├── .gitignore ├── LICENSE ├── README.md ├── intro ├── call_from.cpp ├── lock_basics.cpp └── modern_lock.cpp ├── mpi ├── mpi_hello.cpp └── mpi_hello_synchronize.cpp ├── parallel_algorithms └── gaussian_elimination │ ├── common │ └── common.h │ ├── mpi │ ├── cyclic_striped_mapping │ │ └── gaussian.cpp │ └── naive │ │ └── gaussian.cpp │ └── pthreads │ ├── cyclic_striped_mapping │ ├── gaussian.cpp │ └── utils.h │ └── naive │ ├── gaussian.cpp │ └── utils.h ├── pthreads └── posix_threads.cpp └── windows_threads ├── windows_threads.sln └── windows_threads ├── windows_threads.cpp ├── windows_threads.vcxproj └── windows_threads.vcxproj.filters /.gitignore: -------------------------------------------------------------------------------- 1 | ## Ignore Visual Studio temporary files, build results, and 2 | ## files generated by popular Visual Studio add-ons. 3 | 4 | # User-specific files 5 | *.suo 6 | *.user 7 | *.userosscache 8 | *.sln.docstates 9 | 10 | # User-specific files (MonoDevelop/Xamarin Studio) 11 | *.userprefs 12 | 13 | # Build results 14 | [Dd]ebug/ 15 | [Dd]ebugPublic/ 16 | [Rr]elease/ 17 | [Rr]eleases/ 18 | x64/ 19 | x86/ 20 | bld/ 21 | [Bb]in/ 22 | [Oo]bj/ 23 | [Ll]og/ 24 | 25 | # Visual Studio 2015 cache/options directory 26 | .vs/ 27 | # Uncomment if you have tasks that create the project's static files in wwwroot 28 | #wwwroot/ 29 | 30 | # MSTest test Results 31 | [Tt]est[Rr]esult*/ 32 | [Bb]uild[Ll]og.* 33 | 34 | # NUNIT 35 | *.VisualState.xml 36 | TestResult.xml 37 | 38 | # Build Results of an ATL Project 39 | [Dd]ebugPS/ 40 | [Rr]eleasePS/ 41 | dlldata.c 42 | 43 | # DNX 44 | project.lock.json 45 | project.fragment.lock.json 46 | artifacts/ 47 | 48 | *_i.c 49 | *_p.c 50 | *_i.h 51 | *.ilk 52 | *.meta 53 | *.obj 54 | *.pch 55 | *.pdb 56 | *.pgc 57 | *.pgd 58 | *.rsp 59 | *.sbr 60 | *.tlb 61 | *.tli 62 | *.tlh 63 | *.tmp 64 | *.tmp_proj 65 | *.log 66 | *.vspscc 67 | *.vssscc 68 | .builds 69 | *.pidb 70 | *.svclog 71 | *.scc 72 | 73 | # Chutzpah Test files 74 | _Chutzpah* 75 | 76 | # Visual C++ cache files 77 | ipch/ 78 | *.aps 79 | *.ncb 80 | *.opendb 81 | *.opensdf 82 | *.sdf 83 | *.cachefile 84 | *.VC.db 85 | *.VC.VC.opendb 86 | 87 | # Visual Studio profiler 88 | *.psess 89 | *.vsp 90 | *.vspx 91 | *.sap 92 | 93 | # TFS 2012 Local Workspace 94 | $tf/ 95 | 96 | # Guidance Automation Toolkit 97 | *.gpState 98 | 99 | # ReSharper is a .NET coding add-in 100 | _ReSharper*/ 101 | *.[Rr]e[Ss]harper 102 | *.DotSettings.user 103 | 104 | # JustCode is a .NET coding add-in 105 | .JustCode 106 | 107 | # TeamCity is a build add-in 108 | _TeamCity* 109 | 110 | # DotCover is a Code Coverage Tool 111 | *.dotCover 112 | 113 | # NCrunch 114 | _NCrunch_* 115 | .*crunch*.local.xml 116 | nCrunchTemp_* 117 | 118 | # MightyMoose 119 | *.mm.* 120 | AutoTest.Net/ 121 | 122 | # Web workbench (sass) 123 | .sass-cache/ 124 | 125 | # Installshield output folder 126 | [Ee]xpress/ 127 | 128 | # DocProject is a documentation generator add-in 129 | DocProject/buildhelp/ 130 | DocProject/Help/*.HxT 131 | DocProject/Help/*.HxC 132 | DocProject/Help/*.hhc 133 | DocProject/Help/*.hhk 134 | DocProject/Help/*.hhp 135 | DocProject/Help/Html2 136 | DocProject/Help/html 137 | 138 | # Click-Once directory 139 | publish/ 140 | 141 | # Publish Web Output 142 | *.[Pp]ublish.xml 143 | *.azurePubxml 144 | # TODO: Comment the next line if you want to checkin your web deploy settings 145 | # but database connection strings (with potential passwords) will be unencrypted 146 | #*.pubxml 147 | *.publishproj 148 | 149 | # Microsoft Azure Web App publish settings. Comment the next line if you want to 150 | # checkin your Azure Web App publish settings, but sensitive information contained 151 | # in these scripts will be unencrypted 152 | PublishScripts/ 153 | 154 | # NuGet Packages 155 | *.nupkg 156 | # The packages folder can be ignored because of Package Restore 157 | **/packages/* 158 | # except build/, which is used as an MSBuild target. 159 | !**/packages/build/ 160 | # Uncomment if necessary however generally it will be regenerated when needed 161 | #!**/packages/repositories.config 162 | # NuGet v3's project.json files produces more ignoreable files 163 | *.nuget.props 164 | *.nuget.targets 165 | 166 | # Microsoft Azure Build Output 167 | csx/ 168 | *.build.csdef 169 | 170 | # Microsoft Azure Emulator 171 | ecf/ 172 | rcf/ 173 | 174 | # Windows Store app package directories and files 175 | AppPackages/ 176 | BundleArtifacts/ 177 | Package.StoreAssociation.xml 178 | _pkginfo.txt 179 | 180 | # Visual Studio cache files 181 | # files ending in .cache can be ignored 182 | *.[Cc]ache 183 | # but keep track of directories ending in .cache 184 | !*.[Cc]ache/ 185 | 186 | # Others 187 | ClientBin/ 188 | ~$* 189 | *~ 190 | *.dbmdl 191 | *.dbproj.schemaview 192 | *.jfm 193 | *.pfx 194 | *.publishsettings 195 | node_modules/ 196 | orleans.codegen.cs 197 | 198 | # Since there are multiple workflows, uncomment next line to ignore bower_components 199 | # (https://github.com/github/gitignore/pull/1529#issuecomment-104372622) 200 | #bower_components/ 201 | 202 | # RIA/Silverlight projects 203 | Generated_Code/ 204 | 205 | # Backup & report files from converting an old project file 206 | # to a newer Visual Studio version. Backup files are not needed, 207 | # because we have git ;-) 208 | _UpgradeReport_Files/ 209 | Backup*/ 210 | UpgradeLog*.XML 211 | UpgradeLog*.htm 212 | 213 | # SQL Server files 214 | *.mdf 215 | *.ldf 216 | 217 | # Business Intelligence projects 218 | *.rdl.data 219 | *.bim.layout 220 | *.bim_*.settings 221 | 222 | # Microsoft Fakes 223 | FakesAssemblies/ 224 | 225 | # GhostDoc plugin setting file 226 | *.GhostDoc.xml 227 | 228 | # Node.js Tools for Visual Studio 229 | .ntvs_analysis.dat 230 | 231 | # Visual Studio 6 build log 232 | *.plg 233 | 234 | # Visual Studio 6 workspace options file 235 | *.opt 236 | 237 | # Visual Studio LightSwitch build output 238 | **/*.HTMLClient/GeneratedArtifacts 239 | **/*.DesktopClient/GeneratedArtifacts 240 | **/*.DesktopClient/ModelManifest.xml 241 | **/*.Server/GeneratedArtifacts 242 | **/*.Server/ModelManifest.xml 243 | _Pvt_Extensions 244 | 245 | # Paket dependency manager 246 | .paket/paket.exe 247 | paket-files/ 248 | 249 | # FAKE - F# Make 250 | .fake/ 251 | 252 | # JetBrains Rider 253 | .idea/ 254 | *.sln.iml 255 | 256 | # CodeRush 257 | .cr/ 258 | 259 | # Python Tools for Visual Studio (PTVS) 260 | __pycache__/ 261 | *.pyc -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | GNU GENERAL PUBLIC LICENSE 2 | Version 3, 29 June 2007 3 | 4 | Copyright (C) 2007 Free Software Foundation, Inc. 5 | Everyone is permitted to copy and distribute verbatim copies 6 | of this license document, but changing it is not allowed. 7 | 8 | Preamble 9 | 10 | The GNU General Public License is a free, copyleft license for 11 | software and other kinds of works. 12 | 13 | The licenses for most software and other practical works are designed 14 | to take away your freedom to share and change the works. By contrast, 15 | the GNU General Public License is intended to guarantee your freedom to 16 | share and change all versions of a program--to make sure it remains free 17 | software for all its users. We, the Free Software Foundation, use the 18 | GNU General Public License for most of our software; it applies also to 19 | any other work released this way by its authors. You can apply it to 20 | your programs, too. 21 | 22 | When we speak of free software, we are referring to freedom, not 23 | price. Our General Public Licenses are designed to make sure that you 24 | have the freedom to distribute copies of free software (and charge for 25 | them if you wish), that you receive source code or can get it if you 26 | want it, that you can change the software or use pieces of it in new 27 | free programs, and that you know you can do these things. 28 | 29 | To protect your rights, we need to prevent others from denying you 30 | these rights or asking you to surrender the rights. Therefore, you have 31 | certain responsibilities if you distribute copies of the software, or if 32 | you modify it: responsibilities to respect the freedom of others. 33 | 34 | For example, if you distribute copies of such a program, whether 35 | gratis or for a fee, you must pass on to the recipients the same 36 | freedoms that you received. You must make sure that they, too, receive 37 | or can get the source code. And you must show them these terms so they 38 | know their rights. 39 | 40 | Developers that use the GNU GPL protect your rights with two steps: 41 | (1) assert copyright on the software, and (2) offer you this License 42 | giving you legal permission to copy, distribute and/or modify it. 43 | 44 | For the developers' and authors' protection, the GPL clearly explains 45 | that there is no warranty for this free software. For both users' and 46 | authors' sake, the GPL requires that modified versions be marked as 47 | changed, so that their problems will not be attributed erroneously to 48 | authors of previous versions. 49 | 50 | Some devices are designed to deny users access to install or run 51 | modified versions of the software inside them, although the manufacturer 52 | can do so. This is fundamentally incompatible with the aim of 53 | protecting users' freedom to change the software. The systematic 54 | pattern of such abuse occurs in the area of products for individuals to 55 | use, which is precisely where it is most unacceptable. Therefore, we 56 | have designed this version of the GPL to prohibit the practice for those 57 | products. If such problems arise substantially in other domains, we 58 | stand ready to extend this provision to those domains in future versions 59 | of the GPL, as needed to protect the freedom of users. 60 | 61 | Finally, every program is threatened constantly by software patents. 62 | States should not allow patents to restrict development and use of 63 | software on general-purpose computers, but in those that do, we wish to 64 | avoid the special danger that patents applied to a free program could 65 | make it effectively proprietary. To prevent this, the GPL assures that 66 | patents cannot be used to render the program non-free. 67 | 68 | The precise terms and conditions for copying, distribution and 69 | modification follow. 70 | 71 | TERMS AND CONDITIONS 72 | 73 | 0. Definitions. 74 | 75 | "This License" refers to version 3 of the GNU General Public License. 76 | 77 | "Copyright" also means copyright-like laws that apply to other kinds of 78 | works, such as semiconductor masks. 79 | 80 | "The Program" refers to any copyrightable work licensed under this 81 | License. Each licensee is addressed as "you". "Licensees" and 82 | "recipients" may be individuals or organizations. 83 | 84 | To "modify" a work means to copy from or adapt all or part of the work 85 | in a fashion requiring copyright permission, other than the making of an 86 | exact copy. The resulting work is called a "modified version" of the 87 | earlier work or a work "based on" the earlier work. 88 | 89 | A "covered work" means either the unmodified Program or a work based 90 | on the Program. 91 | 92 | To "propagate" a work means to do anything with it that, without 93 | permission, would make you directly or secondarily liable for 94 | infringement under applicable copyright law, except executing it on a 95 | computer or modifying a private copy. Propagation includes copying, 96 | distribution (with or without modification), making available to the 97 | public, and in some countries other activities as well. 98 | 99 | To "convey" a work means any kind of propagation that enables other 100 | parties to make or receive copies. Mere interaction with a user through 101 | a computer network, with no transfer of a copy, is not conveying. 102 | 103 | An interactive user interface displays "Appropriate Legal Notices" 104 | to the extent that it includes a convenient and prominently visible 105 | feature that (1) displays an appropriate copyright notice, and (2) 106 | tells the user that there is no warranty for the work (except to the 107 | extent that warranties are provided), that licensees may convey the 108 | work under this License, and how to view a copy of this License. If 109 | the interface presents a list of user commands or options, such as a 110 | menu, a prominent item in the list meets this criterion. 111 | 112 | 1. Source Code. 113 | 114 | The "source code" for a work means the preferred form of the work 115 | for making modifications to it. "Object code" means any non-source 116 | form of a work. 117 | 118 | A "Standard Interface" means an interface that either is an official 119 | standard defined by a recognized standards body, or, in the case of 120 | interfaces specified for a particular programming language, one that 121 | is widely used among developers working in that language. 122 | 123 | The "System Libraries" of an executable work include anything, other 124 | than the work as a whole, that (a) is included in the normal form of 125 | packaging a Major Component, but which is not part of that Major 126 | Component, and (b) serves only to enable use of the work with that 127 | Major Component, or to implement a Standard Interface for which an 128 | implementation is available to the public in source code form. A 129 | "Major Component", in this context, means a major essential component 130 | (kernel, window system, and so on) of the specific operating system 131 | (if any) on which the executable work runs, or a compiler used to 132 | produce the work, or an object code interpreter used to run it. 133 | 134 | The "Corresponding Source" for a work in object code form means all 135 | the source code needed to generate, install, and (for an executable 136 | work) run the object code and to modify the work, including scripts to 137 | control those activities. However, it does not include the work's 138 | System Libraries, or general-purpose tools or generally available free 139 | programs which are used unmodified in performing those activities but 140 | which are not part of the work. For example, Corresponding Source 141 | includes interface definition files associated with source files for 142 | the work, and the source code for shared libraries and dynamically 143 | linked subprograms that the work is specifically designed to require, 144 | such as by intimate data communication or control flow between those 145 | subprograms and other parts of the work. 146 | 147 | The Corresponding Source need not include anything that users 148 | can regenerate automatically from other parts of the Corresponding 149 | Source. 150 | 151 | The Corresponding Source for a work in source code form is that 152 | same work. 153 | 154 | 2. Basic Permissions. 155 | 156 | All rights granted under this License are granted for the term of 157 | copyright on the Program, and are irrevocable provided the stated 158 | conditions are met. This License explicitly affirms your unlimited 159 | permission to run the unmodified Program. The output from running a 160 | covered work is covered by this License only if the output, given its 161 | content, constitutes a covered work. This License acknowledges your 162 | rights of fair use or other equivalent, as provided by copyright law. 163 | 164 | You may make, run and propagate covered works that you do not 165 | convey, without conditions so long as your license otherwise remains 166 | in force. You may convey covered works to others for the sole purpose 167 | of having them make modifications exclusively for you, or provide you 168 | with facilities for running those works, provided that you comply with 169 | the terms of this License in conveying all material for which you do 170 | not control copyright. Those thus making or running the covered works 171 | for you must do so exclusively on your behalf, under your direction 172 | and control, on terms that prohibit them from making any copies of 173 | your copyrighted material outside their relationship with you. 174 | 175 | Conveying under any other circumstances is permitted solely under 176 | the conditions stated below. Sublicensing is not allowed; section 10 177 | makes it unnecessary. 178 | 179 | 3. Protecting Users' Legal Rights From Anti-Circumvention Law. 180 | 181 | No covered work shall be deemed part of an effective technological 182 | measure under any applicable law fulfilling obligations under article 183 | 11 of the WIPO copyright treaty adopted on 20 December 1996, or 184 | similar laws prohibiting or restricting circumvention of such 185 | measures. 186 | 187 | When you convey a covered work, you waive any legal power to forbid 188 | circumvention of technological measures to the extent such circumvention 189 | is effected by exercising rights under this License with respect to 190 | the covered work, and you disclaim any intention to limit operation or 191 | modification of the work as a means of enforcing, against the work's 192 | users, your or third parties' legal rights to forbid circumvention of 193 | technological measures. 194 | 195 | 4. Conveying Verbatim Copies. 196 | 197 | You may convey verbatim copies of the Program's source code as you 198 | receive it, in any medium, provided that you conspicuously and 199 | appropriately publish on each copy an appropriate copyright notice; 200 | keep intact all notices stating that this License and any 201 | non-permissive terms added in accord with section 7 apply to the code; 202 | keep intact all notices of the absence of any warranty; and give all 203 | recipients a copy of this License along with the Program. 204 | 205 | You may charge any price or no price for each copy that you convey, 206 | and you may offer support or warranty protection for a fee. 207 | 208 | 5. Conveying Modified Source Versions. 209 | 210 | You may convey a work based on the Program, or the modifications to 211 | produce it from the Program, in the form of source code under the 212 | terms of section 4, provided that you also meet all of these conditions: 213 | 214 | a) The work must carry prominent notices stating that you modified 215 | it, and giving a relevant date. 216 | 217 | b) The work must carry prominent notices stating that it is 218 | released under this License and any conditions added under section 219 | 7. This requirement modifies the requirement in section 4 to 220 | "keep intact all notices". 221 | 222 | c) You must license the entire work, as a whole, under this 223 | License to anyone who comes into possession of a copy. This 224 | License will therefore apply, along with any applicable section 7 225 | additional terms, to the whole of the work, and all its parts, 226 | regardless of how they are packaged. This License gives no 227 | permission to license the work in any other way, but it does not 228 | invalidate such permission if you have separately received it. 229 | 230 | d) If the work has interactive user interfaces, each must display 231 | Appropriate Legal Notices; however, if the Program has interactive 232 | interfaces that do not display Appropriate Legal Notices, your 233 | work need not make them do so. 234 | 235 | A compilation of a covered work with other separate and independent 236 | works, which are not by their nature extensions of the covered work, 237 | and which are not combined with it such as to form a larger program, 238 | in or on a volume of a storage or distribution medium, is called an 239 | "aggregate" if the compilation and its resulting copyright are not 240 | used to limit the access or legal rights of the compilation's users 241 | beyond what the individual works permit. Inclusion of a covered work 242 | in an aggregate does not cause this License to apply to the other 243 | parts of the aggregate. 244 | 245 | 6. Conveying Non-Source Forms. 246 | 247 | You may convey a covered work in object code form under the terms 248 | of sections 4 and 5, provided that you also convey the 249 | machine-readable Corresponding Source under the terms of this License, 250 | in one of these ways: 251 | 252 | a) Convey the object code in, or embodied in, a physical product 253 | (including a physical distribution medium), accompanied by the 254 | Corresponding Source fixed on a durable physical medium 255 | customarily used for software interchange. 256 | 257 | b) Convey the object code in, or embodied in, a physical product 258 | (including a physical distribution medium), accompanied by a 259 | written offer, valid for at least three years and valid for as 260 | long as you offer spare parts or customer support for that product 261 | model, to give anyone who possesses the object code either (1) a 262 | copy of the Corresponding Source for all the software in the 263 | product that is covered by this License, on a durable physical 264 | medium customarily used for software interchange, for a price no 265 | more than your reasonable cost of physically performing this 266 | conveying of source, or (2) access to copy the 267 | Corresponding Source from a network server at no charge. 268 | 269 | c) Convey individual copies of the object code with a copy of the 270 | written offer to provide the Corresponding Source. This 271 | alternative is allowed only occasionally and noncommercially, and 272 | only if you received the object code with such an offer, in accord 273 | with subsection 6b. 274 | 275 | d) Convey the object code by offering access from a designated 276 | place (gratis or for a charge), and offer equivalent access to the 277 | Corresponding Source in the same way through the same place at no 278 | further charge. You need not require recipients to copy the 279 | Corresponding Source along with the object code. If the place to 280 | copy the object code is a network server, the Corresponding Source 281 | may be on a different server (operated by you or a third party) 282 | that supports equivalent copying facilities, provided you maintain 283 | clear directions next to the object code saying where to find the 284 | Corresponding Source. Regardless of what server hosts the 285 | Corresponding Source, you remain obligated to ensure that it is 286 | available for as long as needed to satisfy these requirements. 287 | 288 | e) Convey the object code using peer-to-peer transmission, provided 289 | you inform other peers where the object code and Corresponding 290 | Source of the work are being offered to the general public at no 291 | charge under subsection 6d. 292 | 293 | A separable portion of the object code, whose source code is excluded 294 | from the Corresponding Source as a System Library, need not be 295 | included in conveying the object code work. 296 | 297 | A "User Product" is either (1) a "consumer product", which means any 298 | tangible personal property which is normally used for personal, family, 299 | or household purposes, or (2) anything designed or sold for incorporation 300 | into a dwelling. In determining whether a product is a consumer product, 301 | doubtful cases shall be resolved in favor of coverage. For a particular 302 | product received by a particular user, "normally used" refers to a 303 | typical or common use of that class of product, regardless of the status 304 | of the particular user or of the way in which the particular user 305 | actually uses, or expects or is expected to use, the product. A product 306 | is a consumer product regardless of whether the product has substantial 307 | commercial, industrial or non-consumer uses, unless such uses represent 308 | the only significant mode of use of the product. 309 | 310 | "Installation Information" for a User Product means any methods, 311 | procedures, authorization keys, or other information required to install 312 | and execute modified versions of a covered work in that User Product from 313 | a modified version of its Corresponding Source. The information must 314 | suffice to ensure that the continued functioning of the modified object 315 | code is in no case prevented or interfered with solely because 316 | modification has been made. 317 | 318 | If you convey an object code work under this section in, or with, or 319 | specifically for use in, a User Product, and the conveying occurs as 320 | part of a transaction in which the right of possession and use of the 321 | User Product is transferred to the recipient in perpetuity or for a 322 | fixed term (regardless of how the transaction is characterized), the 323 | Corresponding Source conveyed under this section must be accompanied 324 | by the Installation Information. But this requirement does not apply 325 | if neither you nor any third party retains the ability to install 326 | modified object code on the User Product (for example, the work has 327 | been installed in ROM). 328 | 329 | The requirement to provide Installation Information does not include a 330 | requirement to continue to provide support service, warranty, or updates 331 | for a work that has been modified or installed by the recipient, or for 332 | the User Product in which it has been modified or installed. Access to a 333 | network may be denied when the modification itself materially and 334 | adversely affects the operation of the network or violates the rules and 335 | protocols for communication across the network. 336 | 337 | Corresponding Source conveyed, and Installation Information provided, 338 | in accord with this section must be in a format that is publicly 339 | documented (and with an implementation available to the public in 340 | source code form), and must require no special password or key for 341 | unpacking, reading or copying. 342 | 343 | 7. Additional Terms. 344 | 345 | "Additional permissions" are terms that supplement the terms of this 346 | License by making exceptions from one or more of its conditions. 347 | Additional permissions that are applicable to the entire Program shall 348 | be treated as though they were included in this License, to the extent 349 | that they are valid under applicable law. If additional permissions 350 | apply only to part of the Program, that part may be used separately 351 | under those permissions, but the entire Program remains governed by 352 | this License without regard to the additional permissions. 353 | 354 | When you convey a copy of a covered work, you may at your option 355 | remove any additional permissions from that copy, or from any part of 356 | it. (Additional permissions may be written to require their own 357 | removal in certain cases when you modify the work.) You may place 358 | additional permissions on material, added by you to a covered work, 359 | for which you have or can give appropriate copyright permission. 360 | 361 | Notwithstanding any other provision of this License, for material you 362 | add to a covered work, you may (if authorized by the copyright holders of 363 | that material) supplement the terms of this License with terms: 364 | 365 | a) Disclaiming warranty or limiting liability differently from the 366 | terms of sections 15 and 16 of this License; or 367 | 368 | b) Requiring preservation of specified reasonable legal notices or 369 | author attributions in that material or in the Appropriate Legal 370 | Notices displayed by works containing it; or 371 | 372 | c) Prohibiting misrepresentation of the origin of that material, or 373 | requiring that modified versions of such material be marked in 374 | reasonable ways as different from the original version; or 375 | 376 | d) Limiting the use for publicity purposes of names of licensors or 377 | authors of the material; or 378 | 379 | e) Declining to grant rights under trademark law for use of some 380 | trade names, trademarks, or service marks; or 381 | 382 | f) Requiring indemnification of licensors and authors of that 383 | material by anyone who conveys the material (or modified versions of 384 | it) with contractual assumptions of liability to the recipient, for 385 | any liability that these contractual assumptions directly impose on 386 | those licensors and authors. 387 | 388 | All other non-permissive additional terms are considered "further 389 | restrictions" within the meaning of section 10. If the Program as you 390 | received it, or any part of it, contains a notice stating that it is 391 | governed by this License along with a term that is a further 392 | restriction, you may remove that term. If a license document contains 393 | a further restriction but permits relicensing or conveying under this 394 | License, you may add to a covered work material governed by the terms 395 | of that license document, provided that the further restriction does 396 | not survive such relicensing or conveying. 397 | 398 | If you add terms to a covered work in accord with this section, you 399 | must place, in the relevant source files, a statement of the 400 | additional terms that apply to those files, or a notice indicating 401 | where to find the applicable terms. 402 | 403 | Additional terms, permissive or non-permissive, may be stated in the 404 | form of a separately written license, or stated as exceptions; 405 | the above requirements apply either way. 406 | 407 | 8. Termination. 408 | 409 | You may not propagate or modify a covered work except as expressly 410 | provided under this License. Any attempt otherwise to propagate or 411 | modify it is void, and will automatically terminate your rights under 412 | this License (including any patent licenses granted under the third 413 | paragraph of section 11). 414 | 415 | However, if you cease all violation of this License, then your 416 | license from a particular copyright holder is reinstated (a) 417 | provisionally, unless and until the copyright holder explicitly and 418 | finally terminates your license, and (b) permanently, if the copyright 419 | holder fails to notify you of the violation by some reasonable means 420 | prior to 60 days after the cessation. 421 | 422 | Moreover, your license from a particular copyright holder is 423 | reinstated permanently if the copyright holder notifies you of the 424 | violation by some reasonable means, this is the first time you have 425 | received notice of violation of this License (for any work) from that 426 | copyright holder, and you cure the violation prior to 30 days after 427 | your receipt of the notice. 428 | 429 | Termination of your rights under this section does not terminate the 430 | licenses of parties who have received copies or rights from you under 431 | this License. If your rights have been terminated and not permanently 432 | reinstated, you do not qualify to receive new licenses for the same 433 | material under section 10. 434 | 435 | 9. Acceptance Not Required for Having Copies. 436 | 437 | You are not required to accept this License in order to receive or 438 | run a copy of the Program. Ancillary propagation of a covered work 439 | occurring solely as a consequence of using peer-to-peer transmission 440 | to receive a copy likewise does not require acceptance. However, 441 | nothing other than this License grants you permission to propagate or 442 | modify any covered work. These actions infringe copyright if you do 443 | not accept this License. Therefore, by modifying or propagating a 444 | covered work, you indicate your acceptance of this License to do so. 445 | 446 | 10. Automatic Licensing of Downstream Recipients. 447 | 448 | Each time you convey a covered work, the recipient automatically 449 | receives a license from the original licensors, to run, modify and 450 | propagate that work, subject to this License. You are not responsible 451 | for enforcing compliance by third parties with this License. 452 | 453 | An "entity transaction" is a transaction transferring control of an 454 | organization, or substantially all assets of one, or subdividing an 455 | organization, or merging organizations. If propagation of a covered 456 | work results from an entity transaction, each party to that 457 | transaction who receives a copy of the work also receives whatever 458 | licenses to the work the party's predecessor in interest had or could 459 | give under the previous paragraph, plus a right to possession of the 460 | Corresponding Source of the work from the predecessor in interest, if 461 | the predecessor has it or can get it with reasonable efforts. 462 | 463 | You may not impose any further restrictions on the exercise of the 464 | rights granted or affirmed under this License. For example, you may 465 | not impose a license fee, royalty, or other charge for exercise of 466 | rights granted under this License, and you may not initiate litigation 467 | (including a cross-claim or counterclaim in a lawsuit) alleging that 468 | any patent claim is infringed by making, using, selling, offering for 469 | sale, or importing the Program or any portion of it. 470 | 471 | 11. Patents. 472 | 473 | A "contributor" is a copyright holder who authorizes use under this 474 | License of the Program or a work on which the Program is based. The 475 | work thus licensed is called the contributor's "contributor version". 476 | 477 | A contributor's "essential patent claims" are all patent claims 478 | owned or controlled by the contributor, whether already acquired or 479 | hereafter acquired, that would be infringed by some manner, permitted 480 | by this License, of making, using, or selling its contributor version, 481 | but do not include claims that would be infringed only as a 482 | consequence of further modification of the contributor version. For 483 | purposes of this definition, "control" includes the right to grant 484 | patent sublicenses in a manner consistent with the requirements of 485 | this License. 486 | 487 | Each contributor grants you a non-exclusive, worldwide, royalty-free 488 | patent license under the contributor's essential patent claims, to 489 | make, use, sell, offer for sale, import and otherwise run, modify and 490 | propagate the contents of its contributor version. 491 | 492 | In the following three paragraphs, a "patent license" is any express 493 | agreement or commitment, however denominated, not to enforce a patent 494 | (such as an express permission to practice a patent or covenant not to 495 | sue for patent infringement). To "grant" such a patent license to a 496 | party means to make such an agreement or commitment not to enforce a 497 | patent against the party. 498 | 499 | If you convey a covered work, knowingly relying on a patent license, 500 | and the Corresponding Source of the work is not available for anyone 501 | to copy, free of charge and under the terms of this License, through a 502 | publicly available network server or other readily accessible means, 503 | then you must either (1) cause the Corresponding Source to be so 504 | available, or (2) arrange to deprive yourself of the benefit of the 505 | patent license for this particular work, or (3) arrange, in a manner 506 | consistent with the requirements of this License, to extend the patent 507 | license to downstream recipients. "Knowingly relying" means you have 508 | actual knowledge that, but for the patent license, your conveying the 509 | covered work in a country, or your recipient's use of the covered work 510 | in a country, would infringe one or more identifiable patents in that 511 | country that you have reason to believe are valid. 512 | 513 | If, pursuant to or in connection with a single transaction or 514 | arrangement, you convey, or propagate by procuring conveyance of, a 515 | covered work, and grant a patent license to some of the parties 516 | receiving the covered work authorizing them to use, propagate, modify 517 | or convey a specific copy of the covered work, then the patent license 518 | you grant is automatically extended to all recipients of the covered 519 | work and works based on it. 520 | 521 | A patent license is "discriminatory" if it does not include within 522 | the scope of its coverage, prohibits the exercise of, or is 523 | conditioned on the non-exercise of one or more of the rights that are 524 | specifically granted under this License. You may not convey a covered 525 | work if you are a party to an arrangement with a third party that is 526 | in the business of distributing software, under which you make payment 527 | to the third party based on the extent of your activity of conveying 528 | the work, and under which the third party grants, to any of the 529 | parties who would receive the covered work from you, a discriminatory 530 | patent license (a) in connection with copies of the covered work 531 | conveyed by you (or copies made from those copies), or (b) primarily 532 | for and in connection with specific products or compilations that 533 | contain the covered work, unless you entered into that arrangement, 534 | or that patent license was granted, prior to 28 March 2007. 535 | 536 | Nothing in this License shall be construed as excluding or limiting 537 | any implied license or other defenses to infringement that may 538 | otherwise be available to you under applicable patent law. 539 | 540 | 12. No Surrender of Others' Freedom. 541 | 542 | If conditions are imposed on you (whether by court order, agreement or 543 | otherwise) that contradict the conditions of this License, they do not 544 | excuse you from the conditions of this License. If you cannot convey a 545 | covered work so as to satisfy simultaneously your obligations under this 546 | License and any other pertinent obligations, then as a consequence you may 547 | not convey it at all. For example, if you agree to terms that obligate you 548 | to collect a royalty for further conveying from those to whom you convey 549 | the Program, the only way you could satisfy both those terms and this 550 | License would be to refrain entirely from conveying the Program. 551 | 552 | 13. Use with the GNU Affero General Public License. 553 | 554 | Notwithstanding any other provision of this License, you have 555 | permission to link or combine any covered work with a work licensed 556 | under version 3 of the GNU Affero General Public License into a single 557 | combined work, and to convey the resulting work. The terms of this 558 | License will continue to apply to the part which is the covered work, 559 | but the special requirements of the GNU Affero General Public License, 560 | section 13, concerning interaction through a network will apply to the 561 | combination as such. 562 | 563 | 14. Revised Versions of this License. 564 | 565 | The Free Software Foundation may publish revised and/or new versions of 566 | the GNU General Public License from time to time. Such new versions will 567 | be similar in spirit to the present version, but may differ in detail to 568 | address new problems or concerns. 569 | 570 | Each version is given a distinguishing version number. If the 571 | Program specifies that a certain numbered version of the GNU General 572 | Public License "or any later version" applies to it, you have the 573 | option of following the terms and conditions either of that numbered 574 | version or of any later version published by the Free Software 575 | Foundation. If the Program does not specify a version number of the 576 | GNU General Public License, you may choose any version ever published 577 | by the Free Software Foundation. 578 | 579 | If the Program specifies that a proxy can decide which future 580 | versions of the GNU General Public License can be used, that proxy's 581 | public statement of acceptance of a version permanently authorizes you 582 | to choose that version for the Program. 583 | 584 | Later license versions may give you additional or different 585 | permissions. However, no additional obligations are imposed on any 586 | author or copyright holder as a result of your choosing to follow a 587 | later version. 588 | 589 | 15. Disclaimer of Warranty. 590 | 591 | THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY 592 | APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT 593 | HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY 594 | OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, 595 | THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 596 | PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM 597 | IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF 598 | ALL NECESSARY SERVICING, REPAIR OR CORRECTION. 599 | 600 | 16. Limitation of Liability. 601 | 602 | IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING 603 | WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS 604 | THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY 605 | GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE 606 | USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF 607 | DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD 608 | PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), 609 | EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF 610 | SUCH DAMAGES. 611 | 612 | 17. Interpretation of Sections 15 and 16. 613 | 614 | If the disclaimer of warranty and limitation of liability provided 615 | above cannot be given local legal effect according to their terms, 616 | reviewing courts shall apply local law that most closely approximates 617 | an absolute waiver of all civil liability in connection with the 618 | Program, unless a warranty or assumption of liability accompanies a 619 | copy of the Program in return for a fee. 620 | 621 | END OF TERMS AND CONDITIONS 622 | 623 | How to Apply These Terms to Your New Programs 624 | 625 | If you develop a new program, and you want it to be of the greatest 626 | possible use to the public, the best way to achieve this is to make it 627 | free software which everyone can redistribute and change under these terms. 628 | 629 | To do so, attach the following notices to the program. It is safest 630 | to attach them to the start of each source file to most effectively 631 | state the exclusion of warranty; and each file should have at least 632 | the "copyright" line and a pointer to where the full notice is found. 633 | 634 | 635 | Copyright (C) 636 | 637 | This program is free software: you can redistribute it and/or modify 638 | it under the terms of the GNU General Public License as published by 639 | the Free Software Foundation, either version 3 of the License, or 640 | (at your option) any later version. 641 | 642 | This program is distributed in the hope that it will be useful, 643 | but WITHOUT ANY WARRANTY; without even the implied warranty of 644 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 645 | GNU General Public License for more details. 646 | 647 | You should have received a copy of the GNU General Public License 648 | along with this program. If not, see . 649 | 650 | Also add information on how to contact you by electronic and paper mail. 651 | 652 | If the program does terminal interaction, make it output a short 653 | notice like this when it starts in an interactive mode: 654 | 655 | Copyright (C) 656 | This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'. 657 | This is free software, and you are welcome to redistribute it 658 | under certain conditions; type `show c' for details. 659 | 660 | The hypothetical commands `show w' and `show c' should show the appropriate 661 | parts of the General Public License. Of course, your program's commands 662 | might be different; for a GUI interface, you would use an "about box". 663 | 664 | You should also get your employer (if you work as a programmer) or school, 665 | if any, to sign a "copyright disclaimer" for the program, if necessary. 666 | For more information on this, and how to apply and follow the GNU GPL, see 667 | . 668 | 669 | The GNU General Public License does not permit incorporating your program 670 | into proprietary programs. If your program is a subroutine library, you 671 | may consider it more useful to permit linking proprietary applications with 672 | the library. If this is what you want to do, use the GNU Lesser General 673 | Public License instead of this License. But first, please read 674 | . 675 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # Practical Parallelism in C++ 2 | This repository contains all code from the YouTube series "Practical Parallelism in C++" by CoffeeBeforeArch. 3 | 4 | ## Contact 5 | 6 | Suggestions for specific content can be sent to: CoffeeBeforeArch@gmail.com 7 | 8 | 9 | ## Environment 10 | Operating System: Windows 10 & Ubuntu 18.04 11 | 12 | IDE: Visual Studio 2017 13 | 14 | Text Editor: VIM 15 | 16 | Thread Interfaces: Windows API, Pthreads, C++11 Standard, MPI 17 | 18 | ## Concepts covered in each video 19 | # C++ Threads 20 | | Video | Concepts | Files | 21 | | ----- | -------- | ----- | 22 | |Practical Parallelism in C++: Introduction | Threads, Concurrency, Parallelism | call_from.cpp | 23 | |Practical Parallelism in C++: Mutex Basics | Mutex Locks, Critical Sections | mutex_basics.cpp | 24 | 25 | # Windows Threads 26 | | Video | Concepts | Files | 27 | | ----- | -------- | ----- | 28 | |Practical Parallelism in C++: Windows Threads | Windows Threads, Thread-Safe Functions | windows_threads.cpp | 29 | 30 | # Pthreads 31 | | Video | Concepts | Files | 32 | | ----- | -------- | ----- | 33 | |Practical Parallelism in C++: Basic Pthreads | Pthreads | posix_threads.cpp | 34 | |Practical Parallelism in C++: Broadcast Parallel Gaussian Elimination | Gaussian Elimination, Broadcast Parallel, Pthread Barriers | gaussian.cpp
utils.h
common.h | 35 | |Practical Parallelism in C++: Broadcast-Parallel Gaussian Elimination with Cyclic Mapping | Gaussian Elimination, Broadcast Parallel, Pthread Barriers, Cyclic Striped Mapping | gaussian.cpp
utils.h
common.h | 36 | 37 | # MPI 38 | | Video | Concepts | Files | 39 | | ----- | -------- | ----- | 40 | |Practical Parallelism in C++: MPI Basics | Message Passing, Shared Memory, Ranks, Communicators | mpi_hello.cpp | 41 | |Practical Parallelism in C++: MPI Synchronization | Synchronization, Sending Messages, Receiving Messages, Buffering | mpi_hello_synchronize.cpp | 42 | |Practical Parallelism in C++: MPI Gaussian Elimination Naive | Broadcast, Scatter, Gather | gaussian.cpp | 43 | |Practical Parallelism in C++: MPI Gaussian Elimination Cyclic Striped | Broadcast, Scatter, Gather | gaussian.cpp | 44 | -------------------------------------------------------------------------------- /intro/call_from.cpp: -------------------------------------------------------------------------------- 1 | // This program is a multi-threaded "call from" program in C++ 2 | // By: Nick from CoffeeBeforeArch 3 | 4 | #include 5 | #include 6 | 7 | using std::cout; 8 | using std::endl; 9 | using std::thread; 10 | 11 | // Initial function for each thread 12 | void call_from(int tid) { 13 | cout << "Launched by thread " << tid << endl; 14 | } 15 | 16 | int main() { 17 | // Create an array of 10 thread objects 18 | thread t[10]; 19 | 20 | // Launch 10 threads with initial function "call_from(...)" 21 | for (int i = 0; i < 10; i++) { 22 | t[i] = thread(call_from, i); 23 | } 24 | 25 | // Print from main thread 26 | call_from(10); 27 | 28 | // Wait for all threads to complete (in order) 29 | for (int i = 0; i < 10; i++) { 30 | t[i].join(); 31 | } 32 | 33 | return 0; 34 | } 35 | -------------------------------------------------------------------------------- /intro/lock_basics.cpp: -------------------------------------------------------------------------------- 1 | // This program is a multi-threaded "call from" program in C++ 2 | // By: Nick from CoffeeBeforeArch 3 | 4 | #include 5 | #include 6 | #include 7 | 8 | using std::mutex; 9 | using std::thread; 10 | using std::cout; 11 | using std::endl; 12 | 13 | // Mutex for critical section 14 | mutex mtx; 15 | 16 | // Initial function for each thread 17 | void call_from(int tid) { 18 | mtx.lock(); 19 | cout << "Launched by thread " << tid << endl; 20 | mtx.unlock(); 21 | } 22 | 23 | int main() { 24 | // Create an array of 10 thread objects 25 | thread t[10]; 26 | 27 | // Launch 10 threads with initial function "call_from(...)" 28 | for (int i = 0; i < 10; i++) { 29 | t[i] = thread(call_from, i); 30 | } 31 | 32 | // Print from main thread 33 | call_from(10); 34 | 35 | // Wait for all threads to complete 36 | for (int i = 0; i < 10; i++) { 37 | t[i].join(); 38 | } 39 | 40 | return 0; 41 | } 42 | -------------------------------------------------------------------------------- /intro/modern_lock.cpp: -------------------------------------------------------------------------------- 1 | // This program is a multi-threaded "call from" program in C++ 2 | // By: Nick from CoffeeBeforeArch 3 | 4 | #include 5 | #include 6 | #include 7 | 8 | using std::cout; 9 | using std::endl; 10 | using std::lock_guard; 11 | using std::mutex; 12 | using std::thread; 13 | 14 | // Mutex for critical section 15 | mutex mtx; 16 | 17 | // Initial function for each thread 18 | void call_from(int tid) { 19 | // Use a lock guard to serialize the prints 20 | // Automatically unlocks when it goes out of scope 21 | lock_guard lock(mtx); 22 | std::cout << "Hello from thread - " << tid << endl; 23 | } 24 | 25 | int main() { 26 | // Create an array of 10 thread objects 27 | thread t[10]; 28 | 29 | // Launch 10 threads with initial function "call_from(...)" 30 | for (int i = 0; i < 10; i++) { 31 | t[i] = thread(call_from, i); 32 | } 33 | 34 | // Print from main thread 35 | call_from(10); 36 | 37 | // Wait for all threads to complete 38 | for (int i = 0; i < 10; i++) { 39 | t[i].join(); 40 | } 41 | 42 | return 0; 43 | } 44 | -------------------------------------------------------------------------------- /mpi/mpi_hello.cpp: -------------------------------------------------------------------------------- 1 | // This program shows off the basics of using MPI with C++ 2 | // By: Nick from CoffeeBeforeArch 3 | 4 | #include 5 | #include 6 | 7 | using namespace std; 8 | 9 | int main(int argc, char *argv[]) { 10 | // Unique rank is assigned to each process in a communicator 11 | int rank; 12 | 13 | // Total number of ranks 14 | int size; 15 | 16 | // The machine we are on 17 | char name[80]; 18 | 19 | // Length of the machine name 20 | int length; 21 | 22 | // Initializes the MPI execution environment 23 | MPI_Init(&argc, &argv); 24 | 25 | // Get this process' rank (process within a communicator) 26 | // MPI_COMM_WORLD is the default communicator 27 | MPI_Comm_rank(MPI_COMM_WORLD, &rank); 28 | 29 | // Get the total number ranks in this communicator 30 | MPI_Comm_size(MPI_COMM_WORLD, &size); 31 | 32 | // Gets the name of the processor 33 | // Implementation specific (may be gethostname, uname, or sysinfo) 34 | MPI_Get_processor_name(name, &length); 35 | 36 | // Print out for each rank 37 | cout << "Hello, MPI! Rank: " << rank << " Total: " << size 38 | << " Machine: " << name << endl; 39 | 40 | // Terminate MPI execution environment 41 | MPI_Finalize(); 42 | 43 | return 0; 44 | } 45 | -------------------------------------------------------------------------------- /mpi/mpi_hello_synchronize.cpp: -------------------------------------------------------------------------------- 1 | // This program shows off the basics of using MPI with C++ 2 | // with synchronized output 3 | // By: Nick from CoffeeBeforeArch 4 | 5 | #include 6 | #include 7 | #include 8 | 9 | using namespace std; 10 | 11 | int main(int argc, char *argv[]) { 12 | // Unique rank is assigned to each process in a communicator 13 | int rank; 14 | 15 | // Total number of ranks 16 | int size; 17 | 18 | // The machine we are on 19 | char name[80]; 20 | 21 | // Length of the machine name 22 | int length; 23 | 24 | // Initializes the MPI execution environment 25 | MPI_Init(&argc, &argv); 26 | 27 | // Get this process' rank (process within a communicator) 28 | // MPI_COMM_WORLD is the default communicator 29 | MPI_Comm_rank(MPI_COMM_WORLD, &rank); 30 | 31 | // Get the total number ranks in this communicator 32 | MPI_Comm_size(MPI_COMM_WORLD, &size); 33 | 34 | // Gets the name of the processor 35 | // Implementation specific (may be gethostname, uname, or sysinfo) 36 | MPI_Get_processor_name(name, &length); 37 | 38 | // Pack these values together into a string 39 | int buffer_len = 150; 40 | char buffer[buffer_len]; 41 | sprintf(buffer, "Hello, MPI! Rank: %d Total: %d Machine: %s", rank, size, 42 | name); 43 | 44 | // Synchronize so we can remove interleaved output 45 | if (rank == 0) { 46 | // Always print from rank 0 47 | cout << buffer << endl; 48 | for (int i = 1; i < size; i++) { 49 | // Takes buffer, size, type, source, tag, communicator, and status 50 | MPI_Recv(buffer, buffer_len, MPI_CHAR, i, MPI_ANY_TAG, 51 | MPI_COMM_WORLD, MPI_STATUS_IGNORE); 52 | 53 | // Print our received message 54 | printf("%s\n", buffer); 55 | } 56 | } else { 57 | // If not rank zero, send your message to be printed 58 | MPI_Send(buffer, buffer_len, MPI_CHAR, 0, rank, MPI_COMM_WORLD); 59 | } 60 | 61 | // Terminate MPI execution environment 62 | MPI_Finalize(); 63 | } 64 | -------------------------------------------------------------------------------- /parallel_algorithms/gaussian_elimination/common/common.h: -------------------------------------------------------------------------------- 1 | // This file contains common functions used by the different versions 2 | // of the Gaussian Elimination algorithm 3 | // By: Nick from CoffeeBeforeArch 4 | 5 | #include 6 | #include 7 | #include 8 | #include 9 | #include 10 | 11 | using namespace std; 12 | 13 | // Serial function for computing Gaussian Elimination 14 | // Takes a pointer to a matrix and its dimension as arguments 15 | void ge_serial(float *matrix, int n){ 16 | // Iterate over each row in the matrix 17 | float pivot; 18 | for(int i = 0; i < n - 1; i++){ 19 | // Pivot will be the diagonal 20 | pivot = matrix[i * n + i]; 21 | 22 | // Iterate of the remaining row elements 23 | for(int j = i + 1; j < n; j++){ 24 | // Divide by the pivot 25 | matrix[i * n + j] /= pivot; 26 | } 27 | 28 | // Do direct assignment for trivial case (self-divide) 29 | matrix[i * n + i] = 1.0; 30 | 31 | // Eliminate ith element from the jth row 32 | float scale; 33 | for(int j = i + 1; j < n; j++){ 34 | // Factor we will use to scale subtraction by 35 | scale = matrix[j * n + i]; 36 | 37 | // Iterate over the remaining columns 38 | for(int k = i + 1; k < n; k++){ 39 | matrix[j * n + k] -= matrix[i * n + k] * scale; 40 | } 41 | 42 | // Do direct assignment for trivial case (eliminate position) 43 | matrix[j * n + i] = 0; 44 | } 45 | } 46 | 47 | // Trivial case of last row with only one element 48 | matrix[(n - 1) * n + n - 1] = 1; 49 | } 50 | 51 | // Initialize a matrix with random numbers 52 | // Takes a matrix and its dimension as arguments 53 | void init_matrix(float *matrix, int N){ 54 | srand(time(NULL)); 55 | for(int i = 0; i < N; i++){ 56 | for(int j = 0; j < N; j++){ 57 | // Random number between 100 and -100 58 | matrix[i * N + j] = (float(rand())/float(RAND_MAX)) * (100 - -100) + -100; 59 | } 60 | } 61 | } 62 | 63 | // Prints a matrix 64 | // Takes a matrix and its dimension as arguments 65 | void print_matrix(float *matrix, int N){ 66 | for(int i = 0; i < N; i++){ 67 | for(int j = 0; j < N; j++){ 68 | cout << setprecision(3) << matrix[i * N + j] << "\t"; 69 | 70 | } 71 | cout << endl; 72 | } 73 | cout << endl; 74 | } 75 | 76 | // Verifies the solution of Gaussian Elimination to the serial impl. 77 | // Takes two matrices and a their dimensions as arguments 78 | void verify_solution(float *matrix1, float *matrix2, int N){ 79 | // Error can not exceed this bound 80 | float epsilon = 0.005; 81 | for(int i = 0; i < N; i++){ 82 | for(int j = 0; j < N; j++){ 83 | // Fail if error exceeds epsilon 84 | assert(abs(matrix1[i * N + j] - matrix2[i * N + j]) <= epsilon); 85 | } 86 | } 87 | } 88 | 89 | -------------------------------------------------------------------------------- /parallel_algorithms/gaussian_elimination/mpi/cyclic_striped_mapping/gaussian.cpp: -------------------------------------------------------------------------------- 1 | // This program implements parallel gaussian elimination in C++ using 2 | // MPI and cyclic striped mapping (assumes square matrix) 3 | // By: Nick from CoffeeBeforeArch 4 | 5 | #include 6 | #include 7 | #include 8 | #include "../../common/common.h" 9 | 10 | int main(int argc, char *argv[]){ 11 | // Declare a problem size 12 | int N = 1024; 13 | 14 | // Declate variables for timing 15 | double t_start; 16 | double t_end; 17 | double t_total; 18 | 19 | // Unique rank for this process 20 | int rank; 21 | 22 | // Total number of ranks 23 | int size; 24 | 25 | // Initializes the MPI execution environment 26 | MPI_Init(&argc, &argv); 27 | 28 | // Get the rank 29 | MPI_Comm_rank(MPI_COMM_WORLD, &rank); 30 | 31 | // Get the total number ranks in this communicator 32 | MPI_Comm_size(MPI_COMM_WORLD, &size); 33 | 34 | // Calulate the number of rows based on the number of ranks 35 | int num_rows = N / size; 36 | 37 | /* 38 | * Distribute Work to Ranks: 39 | * Rank 0 needs to send the appropriate rows to each process 40 | * before they are able to proceed 41 | */ 42 | // Declare our problem matrices 43 | // This work is duplicated just for code simplicity 44 | float *matrix; 45 | if(rank == 0){ 46 | // Only rank 0 needs space for the total solution 47 | matrix = new float[N * N]; 48 | 49 | // Initialize the matrix 50 | init_matrix(matrix, N); 51 | } 52 | 53 | // Declare our sub-matrix for each process 54 | float *sub_matrix = new float[N * num_rows]; 55 | 56 | // Cyclic stripe the rows to all the ranks 57 | if(size == 1){ 58 | // All rows to the single rank 59 | memcpy(sub_matrix, matrix, N * N * sizeof(float)); 60 | }else{ 61 | // Scatter "num_rows" rows to "size" processes 62 | for(int i = 0; i < num_rows; i++){ 63 | MPI_Scatter(&matrix[i * N * size], N, MPI_FLOAT, 64 | &sub_matrix[i * N], N, MPI_FLOAT, 0, MPI_COMM_WORLD); 65 | } 66 | } 67 | 68 | /* 69 | * Gaussian Elimination: 70 | * One rank normalizes the pivot row, then sends it to all 71 | * later ranks for elimination 72 | */ 73 | // Allocate space for a single row to be sent to this rank 74 | float *row = new float[N]; 75 | 76 | // Get start time 77 | if(rank == 0){ 78 | t_start = MPI_Wtime(); 79 | } 80 | 81 | // Local variables for code clarity 82 | int local_row; 83 | int which_rank; 84 | int pivot; 85 | int scale; 86 | 87 | // Iterate over all rows 88 | for(int i = 0; i < N; i++){ 89 | // Which row in the sub-matrix are we accessing? 90 | local_row = i / size; 91 | // Which rank does this row belong to? 92 | which_rank = i % size; 93 | 94 | // Eliminate if the pivot belongs to this rank 95 | if(rank == which_rank){ 96 | pivot = sub_matrix[local_row * N + i]; 97 | 98 | // Divide the rest of the row by the pivot 99 | for(int j = i + 1; j < N; j++){ 100 | sub_matrix[local_row * N + j] /= pivot; 101 | } 102 | 103 | // Use assignment for the trivial self-division 104 | sub_matrix[local_row * N + i] = 1; 105 | 106 | // Copy the row into our send buffer 107 | memcpy(row, &sub_matrix[local_row * N], N * sizeof(float)); 108 | 109 | // Broadcast this row to all the ranks 110 | MPI_Bcast(row, N, MPI_FLOAT, which_rank, MPI_COMM_WORLD); 111 | 112 | // Eliminate for the other rows mapped to this rank 113 | for(int j = local_row + 1; j < num_rows; j++){ 114 | scale = sub_matrix[j * N + i]; 115 | 116 | // Subtract to eliminate pivot from later rows 117 | for(int k = i + 1; k < N; k++){ 118 | sub_matrix[j * N + k] -= scale * row[k]; 119 | } 120 | 121 | // Use assignment for the trivial elimination 122 | sub_matrix[j * N + i] = 0; 123 | } 124 | }else{ 125 | // Receive a row to use for elimination 126 | MPI_Bcast(row, N, MPI_FLOAT, which_rank, MPI_COMM_WORLD); 127 | 128 | // Eliminate for all the rows mapped to this rank 129 | for(int j = local_row; j < num_rows; j++){ 130 | if((which_rank < rank) || (j > local_row)){ 131 | scale = sub_matrix[j * N + i]; 132 | 133 | //Subtract to eliminate pivot from later rows 134 | for(int k = i + 1; k < N; k++){ 135 | sub_matrix[j * N + k] -= scale * row[k]; 136 | } 137 | 138 | // Use assignment for the trivial elimination 139 | sub_matrix[j * N + i] = 0; 140 | } 141 | } 142 | } 143 | } 144 | 145 | // Barrier to track when calculations are done 146 | MPI_Barrier(MPI_COMM_WORLD); 147 | 148 | // Stop the time before the gather phase 149 | if(rank == 0){ 150 | t_end = MPI_Wtime(); 151 | t_total = t_end - t_start; 152 | } 153 | 154 | /* 155 | * Collect all Sub-Matrices 156 | * All sub-matrices are gathered using the gather function 157 | */ 158 | if(size == 1){ 159 | memcpy(matrix, sub_matrix, N * N * sizeof(float)); 160 | }else{ 161 | // Gather "size" rows at a time 162 | for(int i = 0; i < num_rows; i++){ 163 | MPI_Gather(&sub_matrix[i * N], N, MPI_FLOAT, 164 | &matrix[i * size * N], N, MPI_FLOAT, 0, MPI_COMM_WORLD); 165 | } 166 | } 167 | MPI_Finalize(); 168 | 169 | // Print the output and the time 170 | if(rank == 0){ 171 | //print_matrix(matrix, N); 172 | cout << t_total << " Seconds" << endl; 173 | } 174 | 175 | // Free heap-allocated memory 176 | if(rank == 0){ 177 | delete[] matrix; 178 | } 179 | delete[] sub_matrix; 180 | delete[] row; 181 | 182 | return 0; 183 | } 184 | 185 | -------------------------------------------------------------------------------- /parallel_algorithms/gaussian_elimination/mpi/naive/gaussian.cpp: -------------------------------------------------------------------------------- 1 | // This program implements parallel gaussian elimination in C++ using 2 | // MPI and block mapping (assumes square matrix) 3 | // By: Nick from CoffeeBeforeArch 4 | 5 | #include 6 | #include 7 | #include 8 | #include "../../common/common.h" 9 | 10 | int main(int argc, char *argv[]){ 11 | // Declare a problem size 12 | int N = 1024; 13 | 14 | // Declate variables for timing 15 | double t_start; 16 | double t_end; 17 | double t_total; 18 | 19 | // Unique rank for this process 20 | int rank; 21 | 22 | // Total number of ranks 23 | int size; 24 | 25 | // Initializes the MPI execution environment 26 | MPI_Init(&argc, &argv); 27 | 28 | // Get the rank 29 | MPI_Comm_rank(MPI_COMM_WORLD, &rank); 30 | 31 | // Get the total number ranks in this communicator 32 | MPI_Comm_size(MPI_COMM_WORLD, &size); 33 | 34 | // Calulate the number of rows based on the number of ranks 35 | int num_rows = N / size; 36 | 37 | /* 38 | * Distribute Work to Ranks: 39 | * Rank 0 needs to send the appropriate rows to each process 40 | * before they are able to proceed 41 | */ 42 | // Declare our problem matrices 43 | // This work is duplicated just for code simplicity 44 | float *matrix; 45 | 46 | // Only rank 0 needs space for the total solution 47 | if(rank == 0){ 48 | matrix = new float [N * N]; 49 | 50 | // Initialize the matrix 51 | init_matrix(matrix, N); 52 | } 53 | 54 | // Declare our sub-matrix for each process 55 | float *sub_matrix = new float[N * num_rows]; 56 | 57 | // Send a sub-matrix to each process 58 | MPI_Scatter(matrix, N * num_rows, MPI_FLOAT, sub_matrix, 59 | N * num_rows, MPI_FLOAT, 0, MPI_COMM_WORLD); 60 | 61 | /* 62 | * Gaussian Elimination: 63 | * One row normalizes the pivot row, then sends it to all 64 | * later ranks for elimination 65 | */ 66 | // Allocate space for a single row to be sent to this rank 67 | float *row = new float[N]; 68 | 69 | // Get start time 70 | if(rank == 0){ 71 | t_start = MPI_Wtime(); 72 | } 73 | 74 | // Variables for code clarity 75 | int pivot; 76 | int scale; 77 | int column; 78 | int start_row; 79 | 80 | // Receivers go here 81 | start_row = rank * num_rows; 82 | for(int i = 0; i < start_row; i++){ 83 | // Wait for the preceeding ranks to forward us a row 84 | MPI_Bcast(row, N, MPI_FLOAT, i / num_rows, MPI_COMM_WORLD); 85 | 86 | // Eliminate from this element from all rows mapped to this 87 | // rank 88 | for(int j = 0; j < num_rows; j++){ 89 | scale = sub_matrix[j * N + i]; 90 | 91 | // Subtract from all other elements in the row 92 | for(int k = i + 1; k < N; k++){ 93 | sub_matrix[j * N + k] -= scale * row[k]; 94 | } 95 | 96 | // Eliminate the element in the same column as the pivot row 97 | sub_matrix[j * N + i] = 0; 98 | } 99 | } 100 | 101 | // Senders go here 102 | for(int i = 0; i < num_rows; i++){ 103 | // Normalize this row to the pivot 104 | column = rank * num_rows + i; 105 | pivot = sub_matrix[i * N + column]; 106 | 107 | // Normalize every other element in this row to the pivot 108 | for(int j = column + 1; j < N; j++){ 109 | sub_matrix[i * N + j] /= pivot; 110 | } 111 | 112 | // Normalize trivial case 113 | sub_matrix[i * N + column] = 1; 114 | 115 | // Fill row to be sent 116 | memcpy(row, &sub_matrix[i * N], N * sizeof(float)); 117 | 118 | // Broadcast the normalized row to everyone else; 119 | MPI_Bcast(row, N, MPI_FLOAT, rank, MPI_COMM_WORLD); 120 | 121 | // Update the rest of the rows for this rank 122 | for(int j = i + 1; j < num_rows; j++){ 123 | scale = sub_matrix[j * N + column]; 124 | 125 | // Subtract from all the elements in a lower row 126 | for(int k = column + 1; k < N; k++){ 127 | sub_matrix[j * N + k] -= scale * row[k]; 128 | } 129 | 130 | // Eliminate the trivial case 131 | sub_matrix[j * N + column] = 0; 132 | } 133 | } 134 | 135 | // Finished ranks must still wait with synchronous broadcast 136 | for(int i = rank * num_rows + 1; i < N; i++){ 137 | MPI_Bcast(row, N, MPI_FLOAT, i / num_rows, MPI_COMM_WORLD); 138 | } 139 | 140 | // Barrier to track when calculations are done 141 | MPI_Barrier(MPI_COMM_WORLD); 142 | 143 | // Stop the time before the gather phase 144 | if(rank == 0){ 145 | t_end = MPI_Wtime(); 146 | t_total = t_end - t_start; 147 | } 148 | 149 | /* 150 | * Collect all Sub-Matrices 151 | * All sub-matrices are gathered using the gather function 152 | */ 153 | MPI_Gather(sub_matrix, N * num_rows, MPI_FLOAT, matrix, 154 | N * num_rows, MPI_FLOAT, 0, MPI_COMM_WORLD); 155 | 156 | 157 | MPI_Finalize(); 158 | 159 | // Check the result, and print the time 160 | if(rank == 0){ 161 | //print_matrix(matrix, N); 162 | cout << t_total << " Seconds" << endl; 163 | } 164 | 165 | // Free heap-allocated memory 166 | if(rank == 0){ 167 | delete[] matrix; 168 | } 169 | delete[] sub_matrix; 170 | delete[] row; 171 | 172 | return 0; 173 | } 174 | 175 | -------------------------------------------------------------------------------- /parallel_algorithms/gaussian_elimination/pthreads/cyclic_striped_mapping/gaussian.cpp: -------------------------------------------------------------------------------- 1 | // This program implements parallel gaussian elimination in C++ using 2 | // Pthreads (assumes square matrix) and assigns rows to each thread 3 | // usign cyclic striped mapping for load balancing 4 | // By: Nick from CoffeeBeforeArch 5 | 6 | #include 7 | #include "utils.h" 8 | 9 | int main(){ 10 | // Number of threads to launch 11 | int num_threads = 8; 12 | 13 | // Dimensions of square matrix 14 | int N = 2048; 15 | 16 | // Declare our problem matrices 17 | float *matrix; 18 | float *matrix_pthread; 19 | 20 | // Declare and initialize the size of the matrix 21 | size_t bytes = N * N * sizeof(float); 22 | 23 | // Allocate space for our matrices 24 | matrix = new float[N * N]; 25 | matrix_pthread = new float[N * N]; 26 | 27 | // Initialize a matrix and copy it 28 | init_matrix(matrix, N); 29 | memcpy(matrix_pthread, matrix, bytes); 30 | 31 | // Launch the threads via a helper function 32 | launch_threads(num_threads, matrix_pthread, N); 33 | 34 | // Create timers for our serial version 35 | high_resolution_clock::time_point start; 36 | high_resolution_clock::time_point end; 37 | 38 | // Call the serial version for our reference solution 39 | start = high_resolution_clock::now(); 40 | ge_serial(matrix, N); 41 | end = high_resolution_clock::now(); 42 | 43 | // Cast timers as double to print 44 | duration elapsed = duration_cast>(end - start); 45 | 46 | // Print out the elapsed time 47 | cout << "Elapsed time serial = " << elapsed.count() << " seconds" << endl; 48 | 49 | // Verify the solution 50 | verify_solution(matrix, matrix_pthread, N); 51 | 52 | // Free our heap-allocated memory 53 | delete[] matrix; 54 | delete[] matrix_pthread; 55 | 56 | return 0; 57 | } 58 | -------------------------------------------------------------------------------- /parallel_algorithms/gaussian_elimination/pthreads/cyclic_striped_mapping/utils.h: -------------------------------------------------------------------------------- 1 | // This file contains utility functions for the pthread parallel 2 | // Gaussian Elimination 3 | // By: Nick from CoffeeBeforeArch 4 | 5 | #include 6 | #include 7 | #include "../../common/common.h" 8 | 9 | using namespace std::chrono; 10 | 11 | struct Args { 12 | // Threaed ID 13 | int tid; 14 | // Number of threads launched 15 | int num_threads; 16 | // Matrix of floating point numbers 17 | float *matrix; 18 | // Dimensions of the square matrix 19 | int N; 20 | // Barrier to synchronize at 21 | pthread_barrier_t *barrier; 22 | // Variables needed for timing 23 | int *counter; 24 | pthread_mutex_t *mtx; 25 | pthread_cond_t *cond; 26 | high_resolution_clock::time_point *start; 27 | high_resolution_clock::time_point *end; 28 | }; 29 | 30 | void perf_cycle(int num_threads, int *counter, pthread_mutex_t *mtx, 31 | pthread_cond_t *cond, 32 | high_resolution_clock::time_point *time){ 33 | // Get the lock 34 | pthread_mutex_lock(mtx); 35 | 36 | // Atomically decrement number of outstanding threads 37 | *counter -= 1; 38 | // Check if we are the last thread 39 | // If not, wait to be signaled 40 | if(*counter == 0){ 41 | // Update a timing variable 42 | *time = high_resolution_clock::now(); 43 | 44 | // Reset the counter 45 | *counter = num_threads; 46 | 47 | // Signal everyone to continue 48 | pthread_cond_broadcast(cond); 49 | }else{ 50 | // Wait for the last thread before continuing 51 | pthread_cond_wait(cond, mtx); 52 | } 53 | 54 | // Everyone unlocks 55 | pthread_mutex_unlock(mtx); 56 | } 57 | 58 | // Pthread function for computing Gaussian Elimination 59 | // Takes a pointer to a struct of args as an argument 60 | void *ge_parallel(void *args){ 61 | // Cast void pointer to struct pointer 62 | Args *local_args = (Args*)args; 63 | 64 | // Unpack the arguments 65 | int tid = local_args->tid; 66 | int num_threads = local_args->num_threads; 67 | float *matrix = local_args->matrix; 68 | int N = local_args->N; 69 | pthread_barrier_t *barrier = local_args->barrier; 70 | 71 | int *counter = local_args->counter; 72 | pthread_mutex_t *mtx = local_args->mtx; 73 | pthread_cond_t *cond = local_args->cond; 74 | high_resolution_clock::time_point *start = local_args->start; 75 | high_resolution_clock::time_point *end = local_args->end; 76 | 77 | // Wait for all threads to be created before profiling 78 | perf_cycle(num_threads, counter, mtx, cond, start); 79 | 80 | // Loop over all rows in the matrix 81 | for(int i = 0; i < N - 1; i++){ 82 | // Check if pivot row belongs to this thread 83 | if((i % num_threads) == tid){ 84 | // Normalize this row to the pivot 85 | float pivot = matrix[i * N + i]; 86 | 87 | // Loop over remaining elements in the pivot row 88 | for(int j = i + 1; j < N; j++){ 89 | matrix[i * N + j] /= pivot; 90 | } 91 | 92 | // Use assignment for trivial case 93 | matrix[i * N + i] = 1; 94 | } 95 | 96 | // All threads must wait for pivot before continuing 97 | pthread_barrier_wait(barrier); 98 | 99 | // Loop over the rest of the rows to eliminate the ith element 100 | for(int j = i + 1; j < N; j++){ 101 | // Check if row belongs to this thread 102 | if((j % num_threads) == tid){ 103 | // Scale the subtraction by the ith element of this row 104 | float scale = matrix[j * N + i]; 105 | 106 | // Subtract from each element of the row 107 | for(int l = i + 1; l < N; l++){ 108 | matrix[j * N + l] -= matrix[i * N + l] * scale; 109 | } 110 | 111 | // Use assignment for trivial case 112 | matrix[j * N + i] = 0; 113 | } 114 | } 115 | } 116 | 117 | // Handle trivial last row with only 1 element 118 | if(tid == num_threads -1){ 119 | matrix[(N - 1) * N + N - 1] = 1; 120 | } 121 | 122 | // Stop monitoring when last thread exits 123 | perf_cycle(num_threads, counter, mtx, cond, end); 124 | 125 | return 0; 126 | } 127 | 128 | // Helper function create thread 129 | void launch_threads(int num_threads, float* matrix, int N){ 130 | // Create array of thread objects we will launch 131 | pthread_t *threads = new pthread_t[num_threads]; 132 | 133 | // Create a barrier and initialize it 134 | pthread_barrier_t barrier; 135 | pthread_barrier_init(&barrier, NULL, num_threads); 136 | 137 | // Create an array of structs to pass to the threads 138 | Args thread_args[num_threads]; 139 | 140 | // Create variables for performance monitoring 141 | int counter = num_threads; 142 | pthread_mutex_t mtx =PTHREAD_MUTEX_INITIALIZER; 143 | pthread_cond_t cond = PTHREAD_COND_INITIALIZER; 144 | high_resolution_clock::time_point start; 145 | high_resolution_clock::time_point end; 146 | 147 | // Launch threads 148 | for(int i = 0; i < num_threads; i++){ 149 | // Pack struct with its arguments 150 | thread_args[i].tid = i; 151 | thread_args[i].num_threads = num_threads; 152 | thread_args[i].matrix = matrix; 153 | thread_args[i].N = N; 154 | thread_args[i].barrier = &barrier; 155 | 156 | thread_args[i].counter = &counter; 157 | thread_args[i].mtx = &mtx; 158 | thread_args[i].cond = &cond; 159 | thread_args[i].start = &start; 160 | thread_args[i].end = &end; 161 | 162 | // Launch the thread 163 | pthread_create(&threads[i], NULL, ge_parallel, (void*)&thread_args[i]); 164 | } 165 | 166 | for(int i = 0; i < num_threads; i++){ 167 | pthread_join(threads[i], NULL); 168 | } 169 | 170 | // Cast timers as double to print 171 | duration elapsed = duration_cast>(end - start); 172 | 173 | // Print out the elapsed time 174 | cout << "Elapsed time parallel = " << elapsed.count() << " seconds" << endl; 175 | } 176 | 177 | -------------------------------------------------------------------------------- /parallel_algorithms/gaussian_elimination/pthreads/naive/gaussian.cpp: -------------------------------------------------------------------------------- 1 | // This program implements parallel gaussian elimination in C++ using 2 | // Pthreads (assumes square matrix) 3 | // By: Nick from CoffeeBeforeArch 4 | 5 | #include 6 | #include "utils.h" 7 | 8 | int main(){ 9 | // Number of threads to launch 10 | int num_threads = 8; 11 | 12 | // Dimensions of square matrix 13 | int N = 2048; 14 | 15 | // Declare our problem matrices 16 | float *matrix; 17 | float *matrix_pthread; 18 | 19 | // Declare and initialize the size of the matrix 20 | size_t bytes = N * N * sizeof(float); 21 | 22 | // Allocate space for our matrices 23 | matrix = new float[N * N]; 24 | matrix_pthread = new float[N * N]; 25 | 26 | // Initialize a matrix and copy it 27 | init_matrix(matrix, N); 28 | memcpy(matrix_pthread, matrix, bytes); 29 | 30 | // Launch the threads via a helper function 31 | // Prints out time in seconds 32 | launch_threads(num_threads, matrix_pthread, N); 33 | 34 | // Create timers for our serial version 35 | high_resolution_clock::time_point start; 36 | high_resolution_clock::time_point end; 37 | 38 | // Call the serial version for our reference solution 39 | start = high_resolution_clock::now(); 40 | ge_serial(matrix, N); 41 | end = high_resolution_clock::now(); 42 | 43 | // Cast timers as double to print 44 | duration elapsed = duration_cast>(end - start); 45 | 46 | // Print out the elapsed time 47 | cout << "Elapsed time serial = " << elapsed.count() << " seconds" << endl; 48 | 49 | // Verify the solution 50 | verify_solution(matrix, matrix_pthread, N); 51 | 52 | // Free heap-allocated memory 53 | delete[] matrix; 54 | delete[] matrix_pthread; 55 | 56 | return 0; 57 | } 58 | 59 | -------------------------------------------------------------------------------- /parallel_algorithms/gaussian_elimination/pthreads/naive/utils.h: -------------------------------------------------------------------------------- 1 | // This file contains utility functions for the pthread parallel 2 | // Gaussian Elimination 3 | // By: Nick from CoffeeBeforeArch 4 | 5 | #include 6 | #include 7 | #include "../../common/common.h" 8 | 9 | using namespace std::chrono; 10 | 11 | struct Args { 12 | // First row assigned to this thread 13 | int start_row; 14 | // One past the last row for this thread 15 | int end_row; 16 | // Matrix of floating point numbers 17 | float *matrix; 18 | // Dimensions of the square matrix 19 | int N; 20 | // Barrier to synchronize at 21 | pthread_barrier_t *barrier; 22 | // Variables needed for timing 23 | int num_threads; 24 | int *counter; 25 | pthread_mutex_t *mtx; 26 | pthread_cond_t *cond; 27 | high_resolution_clock::time_point *start; 28 | high_resolution_clock::time_point *end; 29 | }; 30 | 31 | void perf_cycle(int num_threads, int *counter, pthread_mutex_t *mtx, 32 | pthread_cond_t *cond, 33 | high_resolution_clock::time_point *time){ 34 | // Get the lock 35 | pthread_mutex_lock(mtx); 36 | 37 | // Atomically decrement number of outstanding threads 38 | *counter -= 1; 39 | // Check if we are the last thread 40 | // If not, wait to be signaled 41 | if(*counter == 0){ 42 | // Update a timing variable 43 | *time = high_resolution_clock::now(); 44 | 45 | // Reset the counter 46 | *counter = num_threads; 47 | 48 | // Signal everyone to continue 49 | pthread_cond_broadcast(cond); 50 | }else{ 51 | // Wait for the last thread before continuing 52 | pthread_cond_wait(cond, mtx); 53 | } 54 | 55 | // Everyone unlocks 56 | pthread_mutex_unlock(mtx); 57 | } 58 | 59 | // Pthread function for computing Gaussian Elimination 60 | // Takes a pointer to a struct of args as an argument 61 | void *ge_parallel(void *args){ 62 | // Cast void pointer to struct pointer 63 | Args *local_args = (Args*)args; 64 | 65 | // Unpack the arguments 66 | int start_row = local_args->start_row; 67 | int end_row = local_args->end_row; 68 | float *matrix = local_args->matrix; 69 | int N = local_args->N; 70 | pthread_barrier_t *barrier = local_args->barrier; 71 | 72 | int num_threads = local_args->num_threads; 73 | int *counter = local_args->counter; 74 | pthread_mutex_t *mtx = local_args->mtx; 75 | pthread_cond_t *cond = local_args->cond; 76 | high_resolution_clock::time_point *start = local_args->start; 77 | high_resolution_clock::time_point *end = local_args->end; 78 | 79 | // Wait for all threads to be created before profiling 80 | perf_cycle(num_threads, counter, mtx, cond, start); 81 | 82 | // Loop over all rows in the matrix 83 | for(int i = 0; i < N - 1; i++){ 84 | // Check if pivot row belongs to this thread 85 | if((i >= start_row) && (i < end_row)){ 86 | // Normalize this row to the pivot 87 | float pivot = matrix[i * N + i]; 88 | 89 | // Loop over remaining elements in the pivot row 90 | for(int j = i + 1; j < N; j++){ 91 | matrix[i * N + j] /= pivot; 92 | } 93 | 94 | // Use assignment for trivial case 95 | matrix[i * N + i] = 1; 96 | } 97 | 98 | // All threads must wait for pivot before continuing 99 | pthread_barrier_wait(barrier); 100 | 101 | // Loop over the rest of the rows to eliminate the ith element 102 | for(int j = i + 1; j < end_row; j++){ 103 | // Check if row belongs to this thread 104 | if((j >= start_row) && (j < end_row)){ 105 | // Scale the subtraction by the ith element of this row 106 | float scale = matrix[j * N + i]; 107 | 108 | // Subtract from each element of the row 109 | for(int l = i + 1; l < N; l++){ 110 | matrix[j * N + l] -= matrix[i * N + l] * scale; 111 | } 112 | 113 | // Use assignment for trivial case 114 | matrix[j * N + i] = 0; 115 | } 116 | } 117 | } 118 | 119 | // Handle trivial last row with only 1 element 120 | if((N - 1) >= start_row){ 121 | matrix[(N - 1) * N + N - 1] = 1; 122 | } 123 | 124 | // Stop monitoring when last thread exits 125 | perf_cycle(num_threads, counter, mtx, cond, end); 126 | 127 | return 0; 128 | } 129 | 130 | // Helper function create thread 131 | void launch_threads(int num_threads, float* matrix, int N){ 132 | 133 | // Create array of thread objects we will launch 134 | pthread_t threads[num_threads]; 135 | 136 | // Create a barrier and initialize it 137 | pthread_barrier_t barrier; 138 | pthread_barrier_init(&barrier, NULL, num_threads); 139 | 140 | // Create an array of structs to pass to the threads 141 | Args thread_args[num_threads]; 142 | 143 | // Create variables for performance monitoring 144 | int counter = num_threads; 145 | pthread_mutex_t mtx = PTHREAD_MUTEX_INITIALIZER; 146 | pthread_cond_t cond = PTHREAD_COND_INITIALIZER; 147 | high_resolution_clock::time_point start; 148 | high_resolution_clock::time_point end; 149 | 150 | // Launch threads 151 | for(int i = 0; i < num_threads; i++){ 152 | // Pack struct with its arguments 153 | thread_args[i].start_row = i * (N / num_threads); 154 | thread_args[i].end_row = i * (N / num_threads) + (N / num_threads); 155 | thread_args[i].matrix = matrix; 156 | thread_args[i].N = N; 157 | thread_args[i].barrier = &barrier; 158 | 159 | thread_args[i].num_threads = num_threads; 160 | thread_args[i].counter = &counter; 161 | thread_args[i].mtx = &mtx; 162 | thread_args[i].cond = &cond; 163 | thread_args[i].start = &start; 164 | thread_args[i].end = &end; 165 | 166 | // Launch the thread 167 | pthread_create(&threads[i], NULL, ge_parallel, (void*)&thread_args[i]); 168 | } 169 | 170 | for(int i = 0; i < num_threads; i++){ 171 | pthread_join(threads[i], NULL); 172 | } 173 | 174 | // Cast timers as double to print 175 | duration elapsed = duration_cast>(end - start); 176 | 177 | // Print out the elapsed time 178 | cout << "Elapsed time parallel = " << elapsed.count() << " seconds" << endl; 179 | } 180 | 181 | -------------------------------------------------------------------------------- /pthreads/posix_threads.cpp: -------------------------------------------------------------------------------- 1 | // This program shows off multithreading basics using pthreads 2 | // By: Nick from CoffeeBeforeArch 3 | 4 | #include 5 | #include 6 | 7 | using namespace std; 8 | 9 | #define NUM_THREADS 10 10 | 11 | // Global pthread implementation of mutex lock 12 | pthread_mutex_t mtx; 13 | 14 | // Argument we will cast as void* for your thread function 15 | struct Args { 16 | int tid; 17 | int data; 18 | }; 19 | 20 | // pthread functions must have return type void*, with a void* arg 21 | void *boring_thread_function(void *args) { 22 | Args *local_args = (Args *)args; 23 | // Get the lock before entering the critical section 24 | pthread_mutex_lock(&mtx); 25 | cout << "Printing from thread " << local_args->tid << endl; 26 | cout << "Value of data is " << local_args->data << endl; 27 | // Release the lock to someone else 28 | pthread_mutex_unlock(&mtx); 29 | } 30 | 31 | int main() { 32 | // Create an array of 10 threads 33 | pthread_t threads[NUM_THREADS]; 34 | Args per_thread_args[NUM_THREADS]; 35 | 36 | // Create 10 threads 37 | for (int i = 0; i < NUM_THREADS; i++) { 38 | per_thread_args[i].tid = i; 39 | per_thread_args[i].data = i * i; 40 | // Launch the threads 41 | /* 42 | Arguments: 43 | 1.) Pointer to pthread_t 44 | 2.) Attributes for the thread (NULL means default) 45 | 3.) Entry routine 46 | 4.) Arguments 47 | */ 48 | pthread_create(&threads[i], NULL, boring_thread_function, 49 | (void *)&per_thread_args[i]); 50 | } 51 | 52 | // Return code for the thread 53 | void *ret; 54 | // Wait for all threads to finish before exiting the program 55 | for (int i = 0; i < NUM_THREADS; i++) { 56 | pthread_join(threads[i], &ret); 57 | } 58 | 59 | return 0; 60 | } 61 | -------------------------------------------------------------------------------- /windows_threads/windows_threads.sln: -------------------------------------------------------------------------------- 1 |  2 | Microsoft Visual Studio Solution File, Format Version 12.00 3 | # Visual Studio 15 4 | VisualStudioVersion = 15.0.28307.438 5 | MinimumVisualStudioVersion = 10.0.40219.1 6 | Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "windows_threads", "windows_threads\windows_threads.vcxproj", "{E4493801-7A78-47C0-97BA-1D7F515AA0A8}" 7 | EndProject 8 | Global 9 | GlobalSection(SolutionConfigurationPlatforms) = preSolution 10 | Debug|x64 = Debug|x64 11 | Debug|x86 = Debug|x86 12 | Release|x64 = Release|x64 13 | Release|x86 = Release|x86 14 | EndGlobalSection 15 | GlobalSection(ProjectConfigurationPlatforms) = postSolution 16 | {E4493801-7A78-47C0-97BA-1D7F515AA0A8}.Debug|x64.ActiveCfg = Debug|x64 17 | {E4493801-7A78-47C0-97BA-1D7F515AA0A8}.Debug|x64.Build.0 = Debug|x64 18 | {E4493801-7A78-47C0-97BA-1D7F515AA0A8}.Debug|x86.ActiveCfg = Debug|Win32 19 | {E4493801-7A78-47C0-97BA-1D7F515AA0A8}.Debug|x86.Build.0 = Debug|Win32 20 | {E4493801-7A78-47C0-97BA-1D7F515AA0A8}.Release|x64.ActiveCfg = Release|x64 21 | {E4493801-7A78-47C0-97BA-1D7F515AA0A8}.Release|x64.Build.0 = Release|x64 22 | {E4493801-7A78-47C0-97BA-1D7F515AA0A8}.Release|x86.ActiveCfg = Release|Win32 23 | {E4493801-7A78-47C0-97BA-1D7F515AA0A8}.Release|x86.Build.0 = Release|Win32 24 | EndGlobalSection 25 | GlobalSection(SolutionProperties) = preSolution 26 | HideSolutionNode = FALSE 27 | EndGlobalSection 28 | GlobalSection(ExtensibilityGlobals) = postSolution 29 | SolutionGuid = {E7E7F349-A1C6-4EA6-9215-66B08186089C} 30 | EndGlobalSection 31 | EndGlobal 32 | -------------------------------------------------------------------------------- /windows_threads/windows_threads/windows_threads.cpp: -------------------------------------------------------------------------------- 1 | // This program shows off the basics of windows threads in C++ 2 | // By: Nick from CoffeeBeforeArch 3 | 4 | #include 5 | #include 6 | 7 | #define NUM_THREADS 10 8 | #define BUF_SIZE 255 9 | 10 | // Argument we will be passing to out spawned threads 11 | struct Package { 12 | int tid; 13 | int number; 14 | }; 15 | 16 | void display_message(HANDLE hScreen, int tid, int number, const char *message) { 17 | // A character array 18 | TCHAR message_buffer[BUF_SIZE]; 19 | size_t cchStringSize; 20 | // DWORD is a Windows-specific type (Unsigned 32-bit) 21 | DWORD dwChars; 22 | 23 | // Print using thread-safe functions 24 | // Print a string to the buffer and checks to make sure it doesn't exceed 25 | // length 26 | StringCchPrintf(message_buffer, BUF_SIZE, 27 | TEXT("Printing number %d from thread %d. %s\n"), tid, number, 28 | message); 29 | // Check for a buffer overrun 30 | StringCchLength(message_buffer, BUF_SIZE, &cchStringSize); 31 | // Write character string to console 32 | WriteConsole(hScreen, message_buffer, cchStringSize, &dwChars, NULL); 33 | 34 | // Just put the thread to sleep for a little 35 | Sleep(1000); 36 | } 37 | 38 | // WINAPI is a macro that specifies the use of the Windows calling 39 | // convention 40 | // LPVOID is just a void pointer 41 | DWORD WINAPI boring_thread_function(LPVOID lpParam) { 42 | // Cast our void pointer to our defined struct 43 | Package local_package = *(Package *)lpParam; 44 | 45 | // Get Handle to the screen 46 | HANDLE hStdout = NULL; 47 | // Get handle to stdout 48 | hStdout = GetStdHandle(STD_OUTPUT_HANDLE); 49 | 50 | // Call our function with thread-safe print 51 | display_message(hStdout, local_package.tid, local_package.number, 52 | "Hello, there!"); 53 | 54 | return 0; 55 | } 56 | 57 | int main() { 58 | // Create an array of handles 59 | HANDLE array_of_handles[NUM_THREADS] = {0}; 60 | // Create an array of structs as arguments 61 | Package packages[NUM_THREADS]; 62 | 63 | // Create threads 64 | for (int i = 0; i < NUM_THREADS; i++) { 65 | packages[i].tid = i; 66 | packages[i].number = i * i; 67 | /* 68 | Arguments: 69 | 0.) Security Atrributes (default) 70 | 1.) Stack Size (default) 71 | 2.) Function 72 | 3.) Argument (converted to void pointer) 73 | 4.) Creation flag (0 = run immedately after creation) 74 | 5.) Pointer to variable to receive thread id 75 | */ 76 | array_of_handles[i] = 77 | CreateThread(NULL, 0, boring_thread_function, &packages[i], 0, NULL); 78 | } 79 | 80 | // Wait for a number of threads to be in the signaled state 81 | // Infinite states to wait forever 82 | WaitForMultipleObjects(NUM_THREADS, array_of_handles, true, INFINITE); 83 | 84 | // Clean up by closing open object handles 85 | for (int i = 0; i < NUM_THREADS; i++) { 86 | CloseHandle(array_of_handles[i]); 87 | } 88 | 89 | return 0; 90 | } 91 | -------------------------------------------------------------------------------- /windows_threads/windows_threads/windows_threads.vcxproj: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | Debug 6 | Win32 7 | 8 | 9 | Release 10 | Win32 11 | 12 | 13 | Debug 14 | x64 15 | 16 | 17 | Release 18 | x64 19 | 20 | 21 | 22 | 15.0 23 | {E4493801-7A78-47C0-97BA-1D7F515AA0A8} 24 | windowsthreads 25 | 10.0.17763.0 26 | 27 | 28 | 29 | Application 30 | true 31 | v141 32 | MultiByte 33 | 34 | 35 | Application 36 | false 37 | v141 38 | true 39 | MultiByte 40 | 41 | 42 | Application 43 | true 44 | v141 45 | MultiByte 46 | 47 | 48 | Application 49 | false 50 | v141 51 | true 52 | MultiByte 53 | 54 | 55 | 56 | 57 | 58 | 59 | 60 | 61 | 62 | 63 | 64 | 65 | 66 | 67 | 68 | 69 | 70 | 71 | 72 | 73 | 74 | 75 | Level3 76 | Disabled 77 | true 78 | true 79 | 80 | 81 | 82 | 83 | Level3 84 | Disabled 85 | true 86 | true 87 | 88 | 89 | 90 | 91 | Level3 92 | MaxSpeed 93 | true 94 | true 95 | true 96 | true 97 | 98 | 99 | true 100 | true 101 | 102 | 103 | 104 | 105 | Level3 106 | MaxSpeed 107 | true 108 | true 109 | true 110 | true 111 | 112 | 113 | true 114 | true 115 | 116 | 117 | 118 | 119 | 120 | 121 | 122 | 123 | -------------------------------------------------------------------------------- /windows_threads/windows_threads/windows_threads.vcxproj.filters: -------------------------------------------------------------------------------- 1 |  2 | 3 | 4 | 5 | {4FC737F1-C7A5-4376-A066-2A32D752A2FF} 6 | cpp;c;cc;cxx;def;odl;idl;hpj;bat;asm;asmx 7 | 8 | 9 | {93995380-89BD-4b04-88EB-625FBE52EBFB} 10 | h;hh;hpp;hxx;hm;inl;inc;ipp;xsd 11 | 12 | 13 | {67DA6AB6-F800-4c08-8B7A-83BB121AAD01} 14 | rc;ico;cur;bmp;dlg;rc2;rct;bin;rgs;gif;jpg;jpeg;jpe;resx;tiff;tif;png;wav;mfcribbon-ms 15 | 16 | 17 | 18 | 19 | Source Files 20 | 21 | 22 | --------------------------------------------------------------------------------