[llvm-dev] Intended behavior of CGSCC pass manager.

Hal Finkel via llvm-dev llvm-dev at lists.llvm.org
Wed Jun 8 09:32:39 PDT 2016

----- Original Message -----

> From: "Sean Silva via llvm-dev" <llvm-dev at lists.llvm.org>
> To: "llvm-dev" <llvm-dev at lists.llvm.org>
> Sent: Wednesday, June 8, 2016 6:19:03 AM
> Subject: [llvm-dev] Intended behavior of CGSCC pass manager.

> Hi Chandler, Philip, Mehdi, (and llvm-dev,)

> (this is partially a summary of some discussions that happened at the
> last LLVM bay area social, and partially a discussion about the
> direction of the CGSCC pass manager)

> A the last LLVM social we discussed the progress on the CGSCC pass
> manager. It seems like Chandler has a CGSCC pass manager working,
> but it is still unresolved exactly which semantics we want (more
> about this below) that are reasonably implementable.

> AFAICT, there has been no public discussion about what exact
> semantics we ultimately want to have. We should figure that out.

> The main difficulty which Chandler described is the apparently quite
> complex logic surrounding needing to run function passes nested
> within an SCC pass manager, while providing some guarantees about
> exactly what order the function passes are run. The existing CGSCC
> pass manager just punts on some of the problems that arise (look in
> CGPassManager::runOnModule, CGPassManager::RunAllPassesOnSCC, and
> CGPassManager::RunPassOnSCC in
> llvm/lib/Analysis/CallGraphSCCPass.cpp), and these are the problems
> that Chandler has been trying to solve.

> (
> Why is this "function passes inside CGSCC passes" stuff interesting?
> Because LLVM can do inlining on an SCC (often just a single
> function) and then run function passes to simplify the function(s)
> in the SCC before it tries to inline into a parent SCC. (the SCC
> visitation order is post-order)
> For example, we may inline a bunch of code, but after inlining we can
> tremendously simplify the function, and we want to do so before
> considering this function for inlining into its callers so that we
> get an accurate evaluation of the inline cost.
> Based on what Chandler said, it seems that LLVM is fairly unique in
> this regard and other compilers don't do this (which is why we can't
> just look at how other compilers solve this problem; they don't have
> this problem (maybe they should? or maybe we shouldn't?)). For
> example, he described that GCC uses different inlining "phases";
> e.g. it does early inlining on the entire module, then does
> simplifications on the entire module, then does late inlining on the
> entire module; so it is not able to incrementally simplify as it
> inlines like LLVM does.
This incremental simplification is an important feature of our inliner, and one we should endeavor to keep. We might also want different phases at some point (e.g. a top-down and a bottom-up phase), but that's another story. 

> )

> As background for what is below, the LazyCallGraph tracks two graphs:
> the "call graph" and the "ref graph".
> Conceptually, the call graph is the graph of direct calls, where
> indirect calls and calls to external functions do not appear (or are
> connected to dummy nodes). The ref graph is basically the graph of
> all functions transitively accessible based on the
> globals/constants/etc. referenced by a function (e.g. if a function
> `foo` references a vtable that is defined in the module, there is an
> edge in the ref graph from `foo` to every function in the vtable).
> The call graph is a strict subset of the ref graph.

> Chandler described that he had a major breakthrough in that the CGSCC
> pass manager only had to deal with 3 classes of modifications that
> can occur:
> - a pass may e.g. propagate a load of a function pointer into an
> indirect call, turning it into an direct call. This requires adding
> an edge in the CG but not in the ref graph.
> - a pass may take a direct call and turn it into an indirect call.
> This requires removing an edge from the CG, but not in the ref
> graph.
> - a pass may delete a direct call. This removes an edge in the CG and
> also in the ref graph.

> From the perspective of the CGSCC pass manager, these operations can
> affect the SCC structure. Adding an edge might merge SCC's and
> deleting an edge might split SCC's. Chandler mentioned that
> apparently the issues of splitting and merging SCC's within the
> current infrastructure are actually quite challenging and lead to
> e.g. iterator invalidation issues, and that is what he is working
> on.

> (
> The ref graph is important to guide the overall SCC visitation order
> because it basically represents "the largest graph that the CG may
> turn into due to our static analysis of this module". I.e. no
> transformation we can statically make in the CGSCC passes can ever
> cause us to need to merge SCC's in the ref graph.
> )

> I have a couple overall questions/concerns:

> 1. The ref graph can easily go quadratic. E.g.

> typedef void (*fp)();
> fp funcs[] = {
> &foo1,
> &foo2,
> ...
> &fooN
> }
> void foo1() { funcs[something](); }

> void foo2() { funcs[something](); }
> ...
> void fooN() { funcs[something](); }

> One real-world case where this might come about is in the presence of
> vtables.

> The existing CGSCC pass manager does not have this issue AFAIK
> because it does not consider the ref graph.

> Does anybody have any info/experience about how densely connected the
> ref graph can get in programs that might reasonably be fed to the
> compiler?
> I just did a quick sanity check with LLD/ELF using
> `--lto-newpm-passes=cgscc(no-op-cgscc)` and it at least seemed to
> terminate / not run out of memory. Based on some rough calculations
> looking at the profile, it seem like the entire run of the inliner
> in the old LTO pipeline (which is about 5% of total LTO time on this
> particular example I looked at) is only 2-3x as expensive as just
> `--lto-newpm-passes=cgscc(no-op-cgscc)`, so the LazyCallGraph
> construction is certainly not free.

> 2. What is the intended behavior of CGSCC passes when SCC's are split
> or merged? E.g. a CGSCC pass runs on an SCC (e.g. the inliner). Now
> we run some function passes nested inside the CGSCC pass manager
> (e.g. to simplify things after inlining). Consider:
This is not how I thought the current scheme worked ;) -- I was under the impression that we had a call graph with conservatively-connected dummy nodes for external/indirect functions. As a result, there is no semantics-preserving optimization that will merge SCCs, only split them. In that case, I'd expect that once an SCC is split, we re-run the CGSCC passes over the newly-separated SCCs. But this corresponds to running over the "ref graph", as you describe it. I don't understand why we want the non-conservative graph. 

> a) These function passes are e.g. now able to devirtualize a call,
> adding an edge to the CG, forming a larger CG SCC. Do you re-run the
> CGSCC pass (say, the inliner) on this larger SCC?

> b) These function passes are e.g. able to DCE a call, removing an
> edge from the CG. This converts, say, a CG SCC which is a cycle
> graph (like a->b->c->a) into a path graph (a->b->c, with no edge
> back to a). The inliner had already visited a, b, and c as a single
> SCC. Now does it have to re-visit c, then b, then a, as single-node
> SCC's?

> btw:

> One way that I have found it useful to think about this is in terms
> of the visitation during Tarjan's SCC algorithm. I'll reference the
> pseudocode in
> https://en.wikipedia.org/wiki/Tarjan%27s_strongly_connected_components_algorithm
> . Inside the "strongconnect" routine when we have identified an SCC
> (the true branch of `if (v.lowlink = v.index)` test ) we can visit
> stack[v.index:stack.size()] as an SCC. This may or may not
> invalidate some things on the stack (the variable `S` in the
> pseudocode) and we may need to fix it up (e.g. inlining deleted a
> function, so we can't have an entry on the stack). Then, we can run
> function passes as we pop individual functions off the stack, but it
> is easier to think about IMO than merging of SCC data structures: if
> we add edges to the CG then we have to do more DFS on the new edges
> and if we delete edges then the DFS order of the stack gives us
> certain guarantees.
> Personally I find this much easier to reason about than the
> description in terms of splitting and merging SCC's in the CG and
> ref graph (which the LazyCallGraph API makes one to think about
> since it hides the underlying Tarjan's algorithm).
> The LazyCallGraph API makes the current loop in
> http://reviews.llvm.org/diffusion/L/browse/llvm/trunk/include/llvm/Analysis/CGSCCPassManager.h;272124$100
> very clean, but at least for my thinking about the problem, it seems
> like the wrong abstraction (and most of the LazyCallGraph API seems
> to be unused, so it seems like it may be overly heavyweight).
> E.g. I think that maybe the easiest thing to do is to turn the
> current approach inside out: instead of having the pass manager
> logic be the "normal code" and forcing the Tarjan algorithm to
> become a state machine of iterators, use an open-coded Tarjan
> algorithm with some callbacks and make the pass management logic be
> the state machine.
> This will also open the door to avoiding the potentially quadratic
> size of the ref graph, since e.g. in the example I gave above, we
> can mark the `funcs` array itself as already having been visited
> during the walk. In the current LazyCallGraph, this would require
> adding some sort of notion of hyperedge.
FWIW, I see no purpose in abstracting one algorithm, especially if that makes things algorithmically harder. Also, the LazyCallGraph abstraction and the iterator abstraction seem like separate issues. Iterator abstractions are often useful because you can use them in generic algorithms, etc. 

> Since this is such a high priority (due to blocking PGO inlining), I
> will probably try my hand at implementing the CGSCC pass manager
> sometime soon unless somebody beats me to it. (I'll probably try the
> "open-coded SCC visit" approach).

> Another possibility is implementing the new CGSCC pass manager that
> uses the same visitation semantics as the one in the old PM, and
> then we can refactor that as needed. In fact, that may be the best
> approach so that porting to the new PM is as NFC as possible and we
> can isolate the functional (i.e., need benchmarks, measurements ...)
> changes in separate commits.
I'm in favor of this approach for exactly the reason you mention. Being able to bisect regressions to the algorithmic change, separate from the infrastructure change, will likely make things easier in the long run (and will avoid the problem, to the extent possible, of performance regressions blocking the pass-manager work). 

> Sorry for the wall of text.
No problem; much appreciated. 


> -- Sean Silva
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev


Hal Finkel 
Assistant Computational Scientist 
Leadership Computing Facility 
Argonne National Laboratory 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20160608/e1eb0bc5/attachment.html>

More information about the llvm-dev mailing list