blob: 839875266a2412da2a734508719a9779d5c84b14 [file] [log] [blame]
Lang Hamese038aae2016-06-06 18:35:44 +00001=====================================================================
2Building a JIT: Adding Optimizations -- An introduction to ORC Layers
3=====================================================================
Lang Hamesbe84d2be2016-05-26 00:38:04 +00004
5.. contents::
6 :local:
7
8**This tutorial is under active development. It is incomplete and details may
9change frequently.** Nonetheless we invite you to try it out as it stands, and
10we welcome any feedback.
11
12Chapter 2 Introduction
13======================
14
Lang Hamesc499d2a2016-06-06 03:28:12 +000015Welcome to Chapter 2 of the "Building an ORC-based JIT in LLVM" tutorial. In
16`Chapter 1 <BuildingAJIT1.html>`_ of this series we examined a basic JIT
17class, KaleidoscopeJIT, that could take LLVM IR modules as input and produce
18executable code in memory. KaleidoscopeJIT was able to do this with relatively
19little code by composing two off-the-shelf *ORC layers*: IRCompileLayer and
20ObjectLinkingLayer, to do much of the heavy lifting.
Lang Hamesbe84d2be2016-05-26 00:38:04 +000021
Lang Hamesc499d2a2016-06-06 03:28:12 +000022In this layer we'll learn more about the ORC layer concept by using a new layer,
23IRTransformLayer, to add IR optimization support to KaleidoscopeJIT.
Lang Hamesbe84d2be2016-05-26 00:38:04 +000024
Lang Hamesc499d2a2016-06-06 03:28:12 +000025Optimizing Modules using the IRTransformLayer
26=============================================
Lang Hamesbe84d2be2016-05-26 00:38:04 +000027
Lang Hamesc499d2a2016-06-06 03:28:12 +000028In `Chapter 4 <LangImpl4.html>`_ of the "Implementing a language with LLVM"
29tutorial series the llvm *FunctionPassManager* is introduced as a means for
30optimizing LLVM IR. Interested readers may read that chapter for details, but
Lang Hames8705d112016-06-20 18:34:46 +000031in short: to optimize a Module we create an llvm::FunctionPassManager
Lang Hamesc499d2a2016-06-06 03:28:12 +000032instance, configure it with a set of optimizations, then run the PassManager on
33a Module to mutate it into a (hopefully) more optimized but semantically
34equivalent form. In the original tutorial series the FunctionPassManager was
Lang Hames11c43d52016-06-20 18:37:52 +000035created outside the KaleidoscopeJIT and modules were optimized before being
Lang Hamesc499d2a2016-06-06 03:28:12 +000036added to it. In this Chapter we will make optimization a phase of our JIT
Lang Hames11c43d52016-06-20 18:37:52 +000037instead. For now this will provide us a motivation to learn more about ORC
Lang Hamesc499d2a2016-06-06 03:28:12 +000038layers, but in the long term making optimization part of our JIT will yield an
39important benefit: When we begin lazily compiling code (i.e. deferring
40compilation of each function until the first time it's run), having
41optimization managed by our JIT will allow us to optimize lazily too, rather
42than having to do all our optimization up-front.
Lang Hamesbe84d2be2016-05-26 00:38:04 +000043
Lang Hamesc499d2a2016-06-06 03:28:12 +000044To add optimization support to our JIT we will take the KaleidoscopeJIT from
45Chapter 1 and compose an ORC *IRTransformLayer* on top. We will look at how the
46IRTransformLayer works in more detail below, but the interface is simple: the
47constructor for this layer takes a reference to the layer below (as all layers
48do) plus an *IR optimization function* that it will apply to each Module that
49is added via addModuleSet:
50
Lang Hames38eb0312016-06-06 04:53:59 +000051.. code-block:: c++
Lang Hamesc499d2a2016-06-06 03:28:12 +000052
53 class KaleidoscopeJIT {
54 private:
55 std::unique_ptr<TargetMachine> TM;
56 const DataLayout DL;
57 ObjectLinkingLayer<> ObjectLayer;
58 IRCompileLayer<decltype(ObjectLayer)> CompileLayer;
59
60 typedef std::function<std::unique_ptr<Module>(std::unique_ptr<Module>)>
61 OptimizeFunction;
62
63 IRTransformLayer<decltype(CompileLayer), OptimizeFunction> OptimizeLayer;
64
65 public:
66 typedef decltype(OptimizeLayer)::ModuleSetHandleT ModuleHandle;
67
68 KaleidoscopeJIT()
69 : TM(EngineBuilder().selectTarget()), DL(TM->createDataLayout()),
70 CompileLayer(ObjectLayer, SimpleCompiler(*TM)),
71 OptimizeLayer(CompileLayer,
72 [this](std::unique_ptr<Module> M) {
73 return optimizeModule(std::move(M));
74 }) {
75 llvm::sys::DynamicLibrary::LoadLibraryPermanently(nullptr);
76 }
77
78Our extended KaleidoscopeJIT class starts out the same as it did in Chapter 1,
79but after the CompileLayer we introduce a typedef for our optimization function.
80In this case we use a std::function (a handy wrapper for "function-like" things)
81from a single unique_ptr<Module> input to a std::unique_ptr<Module> output. With
82our optimization function typedef in place we can declare our OptimizeLayer,
83which sits on top of our CompileLayer.
84
85To initialize our OptimizeLayer we pass it a reference to the CompileLayer
86below (standard practice for layers), and we initialize the OptimizeFunction
Lang Hames706db2e2016-06-06 18:07:23 +000087using a lambda that calls out to an "optimizeModule" function that we will
88define below.
Lang Hamesc499d2a2016-06-06 03:28:12 +000089
Lang Hames38eb0312016-06-06 04:53:59 +000090.. code-block:: c++
Lang Hamesc499d2a2016-06-06 03:28:12 +000091
92 // ...
93 auto Resolver = createLambdaResolver(
94 [&](const std::string &Name) {
95 if (auto Sym = OptimizeLayer.findSymbol(Name, false))
Lang Hamesad4a9112016-08-01 20:49:11 +000096 return Sym;
97 return JITSymbol(nullptr);
Lang Hamesc499d2a2016-06-06 03:28:12 +000098 },
99 // ...
Lang Hames3242f652016-06-06 05:07:52 +0000100
101.. code-block:: c++
102
103 // ...
Lang Hamesc499d2a2016-06-06 03:28:12 +0000104 return OptimizeLayer.addModuleSet(std::move(Ms),
105 make_unique<SectionMemoryManager>(),
106 std::move(Resolver));
107 // ...
108
Lang Hames3242f652016-06-06 05:07:52 +0000109.. code-block:: c++
110
Lang Hamesc499d2a2016-06-06 03:28:12 +0000111 // ...
112 return OptimizeLayer.findSymbol(MangledNameStream.str(), true);
113 // ...
114
Lang Hames3242f652016-06-06 05:07:52 +0000115.. code-block:: c++
116
Lang Hamesc499d2a2016-06-06 03:28:12 +0000117 // ...
118 OptimizeLayer.removeModuleSet(H);
119 // ...
120
121Next we need to replace references to 'CompileLayer' with references to
122OptimizeLayer in our key methods: addModule, findSymbol, and removeModule. In
123addModule we need to be careful to replace both references: the findSymbol call
124inside our resolver, and the call through to addModuleSet.
125
Lang Hames38eb0312016-06-06 04:53:59 +0000126.. code-block:: c++
Lang Hamesc499d2a2016-06-06 03:28:12 +0000127
128 std::unique_ptr<Module> optimizeModule(std::unique_ptr<Module> M) {
129 // Create a function pass manager.
130 auto FPM = llvm::make_unique<legacy::FunctionPassManager>(M.get());
131
132 // Add some optimizations.
133 FPM->add(createInstructionCombiningPass());
134 FPM->add(createReassociatePass());
135 FPM->add(createGVNPass());
136 FPM->add(createCFGSimplificationPass());
137 FPM->doInitialization();
138
139 // Run the optimizations over all functions in the module being added to
140 // the JIT.
141 for (auto &F : *M)
142 FPM->run(F);
143
144 return M;
145 }
146
147At the bottom of our JIT we add a private method to do the actual optimization:
148*optimizeModule*. This function sets up a FunctionPassManager, adds some passes
149to it, runs it over every function in the module, and then returns the mutated
Lang Hamesd29ee532016-06-06 18:22:47 +0000150module. The specific optimizations are the same ones used in
Lang Hamesc499d2a2016-06-06 03:28:12 +0000151`Chapter 4 <LangImpl4.html>`_ of the "Implementing a language with LLVM"
Lang Hamesd29ee532016-06-06 18:22:47 +0000152tutorial series. Readers may visit that chapter for a more in-depth
153discussion of these, and of IR optimization in general.
Lang Hamesc499d2a2016-06-06 03:28:12 +0000154
Lang Hamesd29ee532016-06-06 18:22:47 +0000155And that's it in terms of changes to KaleidoscopeJIT: When a module is added via
156addModule the OptimizeLayer will call our optimizeModule function before passing
157the transformed module on to the CompileLayer below. Of course, we could have
158called optimizeModule directly in our addModule function and not gone to the
159bother of using the IRTransformLayer, but doing so gives us another opportunity
160to see how layers compose. It also provides a neat entry point to the *layer*
161concept itself, because IRTransformLayer turns out to be one of the simplest
162implementations of the layer concept that can be devised:
Lang Hamesc499d2a2016-06-06 03:28:12 +0000163
Lang Hames38eb0312016-06-06 04:53:59 +0000164.. code-block:: c++
Lang Hamesc499d2a2016-06-06 03:28:12 +0000165
166 template <typename BaseLayerT, typename TransformFtor>
167 class IRTransformLayer {
168 public:
169 typedef typename BaseLayerT::ModuleSetHandleT ModuleSetHandleT;
170
171 IRTransformLayer(BaseLayerT &BaseLayer,
172 TransformFtor Transform = TransformFtor())
173 : BaseLayer(BaseLayer), Transform(std::move(Transform)) {}
174
175 template <typename ModuleSetT, typename MemoryManagerPtrT,
176 typename SymbolResolverPtrT>
177 ModuleSetHandleT addModuleSet(ModuleSetT Ms,
178 MemoryManagerPtrT MemMgr,
179 SymbolResolverPtrT Resolver) {
180
181 for (auto I = Ms.begin(), E = Ms.end(); I != E; ++I)
182 *I = Transform(std::move(*I));
183
184 return BaseLayer.addModuleSet(std::move(Ms), std::move(MemMgr),
185 std::move(Resolver));
186 }
187
188 void removeModuleSet(ModuleSetHandleT H) { BaseLayer.removeModuleSet(H); }
189
190 JITSymbol findSymbol(const std::string &Name, bool ExportedSymbolsOnly) {
191 return BaseLayer.findSymbol(Name, ExportedSymbolsOnly);
192 }
193
194 JITSymbol findSymbolIn(ModuleSetHandleT H, const std::string &Name,
195 bool ExportedSymbolsOnly) {
196 return BaseLayer.findSymbolIn(H, Name, ExportedSymbolsOnly);
197 }
198
199 void emitAndFinalize(ModuleSetHandleT H) {
200 BaseLayer.emitAndFinalize(H);
201 }
202
203 TransformFtor& getTransform() { return Transform; }
204
205 const TransformFtor& getTransform() const { return Transform; }
206
207 private:
208 BaseLayerT &BaseLayer;
209 TransformFtor Transform;
210 };
211
212This is the whole definition of IRTransformLayer, from
213``llvm/include/llvm/ExecutionEngine/Orc/IRTransformLayer.h``, stripped of its
214comments. It is a template class with two template arguments: ``BaesLayerT`` and
Lang Hamesd29ee532016-06-06 18:22:47 +0000215``TransformFtor`` that provide the type of the base layer and the type of the
216"transform functor" (in our case a std::function) respectively. This class is
217concerned with two very simple jobs: (1) Running every IR Module that is added
218with addModuleSet through the transform functor, and (2) conforming to the ORC
219layer interface. The interface consists of one typedef and five methods:
Lang Hamesc499d2a2016-06-06 03:28:12 +0000220
Lang Hamesd29ee532016-06-06 18:22:47 +0000221+------------------+-----------------------------------------------------------+
Lang Hamesc499d2a2016-06-06 03:28:12 +0000222| Interface | Description |
223+==================+===========================================================+
224| | Provides a handle that can be used to identify a module |
225| ModuleSetHandleT | set when calling findSymbolIn, removeModuleSet, or |
226| | emitAndFinalize. |
227+------------------+-----------------------------------------------------------+
228| | Takes a given set of Modules and makes them "available |
229| | for execution. This means that symbols in those modules |
230| | should be searchable via findSymbol and findSymbolIn, and |
231| | the address of the symbols should be read/writable (for |
232| | data symbols), or executable (for function symbols) after |
233| | JITSymbol::getAddress() is called. Note: This means that |
234| addModuleSet | addModuleSet doesn't have to compile (or do any other |
235| | work) up-front. It *can*, like IRCompileLayer, act |
236| | eagerly, but it can also simply record the module and |
237| | take no further action until somebody calls |
238| | JITSymbol::getAddress(). In IRTransformLayer's case |
239| | addModuleSet eagerly applies the transform functor to |
240| | each module in the set, then passes the resulting set |
241| | of mutated modules down to the layer below. |
242+------------------+-----------------------------------------------------------+
243| | Removes a set of modules from the JIT. Code or data |
244| removeModuleSet | defined in these modules will no longer be available, and |
245| | the memory holding the JIT'd definitions will be freed. |
246+------------------+-----------------------------------------------------------+
247| | Searches for the named symbol in all modules that have |
248| | previously been added via addModuleSet (and not yet |
249| findSymbol | removed by a call to removeModuleSet). In |
250| | IRTransformLayer we just pass the query on to the layer |
251| | below. In our REPL this is our default way to search for |
252| | function definitions. |
253+------------------+-----------------------------------------------------------+
254| | Searches for the named symbol in the module set indicated |
255| | by the given ModuleSetHandleT. This is just an optimized |
256| | search, better for lookup-speed when you know exactly |
257| | a symbol definition should be found. In IRTransformLayer |
258| findSymbolIn | we just pass this query on to the layer below. In our |
259| | REPL we use this method to search for functions |
260| | representing top-level expressions, since we know exactly |
261| | where we'll find them: in the top-level expression module |
262| | we just added. |
263+------------------+-----------------------------------------------------------+
264| | Forces all of the actions required to make the code and |
265| | data in a module set (represented by a ModuleSetHandleT) |
266| | accessible. Behaves as if some symbol in the set had been |
267| | searched for and JITSymbol::getSymbolAddress called. This |
268| emitAndFinalize | is rarely needed, but can be useful when dealing with |
269| | layers that usually behave lazily if the user wants to |
270| | trigger early compilation (for example, to use idle CPU |
271| | time to eagerly compile code in the background). |
272+------------------+-----------------------------------------------------------+
273
274This interface attempts to capture the natural operations of a JIT (with some
275wrinkles like emitAndFinalize for performance), similar to the basic JIT API
276operations we identified in Chapter 1. Conforming to the layer concept allows
277classes to compose neatly by implementing their behaviors in terms of the these
278same operations, carried out on the layer below. For example, an eager layer
279(like IRTransformLayer) can implement addModuleSet by running each module in the
280set through its transform up-front and immediately passing the result to the
281layer below. A lazy layer, by contrast, could implement addModuleSet by
282squirreling away the modules doing no other up-front work, but applying the
283transform (and calling addModuleSet on the layer below) when the client calls
284findSymbol instead. The JIT'd program behavior will be the same either way, but
285these choices will have different performance characteristics: Doing work
286eagerly means the JIT takes longer up-front, but proceeds smoothly once this is
287done. Deferring work allows the JIT to get up-and-running quickly, but will
288force the JIT to pause and wait whenever some code or data is needed that hasn't
Sylvestre Ledru7d540502016-07-02 19:28:40 +0000289already been processed.
Lang Hamesc499d2a2016-06-06 03:28:12 +0000290
291Our current REPL is eager: Each function definition is optimized and compiled as
292soon as it's typed in. If we were to make the transform layer lazy (but not
293change things otherwise) we could defer optimization until the first time we
294reference a function in a top-level expression (see if you can figure out why,
295then check out the answer below [1]_). In the next chapter, however we'll
296introduce fully lazy compilation, in which function's aren't compiled until
297they're first called at run-time. At this point the trade-offs get much more
298interesting: the lazier we are, the quicker we can start executing the first
299function, but the more often we'll have to pause to compile newly encountered
300functions. If we only code-gen lazily, but optimize eagerly, we'll have a slow
301startup (which everything is optimized) but relatively short pauses as each
302function just passes through code-gen. If we both optimize and code-gen lazily
303we can start executing the first function more quickly, but we'll have longer
304pauses as each function has to be both optimized and code-gen'd when it's first
305executed. Things become even more interesting if we consider interproceedural
306optimizations like inlining, which must be performed eagerly. These are
307complex trade-offs, and there is no one-size-fits all solution to them, but by
308providing composable layers we leave the decisions to the person implementing
309the JIT, and make it easy for them to experiment with different configurations.
310
311`Next: Adding Per-function Lazy Compilation <BuildingAJIT3.html>`_
Lang Hamesbe84d2be2016-05-26 00:38:04 +0000312
313Full Code Listing
314=================
315
316Here is the complete code listing for our running example with an
317IRTransformLayer added to enable optimization. To build this example, use:
318
319.. code-block:: bash
320
321 # Compile
322 clang++ -g toy.cpp `llvm-config --cxxflags --ldflags --system-libs --libs core orc native` -O3 -o toy
323 # Run
324 ./toy
325
326Here is the code:
327
328.. literalinclude:: ../../examples/Kaleidoscope/BuildingAJIT/Chapter2/KaleidoscopeJIT.h
329 :language: c++
330
Lang Hamesc499d2a2016-06-06 03:28:12 +0000331.. [1] When we add our top-level expression to the JIT, any calls to functions
332 that we defined earlier will appear to the ObjectLinkingLayer as
333 external symbols. The ObjectLinkingLayer will call the SymbolResolver
334 that we defined in addModuleSet, which in turn calls findSymbol on the
335 OptimizeLayer, at which point even a lazy transform layer will have to
336 do its work.