├── .gitignore
├── .vscode
└── settings.json
├── CMakeLists.txt
├── LICENSE
├── README.md
├── benchtest
├── CMakeLists.txt
└── main.cpp
├── src
├── CMakeLists.txt
├── Chunk.cpp
├── Chunk.hpp
├── ChunkAllocator.cpp
├── ChunkAllocator.hpp
├── Column.cpp
├── Column.hpp
├── Table.cpp
├── Table.hpp
├── main.cpp
└── main.hpp
└── tests
├── CMakeLists.txt
├── FileCreationTest.cpp
├── TableFindRows.cpp
├── TableReadWriteRowsTest.cpp
├── TableReadWriteTest.cpp
└── main.cpp
/.gitignore:
--------------------------------------------------------------------------------
1 | # Prerequisites
2 | *.d
3 |
4 | # Compiled Object files
5 | *.slo
6 | *.lo
7 | *.o
8 | *.obj
9 |
10 | # Precompiled Headers
11 | *.gch
12 | *.pch
13 |
14 | # Compiled Dynamic libraries
15 | *.so
16 | *.dylib
17 | *.dll
18 |
19 | # Fortran module files
20 | *.mod
21 | *.smod
22 |
23 | # Compiled Static libraries
24 | *.lai
25 | *.la
26 | *.a
27 | *.lib
28 |
29 | # Executables
30 | *.exe
31 | *.out
32 | *.app
33 |
34 | build/**
35 | googletest/
36 | benchmark/
37 | liburing/
38 |
--------------------------------------------------------------------------------
/.vscode/settings.json:
--------------------------------------------------------------------------------
1 | {
2 | "files.associations": {
3 | "iostream": "cpp",
4 | "any": "cpp",
5 | "array": "cpp",
6 | "atomic": "cpp",
7 | "bit": "cpp",
8 | "*.tcc": "cpp",
9 | "bitset": "cpp",
10 | "cctype": "cpp",
11 | "cfenv": "cpp",
12 | "charconv": "cpp",
13 | "chrono": "cpp",
14 | "cinttypes": "cpp",
15 | "clocale": "cpp",
16 | "cmath": "cpp",
17 | "codecvt": "cpp",
18 | "compare": "cpp",
19 | "complex": "cpp",
20 | "concepts": "cpp",
21 | "condition_variable": "cpp",
22 | "csetjmp": "cpp",
23 | "csignal": "cpp",
24 | "cstdarg": "cpp",
25 | "cstddef": "cpp",
26 | "cstdint": "cpp",
27 | "cstdio": "cpp",
28 | "cstdlib": "cpp",
29 | "cstring": "cpp",
30 | "ctime": "cpp",
31 | "cuchar": "cpp",
32 | "cwchar": "cpp",
33 | "cwctype": "cpp",
34 | "deque": "cpp",
35 | "forward_list": "cpp",
36 | "list": "cpp",
37 | "map": "cpp",
38 | "set": "cpp",
39 | "unordered_map": "cpp",
40 | "unordered_set": "cpp",
41 | "vector": "cpp",
42 | "exception": "cpp",
43 | "algorithm": "cpp",
44 | "functional": "cpp",
45 | "iterator": "cpp",
46 | "memory": "cpp",
47 | "memory_resource": "cpp",
48 | "numeric": "cpp",
49 | "optional": "cpp",
50 | "random": "cpp",
51 | "ratio": "cpp",
52 | "regex": "cpp",
53 | "string": "cpp",
54 | "string_view": "cpp",
55 | "system_error": "cpp",
56 | "tuple": "cpp",
57 | "type_traits": "cpp",
58 | "utility": "cpp",
59 | "fstream": "cpp",
60 | "future": "cpp",
61 | "initializer_list": "cpp",
62 | "iomanip": "cpp",
63 | "iosfwd": "cpp",
64 | "istream": "cpp",
65 | "limits": "cpp",
66 | "mutex": "cpp",
67 | "new": "cpp",
68 | "numbers": "cpp",
69 | "ostream": "cpp",
70 | "ranges": "cpp",
71 | "scoped_allocator": "cpp",
72 | "shared_mutex": "cpp",
73 | "span": "cpp",
74 | "sstream": "cpp",
75 | "stdexcept": "cpp",
76 | "stop_token": "cpp",
77 | "streambuf": "cpp",
78 | "thread": "cpp",
79 | "typeindex": "cpp",
80 | "typeinfo": "cpp",
81 | "valarray": "cpp",
82 | "variant": "cpp",
83 | "cassert": "cpp",
84 | "cerrno": "cpp"
85 | }
86 | }
--------------------------------------------------------------------------------
/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | cmake_minimum_required(VERSION 3.10)
2 | project(ssd-nvme-database)
3 |
4 | set(CMAKE_CXX_STANDARD 17)
5 |
6 | include_directories(src)
7 |
8 | add_subdirectory(src)
9 | add_subdirectory(tests)
10 | add_subdirectory(googletest)
11 | add_subdirectory(benchmark)
12 | add_subdirectory(benchtest)
--------------------------------------------------------------------------------
/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 | # ssd-nvme-database
2 |
3 | Columnar database on SSD NVMe
4 |
5 | ## Building
6 |
7 | In the root of the project directory, run:
8 |
9 | ```bash
10 | git clone https://github.com/google/googletest.git
11 | git clone https://github.com/google/benchmark.git
12 | git clone https://github.com/axboe/liburing.git
13 | cd liburing
14 | sudo make install
15 | cd ..
16 | mkdir build
17 | cd build
18 | cmake ..
19 | make
20 | ```
21 |
--------------------------------------------------------------------------------
/benchtest/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | set(BINARY ${CMAKE_PROJECT_NAME}_bench)
2 |
3 | file(GLOB_RECURSE BENCH_SOURCES LIST_DIRECTORIES false *.h *.cpp)
4 |
5 | set(SOURCES ${BENCH_SOURCES})
6 |
7 | add_executable(${BINARY} ${BENCH_SOURCES})
8 |
9 | target_link_libraries(${BINARY} PUBLIC ${CMAKE_PROJECT_NAME}_lib benchmark::benchmark)
--------------------------------------------------------------------------------
/benchtest/main.cpp:
--------------------------------------------------------------------------------
1 | #include
2 | #include
3 |
4 | #include "benchmark/benchmark.h"
5 | #include "Table.hpp"
6 |
7 | static void BM_TableWrite(benchmark::State &state)
8 | {
9 | const int table_size = 16 * 16;
10 | Table table{1, 1};
11 | for (auto _ : state)
12 | {
13 | for (int i = 0; i < table_size; i++)
14 | {
15 | table.writeInt(i, 0, i);
16 | table.writeDouble(i, 1, i + 0.1);
17 | }
18 | }
19 | }
20 |
21 | static void BM_TableWriteMultiple(benchmark::State &state)
22 | {
23 | const int table_size = 16 * 16;
24 | Table table{1, 1};
25 | unsigned int attributes[table_size];
26 | double values[table_size];
27 |
28 | for (int i = 0; i < table_size; i++)
29 | {
30 | attributes[i] = i;
31 | values[i] = 0.1 * i;
32 | }
33 | for (auto _ : state)
34 | {
35 | table.writeRows(0, table_size, attributes, values);
36 | }
37 | }
38 |
39 | static void BM_TableRead(benchmark::State &state)
40 | {
41 | const int table_size = 16 * 16;
42 | Table table{1, 0};
43 | unsigned int attributes[table_size];
44 | for (int i = 0; i < table_size; i++)
45 | {
46 | attributes[i] = i;
47 | }
48 | table.writeRows(0, table_size, attributes, nullptr);
49 |
50 | unsigned int i = 0;
51 | for (auto _ : state)
52 | {
53 | table.readInt(i, 0);
54 | i = (i + 1) % table_size;
55 | }
56 | }
57 |
58 | static void BM_TableSearch(benchmark::State &state)
59 | {
60 | const int table_size = 16 * 16;
61 | Table table{1, 0};
62 | unsigned int attributes[table_size];
63 | for (int i = 0; i < table_size; i++)
64 | {
65 | attributes[i] = i;
66 | }
67 | table.writeRows(0, table_size, attributes, nullptr);
68 |
69 | for (auto _ : state)
70 | {
71 | int attribute_predicates[1] = {2};
72 | auto result = table.findRows(attribute_predicates, nullptr, false, false, false);
73 | };
74 | }
75 |
76 | // Test load / comparison / unload
77 | static void BM_ChunkLoad(benchmark::State &state)
78 | {
79 | const int column_size = CHUNK_SIZE;
80 | ChunkAllocator chunk_allocator;
81 | Chunk chunk{chunk_allocator.getChunk(false)};
82 |
83 | unsigned int attributes[column_size];
84 | for (int i = 0; i < column_size; i++)
85 | {
86 | attributes[i] = i;
87 | }
88 | chunk.writeInts(0, column_size, attributes);
89 |
90 | for (auto _ : state)
91 | {
92 | auto start = std::chrono::high_resolution_clock::now();
93 | // Just counting the loading
94 | chunk.load();
95 | auto end = std::chrono::high_resolution_clock::now();
96 | chunk.unload();
97 |
98 | auto elapsed_seconds =
99 | std::chrono::duration_cast>(
100 | end - start);
101 |
102 | state.SetIterationTime(elapsed_seconds.count());
103 | }
104 | }
105 |
106 | static void BM_ChunkComparison(benchmark::State &state)
107 | {
108 | const int column_size = CHUNK_SIZE;
109 | ChunkAllocator chunk_allocator;
110 | Chunk chunk{chunk_allocator.getChunk(false)};
111 |
112 | unsigned int attributes[column_size];
113 | for (int i = 0; i < column_size; i++)
114 | {
115 | attributes[i] = i;
116 | }
117 | chunk.writeInts(0, column_size, attributes);
118 |
119 | for (auto _ : state)
120 | {
121 | chunk.load();
122 | auto start = std::chrono::high_resolution_clock::now();
123 | // Just counting the comparison
124 | chunk.findInt(1);
125 | auto end = std::chrono::high_resolution_clock::now();
126 | chunk.unload();
127 |
128 | auto elapsed_seconds =
129 | std::chrono::duration_cast>(
130 | end - start);
131 |
132 | state.SetIterationTime(elapsed_seconds.count());
133 | }
134 | }
135 |
136 | static void BM_BitsetTest(benchmark::State &state)
137 | {
138 | const int column_size = CHUNK_SIZE;
139 | unsigned int attributes[column_size];
140 | std::bitset results;
141 | for (int i = 0; i < column_size; i++)
142 | {
143 | attributes[i] = i;
144 | }
145 |
146 | for (auto _ : state)
147 | {
148 | for (int i = 0; i < column_size; i++)
149 | {
150 | results[i] = (attributes[i] == 1);
151 | }
152 | }
153 | if (results.count() != 1)
154 | {
155 | std::cout << "should not be printed" << std::endl;
156 | }
157 | }
158 |
159 | BENCHMARK(BM_TableWrite);
160 | BENCHMARK(BM_TableWriteMultiple);
161 | BENCHMARK(BM_TableRead);
162 | BENCHMARK(BM_TableSearch);
163 | BENCHMARK(BM_ChunkLoad)->UseManualTime();
164 | BENCHMARK(BM_ChunkComparison)->UseManualTime();
165 | BENCHMARK(BM_BitsetTest);
166 | BENCHMARK_MAIN();
167 |
--------------------------------------------------------------------------------
/src/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | set(BINARY ${CMAKE_PROJECT_NAME})
2 |
3 | file(GLOB_RECURSE SOURCES LIST_DIRECTORIES true *.h *.hpp *.cpp)
4 |
5 | set(SOURCES ${SOURCES})
6 |
7 | add_executable(${BINARY}_run ${SOURCES})
8 |
9 | add_library(${BINARY}_lib STATIC ${SOURCES})
10 |
11 | find_package(OpenMP REQUIRED)
12 |
13 | target_link_libraries(${BINARY}_lib uring OpenMP::OpenMP_CXX)
14 | target_link_libraries(${BINARY}_run uring OpenMP::OpenMP_CXX)
--------------------------------------------------------------------------------
/src/Chunk.cpp:
--------------------------------------------------------------------------------
1 | #include "Chunk.hpp"
2 |
3 | #include
4 | #include
5 | #include
6 | #include
7 |
8 | Chunk::Chunk(int fd, unsigned long int starting_pos, bool is_double)
9 | : fd(fd),
10 | starting_pos(starting_pos),
11 | is_double(is_double)
12 | {
13 | int number_written = -1;
14 | if (is_double)
15 | {
16 | doubleCache.reset(nullptr);
17 | double zeroes[CHUNK_SIZE] = {};
18 | number_written = pwrite(fd, &zeroes, CHUNK_SIZE * sizeof(double), starting_pos);
19 | }
20 | else
21 | {
22 | intCache.reset(nullptr);
23 | int zeroes[CHUNK_SIZE] = {};
24 | number_written = pwrite(fd, &zeroes, CHUNK_SIZE * sizeof(int), starting_pos);
25 | }
26 |
27 | if (number_written == -1)
28 | {
29 | std::cout << "Error creating a new chunk starting at " << starting_pos << ": " << std::strerror(errno) << std::endl;
30 | }
31 | }
32 |
33 | Chunk::Chunk(Chunk &&other)
34 | : fd(other.fd),
35 | starting_pos(other.starting_pos),
36 | is_double(other.is_double),
37 | nb_element(other.nb_element),
38 | intCache(std::move(other.intCache)),
39 | doubleCache(std::move(other.doubleCache))
40 | {
41 | }
42 |
43 | Chunk &Chunk::operator=(Chunk &&other)
44 | {
45 | fd = other.fd;
46 | starting_pos = other.starting_pos;
47 | is_double = other.is_double;
48 | nb_element = other.nb_element;
49 | intCache = std::move(other.intCache);
50 | doubleCache = std::move(other.doubleCache);
51 |
52 | return *this;
53 | }
54 |
55 | void Chunk::load()
56 | {
57 | int number_written = -1;
58 | if (is_double)
59 | {
60 | doubleCache.reset(new double[CHUNK_SIZE]);
61 | number_written = pread(fd, doubleCache.get(), CHUNK_SIZE * sizeof(double), starting_pos);
62 | }
63 | else
64 | {
65 | intCache.reset(new unsigned int[CHUNK_SIZE]);
66 | number_written = pread(fd, intCache.get(), CHUNK_SIZE * sizeof(int), starting_pos);
67 | }
68 |
69 | if (number_written == -1)
70 | {
71 | std::cout << "Error loading a chunk starting at " << starting_pos << ": " << std::strerror(errno) << std::endl;
72 | }
73 | }
74 |
75 | void Chunk::aload(io_uring *ring, unsigned long int chunk_no)
76 | {
77 | struct io_uring_sqe *sqe;
78 | struct io_data *data = (io_data *)malloc(sizeof(*data));
79 |
80 | sqe = io_uring_get_sqe(ring);
81 | if (is_double)
82 | {
83 | doubleCache.reset(new double[CHUNK_SIZE]);
84 | io_uring_prep_read(sqe, fd, doubleCache.get(), CHUNK_SIZE * sizeof(double), starting_pos);
85 | }
86 | else
87 | {
88 | intCache.reset(new unsigned int[CHUNK_SIZE]);
89 | io_uring_prep_read(sqe, fd, intCache.get(), CHUNK_SIZE * sizeof(int), starting_pos);
90 | }
91 | data->chunk_no = chunk_no;
92 | io_uring_sqe_set_data(sqe, data);
93 | io_uring_submit(ring);
94 | }
95 |
96 | void Chunk::unload()
97 | {
98 | if (is_double)
99 | {
100 | doubleCache.reset(nullptr);
101 | }
102 | else
103 | {
104 | intCache.reset(nullptr);
105 | }
106 | }
107 |
108 | unsigned int Chunk::readInt(unsigned long int chunk_pos) const
109 | {
110 | return intCache[chunk_pos];
111 | }
112 |
113 | double Chunk::readDouble(unsigned long int chunk_pos) const
114 | {
115 | return doubleCache[chunk_pos];
116 | }
117 |
118 | std::bitset Chunk::findInt(unsigned int predicate) const
119 | {
120 | std::bitset result;
121 | unsigned int *intContent = intCache.get();
122 | for (unsigned long int i = 0; i < CHUNK_SIZE; i++)
123 | {
124 | if (predicate == intContent[i])
125 | {
126 | result.set(i);
127 | }
128 | }
129 | return result;
130 | }
131 |
132 | std::bitset Chunk::findDouble(double predicate) const
133 | {
134 | std::bitset result;
135 | double *doubleContent = doubleCache.get();
136 | for (unsigned long int i = 0; i < CHUNK_SIZE; i++)
137 | {
138 | if (predicate == doubleContent[i])
139 | {
140 | result.set(i);
141 | }
142 | }
143 | return result;
144 | }
145 |
146 | void Chunk::writeInts(unsigned int starting_chunk_pos, unsigned int number_of_values, unsigned int *attributes) const
147 | {
148 | unsigned long int file_pos = starting_pos + starting_chunk_pos * sizeof(int);
149 | int number_written = pwrite(fd, attributes, number_of_values * sizeof(int), file_pos);
150 | if (number_written == -1)
151 | {
152 | std::cout << "Error writing on chunk starting at " << starting_pos << ": " << std::strerror(errno) << std::endl;
153 | }
154 | }
155 |
156 | void Chunk::writeDoubles(unsigned int starting_chunk_pos, unsigned int number_of_values, double *values) const
157 | {
158 | unsigned long int file_pos = starting_pos + starting_chunk_pos * sizeof(double);
159 | int number_written = pwrite(fd, values, number_of_values * sizeof(double), file_pos);
160 | if (number_written == -1)
161 | {
162 | std::cout << "Error writing on chunk starting at " << starting_pos << ": " << std::strerror(errno) << std::endl;
163 | }
164 | }
--------------------------------------------------------------------------------
/src/Chunk.hpp:
--------------------------------------------------------------------------------
1 | #ifndef CHUNK_HPP
2 | #define CHUNK_HPP
3 |
4 | #include
5 | #include
6 | #include
7 | #include
8 | #include
9 |
10 | #include "main.hpp"
11 |
12 | struct io_data
13 | {
14 | unsigned long int chunk_no;
15 | };
16 |
17 | class Chunk
18 | {
19 | int fd; // File descriptor where data is stored
20 | unsigned long int starting_pos; // Starting position in file
21 | bool is_double; // Is storing int or double
22 | unsigned int nb_element; // Current number of element in the chunk
23 | std::unique_ptr intCache; // Cache for reading data (could be nullptr when data is not loaded)
24 | std::unique_ptr doubleCache; // Same for double
25 |
26 | public:
27 | Chunk(int fd, unsigned long int starting_pos, bool is_double);
28 | Chunk(Chunk &&);
29 | Chunk &operator=(Chunk &&);
30 | ~Chunk() = default;
31 |
32 | void load(); // Load the content
33 | void aload(io_uring *ring, unsigned long int chunk_no); // Asynchronous load()
34 | void unload(); // Unload the content
35 | unsigned int readInt(unsigned long int chunk_pos) const;
36 | double readDouble(unsigned long int chunk_pos) const;
37 | std::bitset findInt(unsigned int predicate) const;
38 | std::bitset findDouble(double predicate) const;
39 | void writeInts(unsigned starting_chunk_pos, unsigned int number_of_values, unsigned int *attributes) const;
40 | void writeDoubles(unsigned starting_chunk_pos, unsigned int number_of_values, double *values) const;
41 | };
42 |
43 | #endif
44 |
--------------------------------------------------------------------------------
/src/ChunkAllocator.cpp:
--------------------------------------------------------------------------------
1 | #include
2 | #include
3 | #include
4 | #include
5 |
6 | #include "ChunkAllocator.hpp"
7 |
8 | ChunkAllocator::ChunkAllocator()
9 | {
10 | char nameBuff[32];
11 | memset(nameBuff, 0, sizeof(nameBuff));
12 | strncpy(nameBuff, "/tmp/ssd-bdd-XXXXXX", 21);
13 | fd = mkstemp(nameBuff);
14 | unlink(nameBuff);
15 | last_line = 0;
16 | }
17 |
18 | Chunk ChunkAllocator::getChunk(bool is_double)
19 | {
20 | Chunk new_chunk{fd, last_line, is_double};
21 | last_line += CHUNK_SIZE * (is_double ? sizeof(double) : sizeof(int));
22 | return new_chunk;
23 | }
--------------------------------------------------------------------------------
/src/ChunkAllocator.hpp:
--------------------------------------------------------------------------------
1 | #ifndef CHUNK_ALLOCATOR_HPP
2 | #define CHUNK_ALLOCATOR_HPP
3 |
4 | #include "main.hpp"
5 | #include "Chunk.hpp"
6 |
7 | class ChunkAllocator
8 | {
9 | int fd; // File descriptor where data is stored
10 | unsigned long int last_line;
11 |
12 | public:
13 | ChunkAllocator();
14 | ~ChunkAllocator() = default;
15 |
16 | Chunk getChunk(bool is_double);
17 | };
18 |
19 | #endif
--------------------------------------------------------------------------------
/src/Column.cpp:
--------------------------------------------------------------------------------
1 | #include
2 | #include
3 |
4 | #include "Column.hpp"
5 |
6 | Column::Column(ChunkAllocator &chunk_allocator, bool is_double)
7 | : chunk_allocator(chunk_allocator),
8 | is_double(is_double)
9 | {
10 | int ret = io_uring_queue_init(QUEUE_DEPTH, &_ring, 0);
11 | if (ret < 0)
12 | {
13 | std::cout << "Issue on queue init: " << std::strerror(-ret) << std::endl;
14 | }
15 | }
16 |
17 | unsigned int Column::readInt(unsigned long int row_pos)
18 | {
19 | unsigned long int chunk_no = row_pos / CHUNK_SIZE;
20 | unsigned long int chunk_pos = row_pos % CHUNK_SIZE;
21 | chunks[chunk_no].load();
22 | int value = chunks[chunk_no].readInt(chunk_pos);
23 | chunks[chunk_no].unload();
24 | return value;
25 | }
26 |
27 | double Column::readDouble(unsigned long int row_pos)
28 | {
29 | long unsigned long int chunk_no = row_pos / CHUNK_SIZE;
30 | unsigned long int chunk_pos = row_pos % CHUNK_SIZE;
31 | chunks[chunk_no].load();
32 | double value = chunks[chunk_no].readDouble(chunk_pos);
33 | chunks[chunk_no].unload();
34 | return value;
35 | }
36 |
37 | void Column::writeInts(unsigned long int starting_row_pos, unsigned long int number_of_rows, unsigned int *attributes)
38 | {
39 | unsigned long int starting_chunk_pos = starting_row_pos;
40 | while (starting_chunk_pos - starting_row_pos < number_of_rows)
41 | {
42 | unsigned int number_of_values = CHUNK_SIZE - starting_chunk_pos % CHUNK_SIZE;
43 | unsigned long int remaining_number_of_values = number_of_rows - (starting_chunk_pos - starting_row_pos);
44 | if (number_of_values > remaining_number_of_values)
45 | {
46 | number_of_values = remaining_number_of_values;
47 | }
48 | unsigned int chunk_no = starting_chunk_pos / CHUNK_SIZE;
49 | if (chunks.size() <= chunk_no)
50 | {
51 | chunks.reserve(chunk_no);
52 | }
53 | while (chunks.size() <= chunk_no)
54 | {
55 | chunks.push_back(chunk_allocator.getChunk(false));
56 | }
57 | chunks[chunk_no].writeInts(
58 | starting_chunk_pos % CHUNK_SIZE,
59 | number_of_values,
60 | attributes + starting_chunk_pos - starting_row_pos);
61 | starting_chunk_pos += number_of_values;
62 | }
63 | }
64 |
65 | void Column::writeDoubles(
66 | unsigned long int starting_row_pos,
67 | unsigned long int number_of_rows,
68 | double *values)
69 | {
70 | unsigned long int starting_chunk_pos = starting_row_pos;
71 | while (starting_chunk_pos - starting_row_pos < number_of_rows)
72 | {
73 | unsigned int number_of_values = CHUNK_SIZE - starting_chunk_pos % CHUNK_SIZE;
74 | unsigned long int remaining_number_of_values = number_of_rows - (starting_chunk_pos - starting_row_pos);
75 | if (number_of_values > remaining_number_of_values)
76 | {
77 | number_of_values = remaining_number_of_values;
78 | }
79 | unsigned int chunk_no = starting_chunk_pos / CHUNK_SIZE;
80 | if (chunks.size() <= chunk_no)
81 | {
82 | chunks.reserve(chunk_no);
83 | }
84 | while (chunks.size() < chunk_no + 1)
85 | {
86 | chunks.push_back(chunk_allocator.getChunk(true));
87 | }
88 | chunks[chunk_no].writeDoubles(starting_chunk_pos % CHUNK_SIZE, number_of_values, values + starting_chunk_pos - starting_row_pos);
89 | starting_chunk_pos += number_of_values;
90 | }
91 | }
92 |
93 | unsigned long int Column::collectOne(bool is_blocking)
94 | {
95 | struct io_uring_cqe *cqe;
96 |
97 | io_uring_wait_cqe(&_ring, &cqe);
98 | // if (cqe->res < 0)
99 | // {
100 | // std::cout << "IO issue: " << std::strerror(-cqe->res) << std::endl;
101 | // }
102 |
103 | struct io_data *data = (io_data *)io_uring_cqe_get_data(cqe);
104 | io_uring_cqe_seen(&_ring, cqe);
105 |
106 | unsigned long int chunk_no = data->chunk_no;
107 | free(data);
108 |
109 | return chunk_no;
110 | }
111 |
112 | std::vector> Column::findIntRowsSync(int predicate, bool without_loading, bool with_openmp)
113 | {
114 | const unsigned long int number_of_chunks = chunks.size();
115 | std::vector> result = std::vector>();
116 | result.reserve(number_of_chunks);
117 | #pragma omp parallel for schedule(static)
118 | for (unsigned long int i = 0; i < number_of_chunks; i++)
119 | {
120 | if (without_loading)
121 | {
122 | result[i] = chunks[i].findInt(predicate);
123 | }
124 | else
125 | {
126 | chunks[i].load();
127 | result[i] = chunks[i].findInt(predicate);
128 | chunks[i].unload();
129 | }
130 | }
131 | return result;
132 | }
133 |
134 | std::vector> Column::findIntRowsAsync(int predicate)
135 | {
136 | const unsigned long int number_of_chunks = chunks.size();
137 | std::vector> result = std::vector>();
138 | result.reserve(number_of_chunks);
139 | unsigned int current_queue_length = 0;
140 | unsigned long int i = 0;
141 |
142 | while ((i < number_of_chunks) || (current_queue_length > 0))
143 | {
144 | while ((current_queue_length < QUEUE_DEPTH) && (i < number_of_chunks))
145 | {
146 | chunks[i].aload(&_ring, i);
147 | current_queue_length++;
148 | i++;
149 | }
150 | unsigned long int chunk_no = collectOne(true);
151 | result[chunk_no] = chunks[chunk_no].findInt(predicate);
152 | chunks[chunk_no].unload();
153 | current_queue_length--;
154 | }
155 | return result;
156 | }
157 |
158 | std::vector> Column::findIntRows(int predicate, bool use_async, bool without_loading, bool with_openmp)
159 | {
160 | if (use_async)
161 | {
162 | return findIntRowsAsync(predicate);
163 | }
164 | return findIntRowsSync(predicate, without_loading, with_openmp);
165 | }
166 |
167 | std::vector> Column::findDoubleRowsSync(double predicate, bool without_loading, bool with_openmp)
168 | {
169 | const unsigned long int number_of_chunks = chunks.size();
170 | std::vector> result = std::vector>();
171 | result.reserve(number_of_chunks);
172 | #pragma omp parallel for schedule(static) if (with_openmp)
173 | for (unsigned long int i = 0; i < number_of_chunks; i++)
174 | {
175 | if (without_loading)
176 | {
177 | result[i] = chunks[i].findDouble(predicate);
178 | }
179 | else
180 | {
181 | chunks[i].load();
182 | result[i] = chunks[i].findDouble(predicate);
183 | chunks[i].unload();
184 | }
185 | }
186 | return result;
187 | }
188 |
189 | std::vector> Column::findDoubleRowsAsync(double predicate)
190 | {
191 | const unsigned long int number_of_chunks = chunks.size();
192 | std::vector> result = std::vector>();
193 | result.reserve(number_of_chunks);
194 | unsigned int current_queue_length = 0;
195 | unsigned long int i = 0;
196 |
197 | while ((i < number_of_chunks) || (current_queue_length > 0))
198 | {
199 | while ((current_queue_length < QUEUE_DEPTH) && (i < number_of_chunks))
200 | {
201 | chunks[i].aload(&_ring, i);
202 | current_queue_length++;
203 | i++;
204 | }
205 | unsigned long int chunk_no = collectOne(true);
206 | result[chunk_no] = chunks[chunk_no].findDouble(predicate);
207 | chunks[chunk_no].unload();
208 | current_queue_length--;
209 | }
210 | return result;
211 | }
212 |
213 | std::vector> Column::findDoubleRows(double predicate, bool use_async, bool without_loading, bool with_openmp)
214 | {
215 | if (use_async)
216 | {
217 | return findDoubleRowsAsync(predicate);
218 | }
219 | return findDoubleRowsSync(predicate, without_loading, with_openmp);
220 | }
221 |
222 | void Column::loadEverything()
223 | {
224 | for (auto &chunk : chunks)
225 | {
226 | chunk.load();
227 | }
228 | }
229 |
230 | void Column::unloadEverything()
231 | {
232 | for (auto &chunk : chunks)
233 | {
234 | chunk.load();
235 | }
236 | }
--------------------------------------------------------------------------------
/src/Column.hpp:
--------------------------------------------------------------------------------
1 | #ifndef COLUMN_HPP
2 | #define COLUMN_HPP
3 |
4 | #include
5 | #include
6 | #include
7 | #include
8 |
9 | #include "main.hpp"
10 | #include "Chunk.hpp"
11 | #include "ChunkAllocator.hpp"
12 |
13 | // specify if double or single on creation
14 | class Column
15 | {
16 | ChunkAllocator &chunk_allocator; // Memory Allocator to generate new chunk
17 | bool is_double; // Integer or double column
18 | std::vector chunks; // Chunks constituting the column
19 | struct io_uring _ring; // IO_uring ring
20 |
21 | unsigned long int collectOne(bool is_blocking);
22 | std::vector> findIntRowsAsync(int predicate);
23 | std::vector> findIntRowsSync(int predicate, bool without_loading, bool with_openmp);
24 | std::vector> findDoubleRowsAsync(double predicate);
25 | std::vector> findDoubleRowsSync(double predicate, bool without_loading, bool with_openmp);
26 |
27 | public:
28 | Column(ChunkAllocator &chunk_allocator, bool is_double);
29 | Column(Column &&) = default;
30 | Column &operator=(Column &&) = default;
31 | ~Column() = default;
32 |
33 | unsigned int readInt(unsigned long int row_pos);
34 | double readDouble(unsigned long int row_pos);
35 | void writeInts(unsigned long int starting_row_pos, unsigned long int number_of_rows, unsigned int *attributes);
36 | void writeDoubles(unsigned long int starting_row_pos, unsigned long int number_of_rows, double *values);
37 |
38 | std::vector> findIntRows(int predicate, bool use_async, bool without_loading, bool with_openmp);
39 | std::vector> findDoubleRows(double predicate, bool use_async, bool without_loading, bool with_openmp);
40 |
41 | void loadEverything();
42 | void unloadEverything();
43 | };
44 |
45 | #endif
46 |
--------------------------------------------------------------------------------
/src/Table.cpp:
--------------------------------------------------------------------------------
1 | #include "Table.hpp"
2 |
3 | Table::Table(unsigned int attribute_count, unsigned int value_count)
4 | : attribute_count(attribute_count),
5 | value_count(value_count)
6 | {
7 | for (unsigned int i = 0; i < attribute_count + value_count; i++)
8 | {
9 | columns.emplace_back(chunk_allocator, i >= attribute_count);
10 | }
11 | };
12 |
13 | unsigned int Table::readInt(unsigned long int row_pos, unsigned int column)
14 | {
15 | return columns[column].readInt(row_pos);
16 | }
17 |
18 | double Table::readDouble(unsigned long int row_pos, unsigned int column)
19 | {
20 | return columns[column].readDouble(row_pos);
21 | }
22 |
23 | void Table::writeInt(unsigned long int row_pos, unsigned int column, unsigned int attribute)
24 | {
25 | columns[column].writeInts(row_pos, 1, &attribute);
26 | }
27 |
28 | void Table::writeDouble(unsigned long int row_pos, unsigned int column, double value)
29 | {
30 | columns[column].writeDoubles(row_pos, 1, &value);
31 | }
32 |
33 | void Table::writeRows(unsigned long int starting_row_pos, unsigned long int number_of_rows, unsigned int *attributes, double *values)
34 | {
35 | for (unsigned int column = 0; column < attribute_count + value_count; column++)
36 | {
37 | if (column < attribute_count)
38 | {
39 | columns[column].writeInts(starting_row_pos, number_of_rows, attributes + column * number_of_rows);
40 | }
41 | else
42 | {
43 | columns[column].writeDoubles(starting_row_pos, number_of_rows, values + (column - attribute_count) * number_of_rows);
44 | }
45 | }
46 | }
47 |
48 | void Table::writeRow(unsigned long int row_pos, unsigned int *attributes, double *values)
49 | {
50 | return writeRows(row_pos, 1, attributes, values);
51 | };
52 |
53 | std::vector> Table::findRows(int *attribute_predicates, double *value_predicates, bool use_async, bool without_loading, bool with_openmp)
54 | {
55 | std::vector> result;
56 | for (unsigned int i = 0; i < attribute_count; i++)
57 | {
58 | if (result.empty())
59 | {
60 | result = columns[i].findIntRows(attribute_predicates[i], use_async, without_loading, with_openmp);
61 | }
62 | else
63 | {
64 | std::vector> partial_result{columns[i].findIntRows(attribute_predicates[i], use_async, without_loading, with_openmp)};
65 | for (long unsigned int j = 0; j < result.size(); j++)
66 | {
67 | result[j] &= partial_result[j];
68 | }
69 | }
70 | }
71 |
72 | for (unsigned int i = 0; i < value_count; i++)
73 | {
74 | if (result.empty())
75 | {
76 | result = columns[attribute_count + i].findDoubleRows(value_predicates[i], use_async, without_loading, with_openmp);
77 | }
78 | else
79 | {
80 | std::vector> partial_result{columns[attribute_count + i].findDoubleRows(value_predicates[i], use_async, without_loading, with_openmp)};
81 | for (long unsigned int j = 0; j < result.size(); j++)
82 | {
83 | result[j] &= partial_result[j];
84 | }
85 | }
86 | }
87 |
88 | return result;
89 | }
90 |
91 | void Table::loadEverything()
92 | {
93 | for (auto &column : columns)
94 | {
95 | column.loadEverything();
96 | }
97 | }
98 |
99 | void Table::unloadEverything()
100 | {
101 | for (auto &column : columns)
102 | {
103 | column.unloadEverything();
104 | }
105 | }
106 |
--------------------------------------------------------------------------------
/src/Table.hpp:
--------------------------------------------------------------------------------
1 | #ifndef TABLE_HPP
2 | #define TABLE_HPP
3 |
4 | #include "Column.hpp"
5 | #include "ChunkAllocator.hpp"
6 |
7 | class Table
8 | {
9 | ChunkAllocator chunk_allocator; // Chunk Allocator to generate new chunk
10 | unsigned int attribute_count; // Number of int columns
11 | unsigned int value_count; // Number of double columns
12 | std::vector columns; // List of table's column
13 |
14 | public:
15 | Table(unsigned int attribute_count, unsigned int value_count);
16 | ~Table() = default;
17 |
18 | unsigned int readInt(unsigned long int row_pos, unsigned int column);
19 | double readDouble(unsigned long int row_pos, unsigned int column);
20 | void writeInt(unsigned long int row_pos, unsigned int column, unsigned int value);
21 | void writeDouble(unsigned long int row_pos, unsigned int column, double value);
22 | void writeRow(unsigned long int row_pos, unsigned int *attributes, double *values);
23 | void writeRows(unsigned long int starting_row_pos, unsigned long int number_of_rows, unsigned int *attributes, double *values);
24 |
25 | std::vector> findRows(int *attribute_predicates, double *value_predicates, bool use_async, bool without_loading, bool with_openmp);
26 |
27 | void loadEverything();
28 | void unloadEverything();
29 | };
30 |
31 | #endif
32 |
--------------------------------------------------------------------------------
/src/main.cpp:
--------------------------------------------------------------------------------
1 | #include
2 | #include
3 | #include
4 |
5 | #include "Table.hpp"
6 |
7 | int main(int argc, char **argv)
8 | {
9 | std::cout << "Number of thread available: " << omp_get_max_threads() << std::endl;
10 |
11 | const unsigned long int number_of_rows = 1024 * 1024 * 1024;
12 | const unsigned long int write_block_size = 128 * 1024;
13 | const unsigned int attributes_count = 2;
14 | const unsigned int values_count = 2;
15 |
16 | double gigabytes_handled = (double)number_of_rows * (attributes_count * sizeof(int) + values_count * sizeof(double)) / (1024 * 1024 * 1024);
17 | std::cout << "Number of gigabytes handled: " << gigabytes_handled << std::endl;
18 |
19 | Table table{attributes_count, values_count};
20 | auto start_write = std::chrono::high_resolution_clock::now();
21 | for (unsigned long int i = 0; i < number_of_rows; i += write_block_size)
22 | {
23 | unsigned int attributes[write_block_size * attributes_count];
24 | double values[write_block_size * values_count];
25 | for (unsigned long int j = 0; j < write_block_size; j++)
26 | {
27 | attributes[write_block_size * 0 + j] = i + j;
28 | attributes[write_block_size * 1 + j] = 2 * (i + j);
29 | values[write_block_size * 0 + j] = 0.1 + i + j;
30 | values[write_block_size * 1 + j] = 1.0 * (i + j);
31 | }
32 | table.writeRows(i, write_block_size, attributes, values);
33 | }
34 | auto end_write = std::chrono::high_resolution_clock::now();
35 | auto elapsed_seconds_write = std::chrono::duration_cast>(end_write - start_write);
36 | std::cout << "Written in: " << elapsed_seconds_write.count() << "s\t(" << gigabytes_handled / elapsed_seconds_write.count() << " Go/s)" << std::endl;
37 |
38 | const int row_to_find = 19542;
39 | int attribute_predicates[2] = {row_to_find, 2 * row_to_find};
40 | double value_predicates[2] = {0.1 + row_to_find, 1.0 * row_to_find};
41 |
42 | // auto start_find_sync = std::chrono::high_resolution_clock::now();
43 | // auto result_sync = table.findRows(attribute_predicates, value_predicates, false, false, false);
44 | // auto end_find_sync = std::chrono::high_resolution_clock::now();
45 | // // Checking solution
46 | // for (long unsigned int i = 0; i < result_sync.size(); i++)
47 | // {
48 | // for (long unsigned int j = 0; j < CHUNK_SIZE; j++)
49 | // {
50 | // if (result_sync[i][j] != (i * CHUNK_SIZE + j == row_to_find))
51 | // {
52 | // return 1;
53 | // }
54 | // }
55 | // }
56 | // auto elapsed_seconds_find_sync = std::chrono::duration_cast>(end_find_sync - start_find_sync);
57 | // std::cout << "Sync find in: " << elapsed_seconds_find_sync.count() << "s\t(" << gigabytes_handled / elapsed_seconds_find_sync.count() << " Go/s)" << std::endl;
58 |
59 | // auto start_find_async = std::chrono::high_resolution_clock::now();
60 | // auto result_async = table.findRows(attribute_predicates, value_predicates, true, false, false);
61 | // auto end_find_async = std::chrono::high_resolution_clock::now();
62 | // // Checking solution
63 | // for (long unsigned int i = 0; i < result_async.size(); i++)
64 | // {
65 | // if (result_sync[i] != result_async[i])
66 | // {
67 | // return 1;
68 | // }
69 | // }
70 | // auto elapsed_seconds_find_async = std::chrono::duration_cast>(end_find_async - start_find_async);
71 | // std::cout << "Async find in: " << elapsed_seconds_find_async.count() << "s\t(" << gigabytes_handled / elapsed_seconds_find_async.count() << " Go/s)" << std::endl;
72 |
73 | // auto start_load = std::chrono::high_resolution_clock::now();
74 | // table.loadEverything();
75 | // auto end_load = std::chrono::high_resolution_clock::now();
76 | // auto elapsed_seconds_load = std::chrono::duration_cast>(end_load - start_load);
77 | // std::cout << "Loaded in: " << elapsed_seconds_load.count() << "s\t(" << gigabytes_handled / elapsed_seconds_load.count() << " Go/s)" << std::endl;
78 |
79 | // auto start_find_memory = std::chrono::high_resolution_clock::now();
80 | // auto result_memory = table.findRows(attribute_predicates, value_predicates, false, true, false);
81 | // auto end_find_memory = std::chrono::high_resolution_clock::now();
82 | // Checking solution
83 | // for (long unsigned int i = 0; i < result_async.size(); i++)
84 | // {
85 | // if (result_sync[i] != result_memory[i])
86 | // {
87 | // return 1;
88 | // }
89 | // }
90 | // auto elapsed_seconds_find_memory = std::chrono::duration_cast>(end_find_memory - start_find_memory);
91 | // std::cout << "Memory find in: " << elapsed_seconds_find_memory.count() << "s\t(" << gigabytes_handled / elapsed_seconds_find_memory.count() << " Go/s)" << std::endl;
92 |
93 | // table.unloadEverything();
94 |
95 | auto start_find_sync_omp = std::chrono::high_resolution_clock::now();
96 | auto result_sync_omp = table.findRows(attribute_predicates, value_predicates, false, false, true);
97 | auto end_find_sync_omp = std::chrono::high_resolution_clock::now();
98 | // // Checking solution
99 | // for (long unsigned int i = 0; i < result_async.size(); i++)
100 | // {
101 | // if (result_sync[i] != result_sync_omp[i])
102 | // {
103 | // return 1;
104 | // }
105 | // }
106 | auto elapsed_seconds_find_sync_omp = std::chrono::duration_cast>(end_find_sync_omp - start_find_sync_omp);
107 | std::cout << "Sync find (OMP) in: " << elapsed_seconds_find_sync_omp.count() << "s\t(" << gigabytes_handled / elapsed_seconds_find_sync_omp.count() << " Go/s)" << std::endl;
108 |
109 | // table.loadEverything();
110 | // auto start_find_memory_omp = std::chrono::high_resolution_clock::now();
111 | // auto result_memory_omp = table.findRows(attribute_predicates, value_predicates, false, true, true);
112 | // auto end_find_memory_omp = std::chrono::high_resolution_clock::now();
113 | // Checking solution
114 | // for (long unsigned int i = 0; i < result_async.size(); i++)
115 | // {
116 | // if (result_sync[i] != result_memory_omp[i])
117 | // {
118 | // return 1;
119 | // }
120 | // }
121 | // auto elapsed_seconds_find_memory_omp = std::chrono::duration_cast>(end_find_memory_omp - start_find_memory_omp);
122 | // std::cout << "Memory find (OMP) in: " << elapsed_seconds_find_memory_omp.count() << "s\t(" << gigabytes_handled / elapsed_seconds_find_memory_omp.count() << " Go/s)" << std::endl;
123 |
124 | // table.unloadEverything();
125 | return 0;
126 | }
127 |
--------------------------------------------------------------------------------
/src/main.hpp:
--------------------------------------------------------------------------------
1 | #ifndef MAIN_HPP
2 | #define MAIN_HPP
3 |
4 | constexpr int CHUNK_SIZE = 128 * 1024; // = Number of ELEMENT (not size) in a chunk
5 | constexpr int QUEUE_DEPTH = 10;
6 |
7 | #endif
8 |
--------------------------------------------------------------------------------
/tests/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | set(BINARY ${CMAKE_PROJECT_NAME}_tst)
2 |
3 | file(GLOB_RECURSE TEST_SOURCES LIST_DIRECTORIES false *.h *.cpp)
4 |
5 | set(SOURCES ${TEST_SOURCES})
6 |
7 | add_executable(${BINARY} ${TEST_SOURCES})
8 |
9 | add_test(NAME ${BINARY} COMMAND ${BINARY})
10 |
11 | target_link_libraries(${BINARY} PUBLIC ${CMAKE_PROJECT_NAME}_lib gtest)
--------------------------------------------------------------------------------
/tests/FileCreationTest.cpp:
--------------------------------------------------------------------------------
1 | #include "gtest/gtest.h"
2 | #include "Table.hpp"
3 |
4 | TEST(FileCreationTest, FileCreation)
5 | {
6 | Table table{10, 10};
7 | }
8 |
--------------------------------------------------------------------------------
/tests/TableFindRows.cpp:
--------------------------------------------------------------------------------
1 | #include "gtest/gtest.h"
2 | #include "Table.hpp"
3 |
4 | TEST(TableFindRowsTest, TableFindRowsWithWriteRows)
5 | {
6 | const int table_size = 16 * 16;
7 | const int attributes_count = 3;
8 | const int values_count = 2;
9 | Table table{attributes_count, values_count};
10 |
11 | unsigned int attributes[table_size * attributes_count];
12 | double values[table_size * values_count];
13 |
14 | for (unsigned int i = 0; i < table_size; i++)
15 | {
16 | attributes[table_size * 0 + i] = 1;
17 | attributes[table_size * 1 + i] = i;
18 | attributes[table_size * 2 + i] = 2 * i;
19 | values[table_size * 0 + i] = 0.1 + i;
20 | values[table_size * 1 + i] = 1.0 * i;
21 | }
22 | table.writeRows(0, table_size, attributes, values);
23 |
24 | int attribute_predicates[3] = {1, 2, 4};
25 | double value_predicates[2] = {2.1, 2.0};
26 | auto result = table.findRows(attribute_predicates, value_predicates, false, false, false);
27 |
28 | for (long unsigned int i = 0; i < result.size(); i++)
29 | {
30 | for (long unsigned int j = 0; j < CHUNK_SIZE; j++)
31 | {
32 | EXPECT_EQ(result[i][j], i * CHUNK_SIZE + j == 2);
33 | }
34 | }
35 | }
36 |
--------------------------------------------------------------------------------
/tests/TableReadWriteRowsTest.cpp:
--------------------------------------------------------------------------------
1 | #include "gtest/gtest.h"
2 | #include "Table.hpp"
3 |
4 | TEST(TableReadWriteRowsTest, TableReadWriteRows)
5 | {
6 | const unsigned long int number_of_rows = 1024 * 1024;
7 | const unsigned long int write_block_size = 128 * 1024;
8 | const unsigned int attributes_count = 3;
9 | const unsigned int values_count = 2;
10 |
11 | Table table{attributes_count, values_count};
12 | for (unsigned long int i = 0; i < number_of_rows; i += write_block_size)
13 | {
14 | unsigned int attributes[write_block_size * attributes_count];
15 | double values[write_block_size * values_count];
16 | for (unsigned long int j = 0; j < write_block_size; j++)
17 | {
18 | attributes[write_block_size * 0 + j] = 1;
19 | attributes[write_block_size * 1 + j] = i + j;
20 | attributes[write_block_size * 2 + j] = 2 * (i + j);
21 | values[write_block_size * 0 + j] = 0.1 + i + j;
22 | values[write_block_size * 1 + j] = 1.0 * (i + j);
23 | }
24 | table.writeRows(i, write_block_size, attributes, values);
25 | }
26 |
27 | for (unsigned long int i = 0; i < 3; i++)
28 | {
29 | EXPECT_EQ(table.readInt(i, 0), 1);
30 | EXPECT_EQ(table.readInt(i, 1), i);
31 | EXPECT_EQ(table.readInt(i, 2), 2 * i);
32 | EXPECT_EQ(table.readDouble(i, 3), i + 0.1);
33 | EXPECT_EQ(table.readDouble(i, 4), 1.0 * i);
34 | }
35 | for (unsigned long int i = number_of_rows - 3; i < number_of_rows; i++)
36 | {
37 | EXPECT_EQ(table.readInt(i, 0), 1);
38 | EXPECT_EQ(table.readInt(i, 1), i);
39 | EXPECT_EQ(table.readInt(i, 2), 2 * i);
40 | EXPECT_EQ(table.readDouble(i, 3), i + 0.1);
41 | EXPECT_EQ(table.readDouble(i, 4), 1.0 * i);
42 | }
43 |
44 | int row_to_find = 19542;
45 | int attribute_predicates[3] = {1, row_to_find, 2 * row_to_find};
46 | double value_predicates[2] = {0.1 + row_to_find, 1.0 * row_to_find};
47 | auto result = table.findRows(attribute_predicates, value_predicates, false, false, false);
48 |
49 | for (long unsigned int i = 0; i < result.size(); i++)
50 | {
51 | for (long unsigned int j = 0; j < CHUNK_SIZE; j++)
52 | {
53 | EXPECT_EQ(result[i][j], i * CHUNK_SIZE + j == row_to_find);
54 | }
55 | }
56 | }
57 |
58 | TEST(TableReadWriteRowsTest, TableReadWriteRowsWithOffset)
59 | {
60 | const int offset = CHUNK_SIZE / 4;
61 | const int number_of_rows = 2 * CHUNK_SIZE;
62 | Table table{1, 1};
63 | unsigned int attributes[number_of_rows];
64 | double values[number_of_rows];
65 | for (int i = 0; i < number_of_rows - offset; i++)
66 | {
67 | attributes[i] = 1;
68 | values[i] = 0.1 + i;
69 | }
70 | table.writeRows(offset, number_of_rows, attributes, values);
71 | }
72 |
--------------------------------------------------------------------------------
/tests/TableReadWriteTest.cpp:
--------------------------------------------------------------------------------
1 | #include "gtest/gtest.h"
2 | #include "Table.hpp"
3 |
4 | TEST(TableReadWriteTest, TableReadWriteWithWriteRows)
5 | {
6 | const int table_size = 3;
7 | const int attributes_count = 3;
8 | const int values_count = 2;
9 | Table table{attributes_count, values_count};
10 |
11 | unsigned int attributes[table_size * attributes_count];
12 | double values[table_size * values_count];
13 |
14 | for (unsigned int i = 0; i < table_size; i++)
15 | {
16 | attributes[table_size * 0 + i] = 1;
17 | attributes[table_size * 1 + i] = i;
18 | attributes[table_size * 2 + i] = 2 * i;
19 | values[table_size * 0 + i] = 0.1 + i;
20 | values[table_size * 1 + i] = 0.1 * i;
21 | }
22 |
23 | table.writeRows(0, table_size, attributes, values);
24 |
25 | for (int i = 0; i < table_size; i++)
26 | {
27 | EXPECT_EQ(table.readInt(i, 0), 1);
28 | EXPECT_EQ(table.readInt(i, 1), i);
29 | EXPECT_EQ(table.readInt(i, 2), 2 * i);
30 | EXPECT_EQ(table.readDouble(i, 3), i + 0.1);
31 | EXPECT_EQ(table.readDouble(i, 4), 0.1 * i);
32 | }
33 | }
34 |
--------------------------------------------------------------------------------
/tests/main.cpp:
--------------------------------------------------------------------------------
1 | #include "gtest/gtest.h"
2 |
3 | int main(int argc, char **argv) {
4 | ::testing::InitGoogleTest(&argc, argv);
5 | return RUN_ALL_TESTS();
6 | }
7 |
--------------------------------------------------------------------------------