1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
5 <title>LLVM Test Suite Guide</title>
6 <link rel="stylesheet" href="llvm.css" type="text/css">
10 <div class="doc_title">
15 <li><a href="#overview">Overview</a></li>
16 <li><a href="#Requirements">Requirements</a></li>
17 <li><a href="#quick">Quick Start</a></li>
18 <li><a href="#org">LLVM Test Suite Organization</a>
20 <li><a href="#codefragments">Code Fragments</a></li>
21 <li><a href="#wholeprograms">Whole Programs</a></li>
24 <li><a href="#tree">LLVM Test Suite Tree</a></li>
25 <li><a href="#dgstructure">DejaGNU Structure</a></li>
26 <li><a href="#progstructure"><tt>llvm-test</tt> Structure</a></li>
27 <li><a href="#run">Running the LLVM Tests</a>
29 <li><a href="#customtest">Writing custom tests for llvm-test</a></li>
32 <li><a href="#nightly">Running the nightly tester</a></li>
35 <div class="doc_author">
36 <p>Written by John T. Criswell, <a
37 href="http://llvm.x10sys.com/rspencer">Reid Spencer</a>, and Tanya Lattner</p>
40 <!--=========================================================================-->
41 <div class="doc_section"><a name="overview">Overview</a></div>
42 <!--=========================================================================-->
44 <div class="doc_text">
46 <p>This document is the reference manual for the LLVM test suite. It documents
47 the structure of the LLVM test suite, the tools needed to use it, and how to add
52 <!--=========================================================================-->
53 <div class="doc_section"><a name="Requirements">Requirements</a></div>
54 <!--=========================================================================-->
56 <div class="doc_text">
58 <p>In order to use the LLVM test suite, you will need all of the software
59 required to build LLVM, plus the following:</p>
62 <dt><a href="http://www.gnu.org/software/dejagnu/">DejaGNU</a></dt>
63 <dd>The Feature and Regressions tests are organized and run by DejaGNU.</dd>
64 <dt><a href="http://expect.nist.gov/">Expect</a></dt>
65 <dd>Expect is required by DejaGNU.</dd>
66 <dt><a href="http://www.tcl.tk/software/tcltk/">tcl</a></dt>
67 <dd>Tcl is required by DejaGNU. </dd>
69 <dt><a href="http://www.netlib.org/f2c">F2C</a></dt>
70 <dd>For now, LLVM does not have a Fortran front-end, but using F2C, we can run
71 Fortran benchmarks. F2C support must be enabled via <tt>configure</tt> if not
72 installed in a standard place. F2C requires three items: the <tt>f2c</tt>
73 executable, <tt>f2c.h</tt> to compile the generated code, and <tt>libf2c.a</tt>
74 to link generated code. By default, given an F2C directory <tt>$DIR</tt>, the
75 configure script will search <tt>$DIR/bin</tt> for <tt>f2c</tt>,
76 <tt>$DIR/include</tt> for <tt>f2c.h</tt>, and <tt>$DIR/lib</tt> for
77 <tt>libf2c.a</tt>. The default <tt>$DIR</tt> values are: <tt>/usr</tt>,
78 <tt>/usr/local</tt>, <tt>/sw</tt>, and <tt>/opt</tt>. If you installed F2C in a
79 different location, you must tell <tt>configure</tt>:
82 <li><tt>./configure --with-f2c=$DIR</tt><br>
83 This will specify a new <tt>$DIR</tt> for the above-described search
84 process. This will only work if the binary, header, and library are in their
85 respective subdirectories of <tt>$DIR</tt>.</li>
87 <li><tt>./configure --with-f2c-bin=/binary/path --with-f2c-inc=/include/path
88 --with-f2c-lib=/lib/path</tt><br>
89 This allows you to specify the F2C components separately. Note: if you choose
90 this route, you MUST specify all three components, and you need to only specify
91 <em>directories</em> where the files are located; do NOT include the
92 filenames themselves on the <tt>configure</tt> line.</li>
96 <p>Darwin (Mac OS X) developers can simplify the installation of Expect and tcl
97 by using fink. <tt>fink install expect</tt> will install both. Alternatively,
98 Darwinports users can use <tt>sudo port install expect</tt> to install Expect
103 <!--=========================================================================-->
104 <div class="doc_section"><a name="quick">Quick Start</a></div>
105 <!--=========================================================================-->
107 <div class="doc_text">
109 <p>The tests are located in two separate CVS modules. The basic feature and
110 regression tests are in the main "llvm" module under the directory
111 <tt>llvm/test</tt>. A more comprehensive test suite that includes whole
112 programs in C and C++ is in the <tt>llvm-test</tt> module. This module should
113 be checked out to the <tt>llvm/projects</tt> directory. When you
114 <tt>configure</tt> the <tt>llvm</tt> module, the <tt>llvm-test</tt> module
115 will be automatically configured. Alternatively, you can configure the
116 <tt>llvm-test</tt> module manually.</p>
117 <p>To run all of the simple tests in LLVM using DejaGNU, use the master Makefile
118 in the <tt>llvm/test</tt> directory:</p>
127 <p>To run only a subdirectory of tests in llvm/test using DejaGNU (ie.
128 Regression/Transforms), just set the TESTSUITE variable to the path of the
129 subdirectory (relative to <tt>llvm/test</tt>):</p>
131 % gmake -C llvm/test TESTSUITE=Regression/Transforms
134 <p><b>Note: If you are running the tests with <tt>objdir != subdir</tt>, you
135 must have run the complete testsuite before you can specify a
136 subdirectory.</b></p>
138 <p>To run the comprehensive test suite (tests that compile and execute whole
139 programs), run the <tt>llvm-test</tt> tests:</p>
145 % ./configure --with-llvmsrc=$LLVM_SRC_ROOT --with-llvmobj=$LLVM_OBJ_ROOT
151 <!--=========================================================================-->
152 <div class="doc_section"><a name="org">LLVM Test Suite Organization</a></div>
153 <!--=========================================================================-->
155 <div class="doc_text">
157 <p>The LLVM test suite contains two major categories of tests: code
158 fragments and whole programs. Code fragments are in the <tt>llvm</tt> module
159 under the <tt>llvm/test</tt> directory. The whole programs
160 test suite is in the <tt>llvm-test</tt> module under the main directory.</p>
164 <!-- _______________________________________________________________________ -->
165 <div class="doc_subsection"><a name="codefragments">Code Fragments</a></div>
166 <!-- _______________________________________________________________________ -->
168 <div class="doc_text">
170 <p>Code fragments are small pieces of code that test a specific feature of LLVM
171 or trigger a specific bug in LLVM. They are usually written in LLVM assembly
172 language, but can be written in other languages if the test targets a particular
173 language front end.</p>
175 <p>Code fragments are not complete programs, and they are never executed to
176 determine correct behavior.</p>
178 <p>These code fragment tests are located in the <tt>llvm/test/Features</tt> and
179 <tt>llvm/test/Regression</tt> directories.</p>
183 <!-- _______________________________________________________________________ -->
184 <div class="doc_subsection"><a name="wholeprograms">Whole Programs</a></div>
185 <!-- _______________________________________________________________________ -->
187 <div class="doc_text">
189 <p>Whole Programs are pieces of code which can be compiled and linked into a
190 stand-alone program that can be executed. These programs are generally written
191 in high level languages such as C or C++, but sometimes they are written
192 straight in LLVM assembly.</p>
194 <p>These programs are compiled and then executed using several different
195 methods (native compiler, LLVM C backend, LLVM JIT, LLVM native code generation,
196 etc). The output of these programs is compared to ensure that LLVM is compiling
197 the program correctly.</p>
199 <p>In addition to compiling and executing programs, whole program tests serve as
200 a way of benchmarking LLVM performance, both in terms of the efficiency of the
201 programs generated as well as the speed with which LLVM compiles, optimizes, and
204 <p>All "whole program" tests are located in the <tt>llvm-test</tt> CVS
209 <!--=========================================================================-->
210 <div class="doc_section"><a name="tree">LLVM Test Suite Tree</a></div>
211 <!--=========================================================================-->
213 <div class="doc_text">
215 <p>Each type of test in the LLVM test suite has its own directory. The major
216 subtrees of the test suite directory tree are as follows:</p>
219 <li><tt>llvm/test</tt>
220 <p>This directory contains a large array of small tests
221 that exercise various features of LLVM and to ensure that regressions do not
222 occur. The directory is broken into several sub-directories, each focused on
223 a particular area of LLVM. A few of the important ones are:<ul>
224 <li><tt>Analysis</tt>: checks Analysis passes.</li>
225 <li><tt>Archive</tt>: checks the Archive library.</li>
226 <li><tt>Assembler</tt>: checks Assembly reader/writer functionality.</li>
227 <li><tt>Bytecode</tt>: checks Bytecode reader/writer functionality.</li>
228 <li><tt>CodeGen</tt>: checks code generation and each target.</li>
229 <li><tt>Features</tt>: checks various features of the LLVM language.</li>
230 <li><tt>Linker</tt>: tests bytecode linking.</li>
231 <li><tt>Transforms</tt>: tests each of the scalar, IPO, and utility
232 transforms to ensure they make the right transformations.</li>
233 <li><tt>Verifier</tt>: tests the IR verifier.</li>
235 <p>Typically when a bug is found in LLVM, a regression test containing
236 just enough code to reproduce the problem should be written and placed
237 somewhere underneath this directory. In most cases, this will be a small
238 piece of LLVM assembly language code, often distilled from an actual
239 application or benchmark.</p></li>
241 <li><tt>llvm-test</tt>
242 <p>The <tt>llvm-test</tt> CVS module contains programs that can be compiled
243 with LLVM and executed. These programs are compiled using the native compiler
244 and various LLVM backends. The output from the program compiled with the
245 native compiler is assumed correct; the results from the other programs are
246 compared to the native program output and pass if they match.</p>
248 <p>In addition for testing correctness, the <tt>llvm-test</tt> directory also
249 performs timing tests of various LLVM optimizations. It also records
250 compilation times for the compilers and the JIT. This information can be
251 used to compare the effectiveness of LLVM's optimizations and code
254 <li><tt>llvm-test/SingleSource</tt>
255 <p>The SingleSource directory contains test programs that are only a single
256 source file in size. These are usually small benchmark programs or small
257 programs that calculate a particular value. Several such programs are grouped
258 together in each directory.</p></li>
260 <li><tt>llvm-test/MultiSource</tt>
261 <p>The MultiSource directory contains subdirectories which contain entire
262 programs with multiple source files. Large benchmarks and whole applications
265 <li><tt>llvm-test/External</tt>
266 <p>The External directory contains Makefiles for building code that is external
267 to (i.e., not distributed with) LLVM. The most prominent members of this
268 directory are the SPEC 95 and SPEC 2000 benchmark suites. The presence and
269 location of these external programs is configured by the llvm-test
270 <tt>configure</tt> script.</p></li>
275 <!--=========================================================================-->
276 <div class="doc_section"><a name="dgstructure">DejaGNU Structure</a></div>
277 <!--=========================================================================-->
278 <div class="doc_text">
279 <p>The LLVM test suite is partially driven by DejaGNU and partially driven by
280 GNU Make. Specifically, the Features and Regression tests are all driven by
281 DejaGNU. The <tt>llvm-test</tt> module is currently driven by a set of
284 <p>The DejaGNU structure is very simple, but does require some information to
285 be set. This information is gathered via <tt>configure</tt> and is written
286 to a file, <tt>site.exp</tt> in <tt>llvm/test</tt>. The <tt>llvm/test</tt>
287 Makefile does this work for you.</p>
289 <p>In order for DejaGNU to work, each directory of tests must have a
290 <tt>dg.exp</tt> file. DejaGNU looks for this file to determine how to run the
291 tests. This file is just a Tcl script and it can do anything you want, but
292 we've standardized it for the LLVM regression tests. It simply loads a Tcl
293 library (<tt>test/lib/llvm.exp</tt>) and calls the <tt>llvm_runtests</tt>
294 function defined in that library with a list of file names to run. The names
295 are obtained by using Tcl's glob command. Any directory that contains only
296 directories does not need the <tt>dg.exp</tt> file.</p>
298 <p>The <tt>llvm-runtests</tt> function lookas at each file that is passed to
299 it and gathers any lines together that match "RUN:". This are the "RUN" lines
300 that specify how the test is to be run. So, each test script must contain
301 RUN lines if it is to do anything. If there are no RUN lines, the
302 <tt>llvm-runtests</tt> function will issue an error and the test will
305 <p>RUN lines are specified in the comments of the test program using the
306 keyword <tt>RUN</tt> followed by a colon, and lastly the command (pipeline)
307 to execute. Together, these lines form the "script" that
308 <tt>llvm-runtests</tt> executes to run the test case. The syntax of the
309 RUN lines is similar to a shell's syntax for pipelines including I/O
310 redirection and variable substitution. However, even though these lines
311 may <i>look</i> like a shell script, they are not. RUN lines are interpreted
312 directly by the Tcl <tt>exec</tt> command. They are never executed by a
313 shell. Consequently the syntax differs from normal shell script syntax in a
314 few ways. You can specify as many RUN lines as needed.</p>
316 <p>Each RUN line is executed on its own, distinct from other lines unless
317 its last character is <tt>\</tt>. This continuation character causes the RUN
318 line to be concatenated with the next one. In this way you can build up long
319 pipelines of commands without making huge line lengths. The lines ending in
320 <tt>\</tt> are concatenated until a RUN line that doesn't end in <tt>\</tt> is
321 found. This concatenated set or RUN lines then constitutes one execution.
322 Tcl will substitute variables and arrange for the pipeline to be executed. If
323 any process in the pipeline fails, the entire line (and test case) fails too.
326 <p> Below is an example of legal RUN lines in a <tt>.ll</tt> file:</p>
328 ; RUN: llvm-as < %s | llvm-dis > %t1
329 ; RUN: llvm-dis < %s.bc-13 > %t2
334 <!-- _______________________________________________________________________ -->
335 <div class="doc_subsection"><a name="dgvars">Vars And Substitutions</a></div>
336 <div class="doc_text">
337 <p>With a RUN line there are a number of substitutions that are permitted. In
338 general, any Tcl variable that is available in the <tt>substitute</tt>
339 function (in <tt>test/lib/llvm.exp</tt>) can be substituted into a RUN line.
340 To make a substitution just write the variable's name preceded by a $.
341 Additionally, for compatibility reasons with previous versions of the test
342 library, certain names can be accessed with an alternate syntax: a % prefix.
343 These alternates are deprecated and may go away in a future version.
345 Here are the available variable names. The alternate syntax is listed in
347 <dl style="margin-left: 25px">
348 <dt><b>$test</b> (%s)</dt>
349 <dd>The full path to the test case's source. This is suitable for passing
350 on the command line as the input to an llvm tool.</dd>
351 <dt><b>$srcdir</b></dt>
352 <dd>The source directory from where the "<tt>make check</tt>" was run.</dd>
353 <dt><b>objdir</b></dt>
354 <dd>The object directory that corresponds to the </tt>$srcdir</tt>.</dd>
355 <dt><b>subdir</b></dt>
356 <dd>A partial path from the <tt>test</tt> directory that contains the
357 sub-directory that contains the test source being executed.</dd>
358 <dt><b>srcroot</b></dt>
359 <dd>The root directory of the LLVM src tree.</dd>
360 <dt><b>objroot</b></dt>
361 <dd>The root directory of the LLVM object tree. This could be the same
364 <dd>The path to the directory that contains the test case source. This is
365 for locating any supporting files that are not generated by the test, but
366 used by the test.</dd>
368 <dd>The path to a temporary file name that could be used for this test case.
369 The file name won't conflict with other test cases. You can append to it if
370 you need multiple temporaries. This is useful as the destination of some
371 redirected output.</dd>
372 <dt><b>llvmlibsdir</b> (%llvmlibsdir)</dt>
373 <dd>The directory where the LLVM libraries are located.</dd>
374 <dt><b>target_triplet</b> (%target_triplet)</dt>
375 <dd>The target triplet that corresponds to the current host machine (the one
376 running the test cases). This should probably be called "host".<dd>
377 <dt><b>prcontext</b> (%prcontext)</dt>
378 <dd>Path to the prcontext tcl script that prints some context around a
379 line that matches a pattern. This isn't strictly necessary as the test suite
380 is run with its PATH altered to include the test/Scripts directory where
381 the prcontext script is located. Note that this script is similar to
382 <tt>grep -C</tt> but you should use the <tt>prcontext</tt> script because
383 not all platforms support <tt>grep -C</tt>.</dd>
384 <dt><b>llvmgcc</b> (%llvmgcc)</dt>
385 <dd>The full path to the <tt>llvm-gcc</tt> executable as specified in the
386 configured LLVM environment</dd>
387 <dt><b>llvmgxx</b> (%llvmgxx)</dt>
388 <dd>The full path to the <tt>llvm-gxx</tt> executable as specified in the
389 configured LLVM environment</dd>
390 <dt><b>llvmgcc_version</b> (%llvmgcc_version)</dt>
391 <dd>The full version number of the <tt>llvm-gcc</tt> executable.</dd>
392 <dt><b>llvmgccmajvers</b> (%llvmgccmajvers)</dt>
393 <dd>The major version number of the <tt>llvm-gcc</tt> executable.</dd>
394 <dt><b>gccpath</b></dt>
395 <dd>The full path to the C compiler used to <i>build </i> LLVM. Note that
396 this might not be gcc.</dd>
397 <dt><b>gxxpath</b></dt>
398 <dd>The full path to the C++ compiler used to <i>build </i> LLVM. Note that
399 this might not be g++.</dd>
400 <dt><b>compile_c</b> (%compile_c)</dt>
401 <dd>The full command line used to compile LLVM C source code. This has all
402 the configured -I, -D and optimization options.</dd>
403 <dt><b>compile_cxx</b> (%compile_cxx)</dt>
404 <dd>The full command used to compile LLVM C++ source code. This has
405 all the configured -I, -D and optimization options.</dd>
406 <dt><b>link</b> (%link)</dt>
407 <dd>This full link command used to link LLVM executables. This has all the
408 configured -I, -L and -l options.</dd>
409 <dt><b>shlibext</b> (%shlibext)</dt>
410 <dd>The suffix for the host platforms share library (dll) files. This
411 includes the period as the first character.</dd>
413 <p>To add more variables, two things need to be changed. First, add a line in
414 the <tt>test/Makefile</tt> that creates the <tt>site.exp</tt> file. This will
415 "set" the variable as a global in the site.exp file. Second, in the
416 <tt>test/lib/llvm.exp</tt> file, in the substitute proc, add the variable name
417 to the list of "global" declarations at the beginning of the proc. That's it,
418 the variable can then be used in test scripts.</p>
421 <!-- _______________________________________________________________________ -->
422 <div class="doc_subsection"><a name="dgfeatures">Other Features</a></div>
423 <div class="doc_text">
424 <p>To make RUN line writing easier, there are several shell scripts located
425 in the <tt>llvm/test/Scripts</tt> directory. For example:</p>
427 <dt><b>ignore</b></dt>
428 <dd>This script runs its arguments and then always returns 0. This is useful
429 in cases where the test needs to cause a tool to generate an error (e.g. to
430 check the error output). However, any program in a pipeline that returns a
431 non-zero result will cause the test to fail. This script overcomes that
432 issue and nicely documents that the test case is purposefully ignoring the
433 result code of the tool</dd>
435 <dd>This script runs its arguments and then inverts the result code from
436 it. Zero result codes become 1. Non-zero result codes become 0. This is
437 useful to invert the result of a grep. For example "not grep X" means
438 succeed only if you don't find X in the input.</dd>
441 <p>Sometimes it is necessary to mark a test case as "expected fail" or XFAIL.
442 You can easily mark a test as XFAIL just by including <tt>XFAIL: </tt> on a
443 line near the top of the file. This signals that the test case should succeed
444 if the test fails. Such test cases are counted separately by DejaGnu. To
445 specify an expected fail, use the XFAIL keyword in the comments of the test
446 program followed by a colon and one or more regular expressions (separated by
447 a comma). The regular expressions allow you to XFAIL the test conditionally
448 by host platform. The regular expressions following the : are matched against
449 the target triplet or llvmgcc version number for the host machine. If there is
450 a match, the test is expected to fail. If not, the test is expected to
451 succeed. To XFAIL everywhere just specify <tt>XFAIL: *</tt>. When matching
452 the llvm-gcc version, you can specify the major (e.g. 3) or full version
453 (i.e. 3.4) number. Here is an example of an <tt>XFAIL</tt> line:</p>
455 ; XFAIL: darwin,sun,llvmgcc4
458 <p>To make the output more useful, the <tt>llvm_runtest</tt> function wil
459 scan the lines of the test case for ones that contain a pattern that matches
460 PR[0-9]+. This is the syntax for specifying a PR (Problem Report) number that
461 is related to the test case. The numer after "PR" specifies the LLVM bugzilla
462 number. When a PR number is specified, it will be used in the pass/fail
463 reporting. This is useful to quickly get some context when a test fails.</p>
465 <p>Finally, any line that contains "END." will cause the special
466 interpretation of lines to terminate. This is generally done right after the
467 last RUN: line. This has two side effects: (a) it prevents special
468 interpretation of lines that are part of the test program, not the
469 instructions to the test case, and (b) it speeds things up for really big test
470 cases by avoiding interpretation of the remainder of the file.</p>
474 <!--=========================================================================-->
475 <div class="doc_section"><a name="progstructure"><tt>llvm-test</tt>
477 <!--=========================================================================-->
479 <div class="doc_text">
481 <p>As mentioned previously, the <tt>llvm-test</tt> module provides three types
482 of tests: MultiSource, SingleSource, and External. Each tree is then subdivided
483 into several categories, including applications, benchmarks, regression tests,
484 code that is strange grammatically, etc. These organizations should be
485 relatively self explanatory.</p>
487 <p>In addition to the regular "whole program" tests, the <tt>llvm-test</tt>
488 module also provides a mechanism for compiling the programs in different ways.
489 If the variable TEST is defined on the gmake command line, the test system will
490 include a Makefile named <tt>TEST.<value of TEST variable>.Makefile</tt>.
491 This Makefile can modify build rules to yield different results.</p>
493 <p>For example, the LLVM nightly tester uses <tt>TEST.nightly.Makefile</tt> to
494 create the nightly test reports. To run the nightly tests, run <tt>gmake
495 TEST=nightly</tt>.</p>
497 <p>There are several TEST Makefiles available in the tree. Some of them are
498 designed for internal LLVM research and will not work outside of the LLVM
499 research group. They may still be valuable, however, as a guide to writing your
500 own TEST Makefile for any optimization or analysis passes that you develop with
503 <p>Note, when configuring the <tt>llvm-test</tt> module, you might want to
504 specify the following configuration options:</p>
506 <dt><i>--enable-spec2000</i>
507 <dt><i>--enable-spec2000=<<tt>directory</tt>></i>
509 Enable the use of SPEC2000 when testing LLVM. This is disabled by default
510 (unless <tt>configure</tt> finds SPEC2000 installed). By specifying
511 <tt>directory</tt>, you can tell configure where to find the SPEC2000
512 benchmarks. If <tt>directory</tt> is left unspecified, <tt>configure</tt>
513 uses the default value
514 <tt>/home/vadve/shared/benchmarks/speccpu2000/benchspec</tt>.
516 <dt><i>--enable-spec95</i>
517 <dt><i>--enable-spec95=<<tt>directory</tt>></i>
519 Enable the use of SPEC95 when testing LLVM. It is similar to the
520 <i>--enable-spec2000</i> option.
522 <dt><i>--enable-povray</i>
523 <dt><i>--enable-povray=<<tt>directory</tt>></i>
525 Enable the use of Povray as an external test. Versions of Povray written
526 in C should work. This option is similar to the <i>--enable-spec2000</i>
531 <!--=========================================================================-->
532 <div class="doc_section"><a name="run">Running the LLVM Tests</a></div>
533 <!--=========================================================================-->
535 <div class="doc_text">
537 <p>First, all tests are executed within the LLVM object directory tree. They
538 <i>are not</i> executed inside of the LLVM source tree. This is because the
539 test suite creates temporary files during execution.</p>
541 <p>The master Makefile in llvm/test is capable of running only the DejaGNU
542 driven tests. By default, it will run all of these tests.</p>
544 <p>To run only the DejaGNU driven tests, run <tt>gmake</tt> at the
545 command line in <tt>llvm/test</tt>. To run a specific directory of tests, use
546 the TESTSUITE variable.
549 <p>For example, to run the Regression tests, type
550 <tt>gmake TESTSUITE=Regression</tt> in <tt>llvm/tests</tt>.</p>
552 <p>Note that there are no Makefiles in <tt>llvm/test/Features</tt> and
553 <tt>llvm/test/Regression</tt>. You must use DejaGNU from the <tt>llvm/test</tt>
554 directory to run them.</p>
556 <p>To run the <tt>llvm-test</tt> suite, you need to use the following steps:
559 <li>cd into the llvm/projects directory</li>
560 <li>check out the <tt>llvm-test</tt> module with:<br/>
561 <tt>cvs -d :pserver:anon@llvm.org:/var/cvs/llvm co -PR llvm-test</tt><br>
562 This will get the test suite into <tt>llvm/projects/llvm-test</tt></li>
563 <li>configure the test suite. You can do this one of two ways:
565 <li>Use the regular llvm configure:<br/>
566 <tt>cd $LLVM_OBJ_ROOT ; $LLVM_SRC_ROOT/configure</tt><br/>
567 This will ensure that the <tt>projects/llvm-test</tt> directory is also
568 properly configured.</li>
569 <li>Use the <tt>configure</tt> script found in the <tt>llvm-test</tt> source
571 <tt>$LLVM_SRC_ROOT/projects/llvm-test/configure
572 --with-llvmsrc=$LLVM_SRC_ROOT --with-llvmobj=$LLVM_OBJ_ROOT</tt>
577 <p>Note that the second and third steps only need to be done once. After you
578 have the suite checked out and configured, you don't need to do it again (unless
579 the test code or configure script changes).</p>
581 <p>To make a specialized test (use one of the
582 <tt>llvm-test/TEST.<type>.Makefile</tt>s), just run:<br/>
583 <tt>gmake TEST=<type> test</tt><br/>For example, you could run the
584 nightly tester tests using the following commands:</p>
587 % cd llvm/projects/llvm-test
588 % gmake TEST=nightly test
591 <p>Regardless of which test you're running, the results are printed on standard
592 output and standard error. You can redirect these results to a file if you
595 <p>Some tests are known to fail. Some are bugs that we have not fixed yet;
596 others are features that we haven't added yet (or may never add). In DejaGNU,
597 the result for such tests will be XFAIL (eXpected FAILure). In this way, you
598 can tell the difference between an expected and unexpected failure.</p>
600 <p>The tests in <tt>llvm-test</tt> have no such feature at this time. If the
601 test passes, only warnings and other miscellaneous output will be generated. If
602 a test fails, a large <program> FAILED message will be displayed. This
603 will help you separate benign warnings from actual test failures.</p>
607 <!-- _______________________________________________________________________ -->
608 <div class="doc_subsection">
609 <a name="customtest">Writing custom tests for llvm-test</a></div>
610 <!-- _______________________________________________________________________ -->
612 <div class="doc_text">
614 <p>Assuming you can run llvm-test, (e.g. "<tt>gmake TEST=nightly report</tt>"
615 should work), it is really easy to run optimizations or code generator
616 components against every program in the tree, collecting statistics or running
617 custom checks for correctness. At base, this is how the nightly tester works,
618 it's just one example of a general framework.</p>
620 <p>Lets say that you have an LLVM optimization pass, and you want to see how
621 many times it triggers. First thing you should do is add an LLVM
622 <a href="ProgrammersManual.html#Statistic">statistic</a> to your pass, which
623 will tally counts of things you care about.</p>
625 <p>Following this, you can set up a test and a report that collects these and
626 formats them for easy viewing. This consists of two files, an
627 "<tt>llvm-test/TEST.XXX.Makefile</tt>" fragment (where XXX is the name of your
628 test) and an "<tt>llvm-test/TEST.XXX.report</tt>" file that indicates how to
629 format the output into a table. There are many example reports of various
630 levels of sophistication included with llvm-test, and the framework is very
633 <p>If you are interested in testing an optimization pass, check out the
634 "libcalls" test as an example. It can be run like this:<p>
636 <div class="doc_code">
638 % cd llvm/projects/llvm-test/MultiSource/Benchmarks # or some other level
639 % make TEST=libcalls report
643 <p>This will do a bunch of stuff, then eventually print a table like this:</p>
645 <div class="doc_code">
647 Name | total | #exit |
649 FreeBench/analyzer/analyzer | 51 | 6 |
650 FreeBench/fourinarow/fourinarow | 1 | 1 |
651 FreeBench/neural/neural | 19 | 9 |
652 FreeBench/pifft/pifft | 5 | 3 |
653 MallocBench/cfrac/cfrac | 1 | * |
654 MallocBench/espresso/espresso | 52 | 12 |
655 MallocBench/gs/gs | 4 | * |
656 Prolangs-C/TimberWolfMC/timberwolfmc | 302 | * |
657 Prolangs-C/agrep/agrep | 33 | 12 |
658 Prolangs-C/allroots/allroots | * | * |
659 Prolangs-C/assembler/assembler | 47 | * |
660 Prolangs-C/bison/mybison | 74 | * |
665 <p>This basically is grepping the -stats output and displaying it in a table.
666 You can also use the "TEST=libcalls report.html" target to get the table in HTML
667 form, similarly for report.csv and report.tex.</p>
669 <p>The source for this is in llvm-test/TEST.libcalls.*. The format is pretty
670 simple: the Makefile indicates how to run the test (in this case,
671 "<tt>opt -simplify-libcalls -stats</tt>"), and the report contains one line for
672 each column of the output. The first value is the header for the column and the
673 second is the regex to grep the output of the command for. There are lots of
674 example reports that can do fancy stuff.</p>
679 <!--=========================================================================-->
680 <div class="doc_section"><a name="nightly">Running the nightly tester</a></div>
681 <!--=========================================================================-->
683 <div class="doc_text">
686 The <a href="http://llvm.org/nightlytest/">LLVM Nightly Testers</a>
687 automatically check out an LLVM tree, build it, run the "nightly"
688 program test (described above), run all of the feature and regression tests,
689 delete the checked out tree, and then submit the results to
690 <a href="http://llvm.org/nightlytest/">http://llvm.org/nightlytest/</a>.
691 After test results are submitted to
692 <a href="http://llvm.org/nightlytest/">http://llvm.org/nightlytest/</a>,
693 they are processed and displayed on the tests page. An email to
694 <a href="http://lists.cs.uiuc.edu/pipermail/llvm-testresults/">
695 llvm-testresults@cs.uiuc.edu</a> summarizing the results is also generated.
696 This testing scheme is designed to ensure that programs don't break as well
697 as keep track of LLVM's progress over time.</p>
699 <p>If you'd like to set up an instance of the nightly tester to run on your
700 machine, take a look at the comments at the top of the
701 <tt>utils/NewNightlyTest.pl</tt> file. If you decide to set up a nightly tester
702 please choose a unique nickname and invoke <tt>utils/NewNightlyTest.pl</tt>
703 with the "-nickname [yournickname]" command line option. We usually run it
704 from a crontab entry that looks like this:</p>
706 <div class="doc_code">
708 5 3 * * * $HOME/llvm/utils/NewNightlyTest.pl -parallel -nickname Nickname \
709 $CVSROOT $HOME/buildtest $HOME/cvs/testresults
713 <p>Or, you can create a shell script to encapsulate the running of the script.
714 The optimized x86 Linux nightly test is run from just such a script:</p>
716 <div class="doc_code">
719 BASE=/proj/work/llvm/nightlytest
720 export CVSROOT=:pserver:anon@llvm.org:/var/cvs/llvm
721 export BUILDDIR=$BASE/build
722 export WEBDIR=$BASE/testresults
723 export LLVMGCCDIR=/proj/work/llvm/cfrontend/install
724 export PATH=/proj/install/bin:$LLVMGCCDIR/bin:$PATH
725 export LD_LIBRARY_PATH=/proj/install/lib
727 cp /proj/work/llvm/llvm/utils/NewNightlyTest.pl .
728 nice ./NewNightlyTest.pl -nice -release -verbose -parallel -enable-linscan \
729 -nickname NightlyTester -noexternals 2>&1 > output.log
733 <p>It is also possible to specify the the location your nightly test results
734 are submitted. You can do this by passing the command line option
735 "-submit-server [server_address]" and "-submit-script [script_on_server]" to
736 <tt>utils/NewNightlyTest.pl</tt>. For example, to submit to the llvm.org
737 nightly test results page, you would invoke the nightly test script with
738 "-submit-server llvm.org -submit-script /nightlytest/NightlyTestAccept.cgi".
739 If these options are not specified, the nightly test script sends the results
740 to the llvm.org nightly test results page.</p>
742 <p>Take a look at the <tt>NewNightlyTest.pl</tt> file to see what all of the
743 flags and strings do. If you start running the nightly tests, please let us
748 <!-- *********************************************************************** -->
752 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
753 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
754 <a href="http://validator.w3.org/check/referer"><img
755 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
757 John T. Criswell, Reid Spencer, and Tanya Lattner<br>
758 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br/>
759 Last modified: $Date$