Written by John T. Criswell
This document is the reference manual for the LLVM test suite. It documents the structure of the LLVM test suite, the tools needed to use it, and how to add and run tests.
In order to use the LLVM test suite, you will need all of the software required to build LLVM, plus the following:
The tests are located in the LLVM source tree under the directory llvm/test. To run all of the tests in LLVM, use the Master Makefile in that directory:
% gmake -C llvm/test
To run only the code fragment tests (i.e. those that do basic testing of LLVM), run the tests organized by QMTest:
% gmake -C llvm/test qmtest
To run only the tests that compile and execute whole programs, run the Programs tests:
% gmake -C llvm/test/Programs
The LLVM test suite contains two major categories of tests: code fragments and whole programs.
Code fragments are small pieces of code that test a specific feature of LLVM or trigger a specific bug in LLVM. They are usually written in LLVM assembly language, but can be written in other languages if the test targets a particular language front end.
Code fragments are not complete programs, and they are never executed to determine correct behavior.
The tests in the Features and Regression directories contain code fragments.
Whole Programs are pieces of code which can be compiled and linked into a stand-alone program that can be executed. These programs are generally written in high level languages such as C or C++, but sometimes they are written straight in LLVM assembly.
These programs are compiled and then executed using several different methods (native compiler, LLVM C backend, LLVM JIT, LLVM native code generation, etc). The output of these programs is compared to ensure that LLVM is compiling the program correctly.
In addition to compiling and executing programs, whole program tests serve as a way of benchmarking LLVM performance, both in terms of the efficiency of the programs generated as well as the speed with which LLVM compiles, optimizes, and generates code.
The Programs directory contains all tests which compile and benchmark whole programs.
Each type of test in the LLVM test suite has its own directory. The major subtrees of the test suite directory tree are as follows:
This directory contains sample codes that test various features of the LLVM language. These pieces of sample code are run through various assembler, disassembler, and optimizer passes.
This directory contains regression tests for LLVM. When a bug is found in LLVM, a regression test containing just enough code to reproduce the problem should be written and placed somewhere underneath this directory. In most cases, this will be a small piece of LLVM assembly language code, often distilled from an actual application or benchmark.
The Programs directory contains programs that can be compiled with LLVM and executed. These programs are compiled using the native compiler and various LLVM backends. The output from the program compiled with the native compiler is assumed correct; the results from the other programs are compared to the native program output and pass if they match.
In addition for testing correctness, the Programs directory also performs timing tests of various LLVM optimizations. It also records compilation times for the compilers and the JIT. This information can be used to compare the effectiveness of LLVM's optimizations and code generation.
The Programs directory is subdivided into several smaller subdirectories:
The SingleSource directory contains test programs that are only a single source file in size. These are usually small benchmark programs or small programs that calculate a particular value. Several such programs are grouped together in each directory.
The MultiSource directory contains subdirectories which contain entire programs with multiple source files. Large benchmarks and whole applications go here.
The External directory contains Makefiles for building code that is external to (i.e. not distributed with) LLVM. The most prominent member of this directory is the SPEC 2000 benchmark suite. The presence and location of these external programs is configured by the LLVM configure script.
This directory contains the QMTest information files. Inside this directory are QMTest administration files and the Python code that implements the LLVM test and database classes.
The LLVM test suite is partially driven by QMTest and partially driven by GNU Make. Specifically, the Features and Regression tests are all driven by QMTest. The Programs directory is currently driven by a set of Makefiles.
The QMTest system needs to have several pieces of information available; these pieces of configuration information are known collectively as the "context" in QMTest parlance. Since the context for LLVM is relatively large, the master Makefile in llvm/test sets it for you.
The LLVM database class makes the subdirectories of llvm/test a QMTest test database. For each directory that contains tests driven by QMTest, it knows what type of test the source file is and how to run it.
Hence, the QMTest namespace is essentially what you see in the Feature and Regression directories, but there is some magic that the database class performs (as described below).
The QMTest namespace is currently composed of the following tests and test suites:
These are the feature tests found in the Feature directory. They are broken up into the following categories:
Assembler/Disassembler tests. These tests verify that a piece of LLVM assembly language can be assembled into bytecode and then disassembled into the original assembly language code. It does this several times to ensure that assembled output can be disassembled and disassembler output can be assembled. It also verifies that the give assembly language file can be assembled correctly.
Optimizer tests. These tests verify that two of the optimizer passes completely optimize a program (i.e. after a single pass, they cannot optimize a program any further).
Machine code tests. These tests verify that the LLVM assembly language file can be translated into native assembly code.
C code tests. These tests verify that the specified LLVM assembly code can be converted into C source code using the C backend.
The LLVM database class looks at every file in the Feature directory and creates a fake test hierarchy containing Feature.<testtype>.<testname>. So, if you add an LLVM assembly language file to the Feature directory, it actually creates 5 new tests: assembler/disassembler, assembler, optimizer, machine code, and C code.
These are the regression tests. There is one suite for each subdirectory of the Regression directory. If you add a new subdirectory there, you will need to modify, at least, the RegressionMap variable in QMTest/llvmdb.py so that QMTest knows how to run the tests in the new subdirectory.
As mentioned previously, the Programs tree in llvm/test provides three types of tests: MultiSource, SingleSource, and External. Each tree is then subdivided into several categories, including applications, benchmarks, regression tests, code that is strange grammatically, etc. These organizations should be relatively self explanatory.
In addition to the regular Programs tests, the Programs tree also provides a mechanism for compiling the programs in different ways. If the variable TEST is defined on the gmake command line, the test system will include a Makefile named TEST.<value of TEST variable>.Makefile. This Makefile can modify build rules to yield different results.
For example, the LLVM nightly tester uses TEST.nightly.Makefile to create the nightly test reports. To run the nightly tests, run gmake TEST=nightly.
There are several TEST Makefiles available in the tree. Some of them are designed for internal LLVM research and will not work outside of the LLVM research group. They may still be valuable, however, as a guide to writing your own TEST Makefile for any optimization or analysis passes that you develop with LLVM.
First, all tests are executed within the LLVM object directory tree. They are not executed inside of the LLVM source tree. This is because the test suite creates temporary files during execution.
The master Makefile in llvm/test is capable of running both the QMTest driven tests and the Programs tests. By default, it will run all of the tests.
To run only the QMTest driven tests, run gmake qmtest at the command line in llvm/tests. To run a specific qmtest, suffix the test name with ".t" when running gmake.
For example, to run the Regression.LLC tests, type gmake Regression.LLC.t in llvm/tests.
Note that the Makefiles in llvm/test/Features and llvm/test/Regression are gone. You must now use QMTest from the llvm/test directory to run them.
To run the Programs test, cd into the llvm/test/Programs directory and type gmake. Alternatively, you can type gmake TEST=<type> test to run one of the specialized tests in llvm/test/Programs/TEST.<type>.Makefile. For example, you could run the nightly tester tests using the following commands:
% cd llvm/test/Programs % gmake TEST=nightly test
Regardless of which test you're running, the results are printed on standard output and standard error. You can redirect these results to a file if you choose.
Some tests are known to fail. Some are bugs that we have not fixed yet; others are features that we haven't added yet (or may never add). In QMTest, the result for such tests will be XFAIL (eXpected FAILure). In this way, you can tell the difference between an expected and unexpected failure.
The Programs tests have no such feature as of this time. If the test passes, only warnings and other miscellaneous output will be generated. If a test fails, a large <program> FAILED message will be displayed. This will help you separate benign warnings from actual test failures.