search for: smallset

...nk we have too many different set classes. They have incompatible APIs and surprising behavior. I think we can reduce their number substantially. Dear all, The following is a subset of the set classes we have in ADT: * DenseSet * SmallDenseSet (https://reviews.llvm.org/D25628) * SetVector * SmallSetVector * SmallSet * SmallPtrSet * StringSet * FoldingSet * ContextualFoldingSet * FoldingSetVector Every one of these classes has quirks. Here are some particularly bad ones: * SmallSet falls back to std::set<T> when it runs out of inline storage (which itself is pretty bad because...

...9;m filtering, although I'm concerned about the overhead of copying the set. > > What I'm wondering is what's the best practice for these kinds of operations? I'm not sure if LLVM has a preferred idiom, although I'm sure copying elements is fine in most cases. Ideally, SmallSet iterators would work just like std::set iterators so the containers would be interchangeable. But I don't see an easy way to implement those semantics. Unless I'm missing something, it seems fairly easy to implement it = container.erase(it) in SmallSet with some care to check for underlying...

I'm using the various LLVM set containers (SmallPtrSet and the like), and something I need to do fairly often is iterate over a set, removing all elements that don't pass a test. Now, doing this on lists is fairly straightforward using the "it = container.erase(it)" idiom so as to keep the iterator valid during the delete. However, the set classes don't have a means to both

...good to me. I have a couple of minor comments: > > +void State::AddInsnClass(unsigned InsnClass, > Add a top level comment describing the function > > + std::map<State*, std::set<Transition*, ltTransition>, ltState> > stateTransitions; > You should be able to use SmallSet here. Also, this line exceeds 80 > columns. I tried but SmallSet is not iterable. SmallSetPtr could be useful here but it doesn't allow custom sorting. > > > On a related note, is the CachedTable mechanism in DFAPacketizer.h > useful for your architecture? Currently the DFA...

...> >>> +void State::AddInsnClass(unsigned InsnClass, >>> Add a top level comment describing the function >>> >>> + std::map<State*, std::set<Transition*, ltTransition>, ltState> >>> stateTransitions; >>> You should be able to use SmallSet here. Also, this line exceeds 80 >>> columns. >> >> I tried but SmallSet is not iterable. SmallSetPtr could be useful >> here but it doesn't allow custom sorting. >> >>> >>> >>> On a related note, is the CachedTable mechanism in DFAP...

UniqueVector is an stl map coupled with an stl vector[1], which seems very undesirable for you. As far as SmallVector + slow search, I believe that is precisely what the implementation of SmallSet does until it has to grow, after which it uses an stl set[2]. If you're always staying within your small number, SmallSetVector would be wasteful as the SmallSet would itself also have a SmallVector. If you need to access in insertion order then go with SmallVector + slow search. If you need s...

...alias sets from AliasSetTracker::findAliasSetForUnknownInst. I think that Philip proposed something similar to your approach in https://llvm.org/bugs/show_bug.cgi?id=23077. 2016-01-24 22:44 GMT+05:00 Daniel Berlin <dberlin at dberlin.org>: > It sounds like UnknownInsts should really be a SmallSet, SmallPtrSet, or > DenseSet (if you can't do the others). > > Otherwise, even with your patch, we are still wasting time traversing and > copying and .... the unknown instructions. > > If that doesn't work, I suspect the way you get dupes is through mergeSetIn, > so you...

...of minor comments: >> >> +void State::AddInsnClass(unsigned InsnClass, >> Add a top level comment describing the function >> >> + std::map<State*, std::set<Transition*, ltTransition>, ltState> >> stateTransitions; >> You should be able to use SmallSet here. Also, this line exceeds 80 >> columns. > > I tried but SmallSet is not iterable. SmallSetPtr could be useful here > but it doesn't allow custom sorting. > >> >> >> On a related note, is the CachedTable mechanism in DFAPacketizer.h >> useful for...

...void State::AddInsnClass(unsigned InsnClass, >>>> Add a top level comment describing the function >>>> >>>> + std::map<State*, std::set<Transition*, ltTransition>, ltState> >>>> stateTransitions; >>>> You should be able to use SmallSet here. Also, this line exceeds 80 >>>> columns. >>> >>> I tried but SmallSet is not iterable. SmallSetPtr could be useful >>> here but it doesn't allow custom sorting. >>> >>>> >>>> >>>> On a related note, is the...

Hi Ivan, The patch looks good to me. I have a couple of minor comments: +void State::AddInsnClass(unsigned InsnClass, Add a top level comment describing the function + std::map<State*, std::set<Transition*, ltTransition>, ltState> stateTransitions; You should be able to use SmallSet here. Also, this line exceeds 80 columns. On a related note, is the CachedTable mechanism in DFAPacketizer.h useful for your architecture? Currently the DFA generator generates one table for a given architecture. I had originally added the CachedTable mechanism since for a given compilation an...

Dear llvm contributors, Could you please advise me how to skip checks, which are performed in AliasSet::aliasesUnknownInst, of unknown instructions from different alias sets of an alias set tracker that is a parameter of ‘AliasSetTracker::add(const AliasSetTracker &AST)’? If this wasn’t available at the moment and someone could review me, I would try to implement it. A temporary patch can be

...lass(unsigned InsnClass, >>>>> Add a top level comment describing the function >>>>> >>>>> + std::map<State*, std::set<Transition*, ltTransition>, ltState> >>>>> stateTransitions; >>>>> You should be able to use SmallSet here. Also, this line exceeds 80 >>>>> columns. >>>> I tried but SmallSet is not iterable. SmallSetPtr could be useful >>>> here but it doesn't allow custom sorting. >>>> >>>>> >>>>> On a related note, is the Cache...

What are the rules regarding iterator stability for SmallPtrSet? I'm running into a problem where erasing an element seems to invalidate iterators that are not pointing at the erased element. Specifically, the set is in small mode, I advance an iterator to the last element in the set (not end(), but one before) and then erase the second-to-last element. I understand that the last element

...but one before) and then erase the > second-to-last element. I understand that the last element will replace Right, deleting an element will invalidate any iterators past it, just like a vector. I don't know of a good solution to this that doesn't break the efficiency properties of smallset. Do you? -Chris -- http://nondot.org/sabre/ http://llvm.org/

...hen erase the > > second-to-last element. I understand that the last element will replace > > Right, deleting an element will invalidate any iterators past it, just > like a vector. I don't know of a good solution to this that doesn't break > the efficiency properties of smallset. Do you? What I've done is code up a version of erase() that returns an iterator to the next element. There still seem to be some problems that I'm looking into. This interface has the side benefit of being standard-library compatible and can be used with standard algorithms. If I get...

...ng up redundant information. Transitions doesn't contains 'From' state anymore because they are folded into it. - Using STLExtras functions. - Removed old and unused API's (e.g. addTransition in DFA) The new patch is attached but if you prefer (2) I can remake it. I failed to use SmallSet to store the transitions in State because I needed to iterate on it. What kind of llvm structure may I use? Ivan > > -Anshu > -------------- next part -------------- A non-text attachment was scrubbed... Name: DFAPacketizerEmitter.new.patch Type: text/x-patch Size: 8628 bytes Desc: no...

...ass, >>>>>> Add a top level comment describing the function >>>>>> >>>>>> + std::map<State*, std::set<Transition*, ltTransition>, ltState> >>>>>> stateTransitions; >>>>>> You should be able to use SmallSet here. Also, this line exceeds 80 >>>>>> columns. >>>>> I tried but SmallSet is not iterable. SmallSetPtr could be useful >>>>> here but it doesn't allow custom sorting. >>>>> >>>>>> >>>>>> On a re...

...ntEqClasses.h" #include "llvm/ADT/MapVector.h" #include "llvm/ADT/Optional.h" #include "llvm/ADT/PointerUnion.h" #include "llvm/ADT/STLExtras.h" #include "llvm/ADT/SetVector.h" #include "llvm/ADT/SmallPtrSet.h" #include "llvm/ADT/SmallSet.h" #include "llvm/ADT/SmallVector.h" #include "llvm/ADT/SparseBitVector.h" #include "llvm/ADT/Statistic.h" #include "llvm/ADT/StringExtras.h" #include "llvm/ADT/StringMap.h" #include "llvm/ADT/StringRef.h" #include "llvm/ADT/Stri...

...;>>> Add a top level comment describing the function >>>>>>> >>>>>>> + std::map<State*, std::set<Transition*, ltTransition>, ltState> >>>>>>> stateTransitions; >>>>>>> You should be able to use SmallSet here. Also, this line exceeds 80 >>>>>>> columns. >>>>>> I tried but SmallSet is not iterable. SmallSetPtr could be useful >>>>>> here but it doesn't allow custom sorting. >>>>>> >>>>>>> >>>...

...; ours is compiled to a Debugging version. llvm-gcc, though, is angry. I get the following error (with path prefixes shortened for convenience) when compiling llvm-gcc: > In file included from [...]/llvm/include/llvm/ADT/SmallPtrSet.h:20, > from [...]/llvm/include/llvm/ADT/SmallSet.h:18, > from [...]/llvm/include/llvm/ADT/SetVector.h:23, > from ../../llvm-gcc-src/gcc/llvm-internal.h:38, > from ../../llvm-gcc-src/gcc/llvm-backend.cpp:23: > [...]/llvm/include/llvm/Support/DataTypes.h:38:3: error: #error "Must #...