Donggyu | |
| 2016-09-07 | clean up SplitExpressions | Donggyu Kim | |
| 2016-09-07 | clean up LowerTypes | Donggyu Kim | |
| 2016-09-07 | clean up PullMuxes & ExpandConnects | Donggyu Kim | |
| 2016-09-07 | clean up Utils.scala | Donggyu Kim | |
| remove unnecessary functions & change spaces | |||
| 2016-09-07 | remove Utils.ONE | Donggyu Kim | |
| 2016-09-07 | remove Utils.tpe | Donggyu Kim | |
| 2016-09-07 | Utils.scala: remove vars with fold | Donggyu Kim | |
| 2016-09-07 | Added ReplaceSubAccesses before RemoveSubAccesses | azidar | |
| 2016-09-07 | add caches for create_exps in RemoveAccess | Donggyu Kim | |
| 2016-09-07 | clean up RemoveAccesses | Donggyu Kim | |
| 2016-09-06 | optimize equals of WrappedExpression & WrappedType | Donggyu Kim | |
| 2016-09-06 | remove unnecessary mappers in Namespace | Donggyu Kim | |
| there's a big stack overhead with mappers | |||
| 2016-09-06 | replace flatMap with foldLeft in create_exps | Donggyu Kim | |
| internal implementation for flatMap seems to be inefficient | |||
| 2016-09-06 | Address style feedback and add tests for getConnectOrigin utility | Angie | |
| 2016-09-06 | Support optionally filling write mask to data width via transform input ↵ | Angie | |
| config file | |||
| 2016-09-06 | Expanded annotations for valid memory sizes | Angie | |
| 2016-09-06 | Edited conf generation to handle mem namespace collision | Angie | |
| * Also started separate pass for annotating valid memory | |||
| 2016-09-06 | Made the connect origin function more powerful | Angie | |
| * It analyzes through statements that ConstProp would've optimized * Edge case wmask can be removed (pass tries harder to figure out that wmask = wen) | |||
| 2016-09-06 | Added simple unit test for ReplSeqMem | Angie | |
| 2016-09-06 | Added back support for conf writing. | Angie | |
| * Conf file info is passed in through annotations. * A pass should have its own set of sub-arguments delimited by : | |||
| 2016-09-06 | Changed wmask to convert from VecType to UInt | Angie | |
| * Instead of filling the whole data width * Added helper functions in MemUtils | |||
| 2016-09-06 | Replace DefMemories with wrapped black box | Angie | |
| * Note, this version uses Albert's toBitMask function, * which expands the bit mask to be the full data width (similar to Chisel2 output) * Black boxes only have wmasks as needed | |||
| 2016-09-06 | Pulled out duplicate memory annotations | Angie | |
| * Annotate reference * Changed memory port names to RWx, Wx, Rx, etc. and reconnected nodes | |||
| 2016-09-06 | Corrected counting for VectorTypes in MemUtils | Angie | |
| * Was originally adding one extra set of things (to -> until) * MemPortUtil conditionally includes wmask, if necessary Changed endian-ness of write data/mask to match convention (little endian) | |||
| 2016-09-06 | Pulled out memory annotation (write mask) | Angie | |
| * Annotates all sequential memories that can be black boxed. * Annotates sequential memories that don't need write masks. | |||
| 2016-09-06 | Generated *.conf file for unique srams | Angie | |
| 2016-09-06 | Minor utility changes. | Angie | |
| * Corrected names to match current RW port spec * Added Jack's Namespace on Circuit | |||
| 2016-09-06 | Added starter code for SMem replacement | Angie | |
| 2016-09-05 | Change null statement to empty begin end (#264) | Colin Schmidt | |
| this eliminates warnings in recent versions of VCS | |||
| 2016-08-25 | emit wires instead of registers for invalid randomization | Howard Mao | |
| Before, the verilog emitter would connect registers to the invalid ports and use random initialization on the generated registers. It is better to generate wires instead and use random assignment on the wires. | |||
| 2016-08-25 | Finer grained control over randomization | Howard Mao | |
| We previously had `ifdef guards on some parts of the emitted verilog to control whether some registers or nets should be given random initial values. These guards were all dependent on the RANDOMIZE macro. However, there were actually three separate cases being controlled 1. Giving random values to disconnected wires 2. Random initialization of registers 3. Random initialization of memories It is possible that the designer would want to switch these three on or off independently in simulation. For instance, the latter two are usually safe because registers and memories will get some definite binary value at power on in the actual circuit, but the first one can be quite dangerous because the undriven wire could be metastable. This change provides separate macros for each of the three sets of guards so that they can be controlled independently. | |||
