Difference between revisions of "Winter 2015 SPO600 Weekly Schedule"

From CDOT Wiki
Jump to: navigation, search
(Course Projects)
 
(39 intermediate revisions by 3 users not shown)
Line 1: Line 1:
 
[[Category:Winter 2015 SPO600]]
 
[[Category:Winter 2015 SPO600]]
{{Admon/important|It's Alive!|This [[SPO600]] weekly schedule will be updated as the course proceeds - dates and content are subject to change. The cells in the summary table will be linked to relevant resources and labs as the course progresses.}}
+
{{Admon/obsolete|[[Current SPO600 Weekly Schedule]]}}
 
 
For the Fall 2014 version of the weekly schedule, see [[Fall 2014 SPO600 Weekly Schedule]].
 
  
 
== Schedule Summary Table ==
 
== Schedule Summary Table ==
Line 18: Line 16:
 
|-
 
|-
  
|3||Jan 26||[[#Tuesday (Jan 27|Profiling: Finding the pain points (Lab 3)]]||[[#Thursday (Jan 29)|Computer architecture overview]]||[[#Week 3 Deliverables|Blog about the profiling results and pain points discovered (Lab 3)]]
+
|3||Jan 26||[[#Tuesday (Jan 27)|Profiling: Finding the pain points (Lab 3)]]||[[#Thursday (Jan 29)|Computer architecture overview]]||[[#Week 3 Deliverables|Blog about the profiling results and pain points discovered (Lab 3)]]
 
|-
 
|-
  
|4||Feb 2||[[#Tuesday (Feb 3)|Compiled C Lab (Lab 4)]]||[[#Thursday (Feb 5)|Compiler optimizations]]||[[#Week 4 Deliverables|Blog a commentary on the Compiled C Lab (Lab 3).]]
+
|4||Feb 2||[[#Tuesday (Feb 3)|Compiled C Lab (Lab 4)]]||[[#Thursday (Feb 5)|Compiler optimizations]]||[[#Week 4 Deliverables|Blog a commentary on the Compiled C Lab (Lab 4).]]
 
|-
 
|-
  
Line 30: Line 28:
 
|-
 
|-
  
|7||Feb 23 ||[[#Tuesday (Feb 24)|Project presentations - Stage 1.]]||[[#Thursday (Feb 26)|Prepare a presentation on a selected aspect of assembly language.]]||[[#Week 6 Deliverables|Blog about your project plans - this will be used to assign your 1st Project Marks. Be prepared to give your presentation.]]
+
|7||Feb 23||[[#Tuesday (Feb 24)|Assembly lab (Lab 5)]]||[[#Thursday (Feb 26)|Inline assembly lab and Course Projects]]||[[#Week 7 Deliverables|Blog about the Assembly Lab (Lab 5) and start project investigation.]]
  
 
|-style="background: #f0f0ff"
 
|-style="background: #f0f0ff"
|Study Week||Mar 2||colspan="3" align="center"|Study Week - No classes!
+
|Study Week||Mar 2||colspan="3" align="center"|Study Week - No classes! Please work on your initial project investigation.
 
|-
 
|-
  
|8||Mar 9||[[#Tuesday (Mar 10)|Assembly language presentations]]||[[#Thursday (Mar 12)|Assembly language presentations]]||[[#Week 7 Deliverables|Blog about your assembly language presentation.]]
+
|8||Mar 9||[[#Tuesday (Mar 10)|Project discussion, and Reading Reference Documentation.]]||[[#Thursday (Mar 12)|Project presentations - Stage 1]]||[[#Week 8 Deliverables|Blog about your project progress - this will be used to assign your 1st Project marks.]]
 
|-
 
|-
  
|9||Mar 16||[[#Tuesday (Mar 17)|Assembly lab (Lab 5)]]||[[#Thursday (Mar 19)|Reading reference documentation (and Lab 5 Continued)]]||[[#Week 9 Deliverables|Blog about the assembly lab (Lab 5).]]
+
|9||Mar 16||[[#Tuesday (Mar 17)|Upstreaming]]||[[#Thursday (Mar 19)|Spinlocks and Tail Call Optimization]]||[[#Week 9 Deliverables|Blog about your project work.]]
 
|-
 
|-
  
|10||Mar 23||[[#Tuesday (Mar 24)|Project Presentations - Stage 2]]||[[#Thursday (Mar 26)|Uptreaming]]||[[#Week 8 Deliverables|Blog about your project progress - this will be used to assign your 2nd Project Marks.]]
+
|10||Mar 23||[[#Tuesday (Mar 24)|Portable alternatives to Assembler]]||[[#Thursday (Mar 26)|Project Presentations - Stage 2]]||[[#Week 10 Deliverables|Blog about your project progress - this will be used to assign your 2nd Project Marks.]]
 
|-
 
|-
  
|11||Mar 30||[[#Tuesday (Mar 31)|Inline Assembler]]||[[#Thursday (Apr 2)|Find inline assembler in 3 packages (Lab 6)]]||[[#Week 9 Deliverables|Blog about the assembler you found (Lab 6).]]
+
|11||Mar 30||[[#Tuesday (Mar 31)|Hack Session]] (Professor will not be present)||[[#Thursday (Apr 2)|SIMD]]||[[#Week 11 Deliverables|Blog about SIMD/vectorization and your project.]]
 
|-
 
|-
  
|12||Apr 6||[[#Tuesday (Apr 7)|Discussion & Hack Session]]||[[#Thursday (Apr 9)|Discussion & Hack Session]]||[[#Week 10 Deliverables|Blog about your progress.]]
+
|12||Apr 6||[[#Tuesday (Apr 7)|Discussion & Hack Session]]||[[#Thursday (Apr 9)|Vectorization]]||[[#Week 12 Deliverables|Blog about your work.]]
 
|-
 
|-
  
|13||Apr 13||[[#Tuesday (Apr 14)|Project Presentations - Stage 3]]||[[#Thursday (Apr 16)|Wrap-up Session]]||[[#Week 11 Deliverables|Blog about your project progress - this will be used to assign your 3rd project marks.]]
+
|13||Apr 13||[[#Tuesday (Apr 14)|Project Presentations - Stage 3]]||[[#Thursday (Apr 16)|Wrap-up Session]]||[[#Week 13 Deliverables|Blog about your project progress - this will be used to assign your 3rd project marks.]]
 
|-
 
|-
  
Line 68: Line 66:
 
|Labs||align="right"|10%||See deliverables column above.
 
|Labs||align="right"|10%||See deliverables column above.
 
|-
 
|-
|Project work||align="right"|60%||Feb 27 (15%), March 27 (20%), April 22 (25%)
+
|Project work||align="right"|60%||March 16 (15%), March 31 (20%), April 22 (25%)
 
|}
 
|}
  
Line 75: Line 73:
 
=== Tuesday (Jan 13) ===
 
=== Tuesday (Jan 13) ===
  
==== Introduction to the Problem ====
+
==== Introduction to the Problems ====
  
* Most software is written in a '''high-level language''' which can be compiled into [[Machine Language|machine code]] for a specific architecture. However, there is a lot of existing code that contains some architecture-specific code fragments written in [[Assembly Language]] (or, in some cases, machine-specific high-level code).
+
===== Porting and Portability =====
 +
* Most software is written in a '''high-level language''' which can be compiled into [[Machine Language|machine code]] for a specific computer architecture. In many cases, this code can be compiled for multiple architectures. However, there is a lot of existing code that contains some architecture-specific code fragments written in [[Assembly Language]] (or, in some cases, machine-specific high-level code).
 
* Reasons for writing code in Assembly Langauge include:
 
* Reasons for writing code in Assembly Langauge include:
 
** Performance
 
** Performance
Line 84: Line 83:
 
* Most of the historical reasons for including assembler are no longer valid. Modern compilers can out-perform most hand-optimized assembly code, atomic operations can be handled by libraries or [[Compiler Intrinsics|compiler intrinsics]], and most hardware access should be performed through the operating system or appropriate libraries.
 
* Most of the historical reasons for including assembler are no longer valid. Modern compilers can out-perform most hand-optimized assembly code, atomic operations can be handled by libraries or [[Compiler Intrinsics|compiler intrinsics]], and most hardware access should be performed through the operating system or appropriate libraries.
 
* A new architecture has appeared: Aarch64, which is part of [http://www.arm.com/products/processors/instruction-set-architectures/armv8-architecture.php ARMv8]. This is the first new [[Computer Architecture|computer architecture]] to appear in several years (at least, the first mainstream computer architecture).
 
* A new architecture has appeared: Aarch64, which is part of [http://www.arm.com/products/processors/instruction-set-architectures/armv8-architecture.php ARMv8]. This is the first new [[Computer Architecture|computer architecture]] to appear in several years (at least, the first mainstream computer architecture).
* There are over 1400 software packages/modules present in GNU Linux systems which contain architecture-specific [[Assembly Language|assembly language code]] or have other portability issues. Most of these packages cannot be built on Aarch64 systems without modification.
+
* At this point, most key open source software (the software typically present in a Linux distribution such as Ubuntu or Fedora, for example) now runs on AArch64. However, it may not run as well as on older architectures (such as x86_64).
 +
 
 +
===== Benchmarking and Profiling =====
 +
Benchmarking involves testing software performance under controlled conditions so that the performance can be compared to other software, the same software operating on other types of computers, or so that the impact of a change to the software can be gauged.
 +
 
 +
Profiling is the process of analyzing software performance on finer scale, determining resource usage per program part (typically per function/method). This can identify software bottlenecks and potential targets for optimization.
 +
 
 +
===== Optimization =====
 +
Optimization is the process of evaluating different ways that software can be written or built and selecting the option that has the best performance tradeoffs.
 +
 
 +
Optimization may involve substituting software algorithms, altering the sequence of operations, using architecture-specific code, or altering the build process. It is important to ensure that the optimized software produces correct results and does not cause an unacceptable performance regression for other use-cases, system configurations, operating systems, or architectures.
 +
 
 +
The definition of "performance" varies according to the target system and the operating goals. For example, in some contexts, low memory or storage usage is important; in other cases, fast operation; and in other cases, low CPU utilization or long battery life may be the most important factor. It is often possible to trade off performance in one area for another; using a lookup table, for example, can reduce CPU utilization and improve battery life in some algorithms, in return for increased memory consumption.
 +
 
 +
Most advanced compilers perform some level of optimization, and the options selected for compilation can have a significant effect on the trade-offs made by the compiler, affecting memory usage, execution speed, executable size, power consumption, and debuggability.
 +
 
 +
===== Build Process =====
 +
Building software is a complex task that many developers gloss over. The simple act of compiling a program invokes a process with five or more stages, including pre-proccessing, compiling, optimizing, assembling, and linking. However, a complex software system will have hundreds or even thousands of source files, as well as dozens or hundreds of build configuration options, auto configuration scripts (cmake, autotools), build scripts (such as Makefiles) to coordinate the process, test suites, and more.
 +
 
 +
The build process varies significantly between software packages. Most software distribution projects (including Linux distributions such as Ubuntu and Fedora) use a packaging system that further wraps the build process in a standardized script format, so that different software packages can be built using a consistent process.
 +
 
 +
In order to get consistent and comparable benchmark results, you need to ensure that the software is being built in a consistent way. Altering the build process is one way of optimizing software.
 +
 
 +
Note that the build time for a complex package can range up to hours or even days!
  
 
==== Course Projects ====
 
==== Course Projects ====
Line 92: Line 114:
 
# Determine how the performance of one specific part of the stack could be improved, either specifically on AArch64 systems (without detriment to other systems) or on all systems.
 
# Determine how the performance of one specific part of the stack could be improved, either specifically on AArch64 systems (without detriment to other systems) or on all systems.
 
# Prepare a fix/patch for the software to implement that performance improvement.
 
# Prepare a fix/patch for the software to implement that performance improvement.
# Prove that your changes do not cause any performance regressions.
+
# Prove that your changes do not cause any unacceptable performance regressions (Note: there might be performance regressions which are considered acceptable!).
 
# Upstream your Code - Submit your code to the upstream (originating) software project so that it can be incorporated into future versions of the software. This will involve going through a code review to ensure that your code is compatible with and acceptable to the upstream community.
 
# Upstream your Code - Submit your code to the upstream (originating) software project so that it can be incorporated into future versions of the software. This will involve going through a code review to ensure that your code is compatible with and acceptable to the upstream community.
  
Line 115: Line 137:
 
=== Thursday (Jan 15) ===
 
=== Thursday (Jan 15) ===
  
* Benchmarking and Profiling
+
* Benchmarking
 
+
* Profiling
* Profiling with <code>gprof</code>
 
** Build with profiling enabled (use the option <code>-pg</code> with both gcc and ld)
 
** Run the profile-enabled executable
 
** Analyze the data in the <code>gmon.out</code> file
 
*** <code>gprof ''nameOfBinary''</code> # Displays text profile including call graph
 
*** <code>gprof ''nameOfBinary'' | gprof2dot | dot | display -</code> # Displays visualization of call graph
 
 
 
Resources
 
* [https://sourceware.org/binutils/docs-2.16/gprof/ GProf Manual]
 
* [http://www.thegeekstuff.com/2012/08/gprof-tutorial/ Profiling with GProf]
 
  
 
=== Week 1 Deliverables ===
 
=== Week 1 Deliverables ===
  
 
# Set up your [[SPO600 Communication Tools]] - in particular, set up a blog and add it to [http://zenit.senecac.on.ca/~chris.tyler/planet/ Planet CDOT] (via the [[Planet CDOT Feed List]]).
 
# Set up your [[SPO600 Communication Tools]] - in particular, set up a blog and add it to [http://zenit.senecac.on.ca/~chris.tyler/planet/ Planet CDOT] (via the [[Planet CDOT Feed List]]).
# Add yourself to the [[Fall 2014 SPO600 Participants]] page (leave the projects columns blank).
+
# Add yourself to the [[Winter 2015 SPO600 Participants]] page (leave the projects columns blank).
 
# Generate a [[SSH#Using_Public_Keys_with_SSH|pair of keys]] for [[SSH]] and email the public key to your professor.
 
# Generate a [[SSH#Using_Public_Keys_with_SSH|pair of keys]] for [[SSH]] and email the public key to your professor.
 
# Sign and return the [[Open Source Professional Option Student Agreement]].
 
# Sign and return the [[Open Source Professional Option Student Agreement]].
# Optional but recommended: [[SPO600 Host Setup|Set up a Fedora 20 system]].
+
# Optional but recommended: [[SPO600 Host Setup|Set up a personal Fedora system]].
  
 
== Week 2 ==
 
== Week 2 ==
  
 
=== Tuesday (Jan 20) ===
 
=== Tuesday (Jan 20) ===
{{Admon/tip|Bring Your Laptop|Classes are held in a [[Active Learning Classroom]]. If you have a laptop or other device with a VGA or HDMI output (such as a smartphone!) please bring it. You'll need either a local linux environment or an SSH client -- which is built-in to Linux, Mac, and Chromebook systems, and readily available for Windows, Android, and iOS devices.}}
+
{{Admon/tip|Bring Your Laptop|Classes are held in a [[Active Learning Classroom]]. If you have a laptop or other device with a VGA or HDMI output (such as a smartphone!) please bring it. You'll need either a local linux environment or an [[SSH]] client -- which is built-in to Linux, Mac, and Chromebook systems, and readily available for Windows, Android, and iOS devices.}}
  
 +
Status Check
 +
* Introductions around the Room
 +
** Name
 +
** Program
 +
** Interest in the course
 +
** Results of <code>gzip</code> benchmarking and most interesting thing you observed
 +
* Check the [[SPO600 Servers|server accounts]] that were set up with your [[SSH]] keys
 +
* Sheets from Last Week
 +
** Open Source Student Agreement
 +
** Survey (Optional)
  
 
Working with the Code
 
Working with the Code
* Working with GIT
+
* Getting Code
* Working with other version control systems
+
** In a tarball
 
+
** From git
 
+
*** Git basics
Building the Code
+
** Working with other version control systems
* Make
+
* Building the Code
* Configuration tools (autotools, cmake)
+
** Make
* The compiler toolchain
+
** Configuration tools (autotools, cmake)
** Preprocessor
+
** The compiler toolchain
** Compiler
+
*** Preprocessor
** Assembler
+
*** Compiler
** Linker
+
*** Assembler
* Debug vs. Non-debug/Stripped binaries
+
*** Linker
 
+
** Debug vs. Non-debug/Stripped binaries
  
 
Looking at How Distributions Package the Code
 
Looking at How Distributions Package the Code
 
* Using fedpkg
 
* Using fedpkg
 +
 +
==== Resources ====
 +
* [https://sourceware.org/binutils/docs-2.16/gprof/ GProf Manual]
 +
* [http://www.thegeekstuff.com/2012/08/gprof-tutorial/ Profiling with GProf]
  
 
=== Thursday (Jan 22) ===
 
=== Thursday (Jan 22) ===
Line 168: Line 194:
 
* Complete and blog your conclusion to the [[SPO600 Code Review Lab|Code Review Lab (Lab 1)]].
 
* Complete and blog your conclusion to the [[SPO600 Code Review Lab|Code Review Lab (Lab 1)]].
 
* Blog your baseline data from the [[SPO600 Baseline Builds and Benchmarking Lab|Baseline Builds and Benchmarking lab (Lab 2)]].
 
* Blog your baseline data from the [[SPO600 Baseline Builds and Benchmarking Lab|Baseline Builds and Benchmarking lab (Lab 2)]].
 +
 +
 +
== Week 3 ==
 +
 +
=== Tuesday (Jan 27) ===
 +
 +
Profiling
 +
* Profiling with <code>gprof</code>
 +
** Build with profiling enabled (use the option <code>-pg</code> with both gcc and ld)
 +
** Run the profile-enabled executable
 +
** Analyze the data in the <code>gmon.out</code> file
 +
*** <code>gprof ''nameOfBinary''</code> # Displays text profile including call graph
 +
*** <code>gprof ''nameOfBinary'' | gprof2dot | dot | display -</code> # Displays visualization of call graph
 +
* Other profiling tools
 +
** OProf, SystemTap, and others
 +
* [[SPO600 Profiling Lab]] (Lab 3)
 +
 +
=== Thursday (Jan 29) ===
 +
 +
* [[Computer Architecture]] overview (see also the [[:Category:Computer Architecture|Computer Architecture Category]])
 +
 +
=== Week 3 Deliverables ===
 +
 +
* Complete and blog your conclusions to the [[SPO600 Profiling Lab|Profiling Lab (Lab 3)]].
 +
 +
 +
== Week 4 ==
 +
 +
=== Tuesday (Feb 3) ===
 +
* [[SPO600 Compiled C Lab|Compiled C Lab (Lab 4)]]
 +
 +
=== Thursday (Feb 5) ===
 +
* [[Compiler Optimizations]]
 +
 +
=== Week 4 Deliverables ===
 +
* Blog your [[SPO600 Compiled C Lab|Compiled C Lab (Lab 4)]] results. Be sure to include a reflective section on what you learned.
 +
* Select your topic for the [[Winter 2015 SPO600 Platform Specific Code Presentation|Platform-Specific Code Presentation]].
 +
 +
== Week 5 ==
 +
* Prepare your [[Winter 2015 SPO600 Platform Specific Code Presentation|Platform-Specific Code Presentation]].
 +
 +
=== Week 5 Deliverables ===
 +
* Be ready to [[Winter 2015 SPO600 Platform Specific Code Presentation|present]].
 +
 +
== Week 6 ==
 +
=== Tuesday (Feb 17) ===
 +
* Introduction to the [http://www.96boards.org/products/hikey/ HiKey] board and the [http://www.96boards.org/ 96Boards] project.
 +
* [[Winter 2015 SPO600 Platform Specific Code Presentation|Presentations on platform-specific code]]
 +
 +
=== Thursday (Feb 19) ===
 +
* [[Winter 2015 SPO600 Platform Specific Code Presentation|Presentations on platform-specific code]]
 +
 +
=== Week 6 Deliverables ===
 +
* Blog about your presentation.
 +
 +
== Week 7 ==
 +
=== Tuesday (Feb 24) ===
 +
* [[SPO600 Assembler Lab|Assembly language lab (Lab 5)]]
 +
 +
=== Thursday (Feb 26) ===
 +
* Remaining [[Winter 2015 SPO600 Platform Specific Code Presentation|Presentations on platform-specific code]]
 +
* [[Inline Assembly Language]]
 +
* [[Winter 2015 SPO600 Project|Course Projects]]
 +
 +
=== Week 7 Deliverables ===
 +
* Blog your conclusion to the [[SPO600 Assembler Lab|assembly language lab (Lab 5)]] - See the ''Deliverables'' section in the lab instructions for details on what to include in your blog post.
 +
 +
== Week 8 ==
 +
 +
=== Tuesday (Mar 10) ===
 +
* Project discussion
 +
** Discussion of project ideas and fine-tuning of project plans
 +
** Next steps
 +
** Addressing Problem Areas
 +
* Accessing and Reading Reference Documentation
 +
** AArch64 ISA
 +
 +
=== Thursday (Mar 12) ===
 +
* Project presentations
 +
** Provide a very short (2 minute) overview of your project. Include:
 +
*** Which piece of software you are working on
 +
*** What area of that software needs optimization/performance tuning
 +
*** How you are going to perform the optimization/tuning (algorithm replacement, platform-specific code, removing platform-specific code, build options, and so forth)
 +
*** Plans for the next step
 +
*** Engagement with the upstream community
 +
* Use this project presentation to tell the class what you're working on and incorporate feedback into your blog post
 +
 +
=== Week 8 Deliverables ===
 +
* Blog about your project and plans, incorporating feedback from your presentation
 +
* Add your project to the [[Winter 2015 SPO600 Participants|Participants and Project Table]].
 +
 +
== Week 9 ==
 +
 +
=== Tuesday (Mar 17) ===
 +
* Upstreaming
 +
** Git Discussion
 +
** Creating a Patch
 +
** Various open source community workflows
 +
 +
=== Thursday (Mar 19) ===
 +
* Spinlocks
 +
** The need for Atomics in Spinlocks
 +
** Using Intrinsics for Atomic Operations
 +
* Tail Call Optimization (TCO)
 +
** gcc will perform TCO at -O2 and higher
 +
** Important to ensure that gcc recognizes code patterns that permit TCO to be applied
 +
 +
=== Week 9 Deliverables ===
 +
* Blog at least once (and ideally more than once) about your Project
 +
** Frequent shorter posts are better than rare long ones
 +
** Aim to make steady progress on your project, a bit each day
 +
** Blog about your evolving project plan and the steps you're making along the way
 +
** Keep the work short, and focus on narrowing the scope of work as early as possible
 +
** Keep talking to the community
 +
 +
== Week 10 ==
 +
 +
=== Tuesday (Mar 24) ===
 +
* Discussion of GCC Intrinsics
 +
** Where to find documentation: [https://gcc.gnu.org/onlinedocs/ GCC Manual]
 +
** __atomic vs __sync intrinsic families
 +
* Examination of Build Files
 +
** Differential analysis of successful and unsuccessfull build
 +
 +
=== Thursday (Mar 26) ===
 +
* Project Presentations - Stage 2
 +
** At this point, you should have a proposed patch that you're working to get upstream
 +
 +
=== Week 10 Deliverables ===
 +
* Blog posts about your project
 +
** Include information about your patch(es)
 +
** Include links to discussion about the patch(es) with the community (e.g., link to bug/issue tracker entries or email archives)
 +
** If you don't have patches and/or have not pushed them upstream yet, describe the state of your project and your plan to get changes upstream.
 +
* Blogs will be marked Monday for Stage 2 project work (20%)
 +
 +
== Week 11 ==
 +
 +
=== Tuesday (Mar 31) ===
 +
Your professor will be away. You are welcome to use the classroom for a project hacking time and collaboration with your colleagues.
 +
 +
=== Thursday (Apr 2) ===
 +
* Single Instruction / Multiple Data (SIMD)
 +
** Brief look at x86_64 & AArch64 implementations of SIMD
 +
 +
=== Week 11 Deliverables ===
 +
* Blog about your ongoing project work. Please post at least 1-2 entries per week.
 +
 +
== Week 12 ==
 +
 +
=== Tuesday (Apr 7) ===
 +
* Discussion & Hacking Session
 +
 +
=== Thursday (Apr 9) ===
 +
* Vectorization
 +
** Using Intrinsics
 +
** Using Auto-Vectorization
 +
 +
=== Week 12 Deliverables ===
 +
* Blog about your ongoing project work. Please post at least 1-2 entries per week.
 +
* You should have patches (code, makefiles, documentation, or tests) in the upstream review process at or before this point.
 +
* Blog about vectorization - specifically:
 +
** What -O level and/or options are needed to turn on the auto-vectorizer in GCC
 +
** At least three limitations -- conditions under which the vectorization will not be performed
 +
** The significance of alignment and non-overlapping pointers for vectorization
 +
 +
== Week 13 ==
 +
 +
=== Tuesday (Apr 14) ===
 +
* Informal Project Presentations - Phase 3
 +
 +
=== Thursday (Apr 16) ===
 +
* Wrap-up Session
 +
 +
=== Week 13 Deliverables ===
 +
* Blog about your ongoing project work. Your patches should be through the upstream review process.
 +
* Final date for posting about your project: April 22.
  
 
<!--
 
<!--

Latest revision as of 12:21, 29 August 2015

Important.png
This page may be obsolete.
It contains historical information. For current information, please see Current SPO600 Weekly Schedule.

Schedule Summary Table

This is a summary/index table. Please follow the links in each cell for additional detail -- especially for the Deliverables column.

Week Week of... Tuesday Thursday Deliverables
(Summary - click for details)
1 Jan 12 Introduction to Software Porting, Portability, Benchmarking, and Optimization / How is code accepted into an open source project? (Lab 1) Benchmarking and Profiling Set up accounts, and blog a comparision of code reviews in two communities (Lab 1).
2 Jan 19 Working with the code / Building Build and benchmark a portion of the LAMP stack (Lab 2) Blog about your baseline data (Lab 2)
3 Jan 26 Profiling: Finding the pain points (Lab 3) Computer architecture overview Blog about the profiling results and pain points discovered (Lab 3)
4 Feb 2 Compiled C Lab (Lab 4) Compiler optimizations Blog a commentary on the Compiled C Lab (Lab 4).
5 Feb 9 Linaro Connect - No classes scheduled
Prepare a presentation on a selected topic about platform-specific code.
Be prepared to present your topic.
6 Feb 16 Presentations on platform-specific code. Presentations on platform-specific code. Blog about your presentation.
7 Feb 23 Assembly lab (Lab 5) Inline assembly lab and Course Projects Blog about the Assembly Lab (Lab 5) and start project investigation.
Study Week Mar 2 Study Week - No classes! Please work on your initial project investigation.
8 Mar 9 Project discussion, and Reading Reference Documentation. Project presentations - Stage 1 Blog about your project progress - this will be used to assign your 1st Project marks.
9 Mar 16 Upstreaming Spinlocks and Tail Call Optimization Blog about your project work.
10 Mar 23 Portable alternatives to Assembler Project Presentations - Stage 2 Blog about your project progress - this will be used to assign your 2nd Project Marks.
11 Mar 30 Hack Session (Professor will not be present) SIMD Blog about SIMD/vectorization and your project.
12 Apr 6 Discussion & Hack Session Vectorization Blog about your work.
13 Apr 13 Project Presentations - Stage 3 Wrap-up Session Blog about your project progress - this will be used to assign your 3rd project marks.
Exam Week Apr 20 Exam Week - No exam in this course!

Evaluation

Category Percentage Evaluation Dates
Communication 20% January 31, February 28, March 27, April 22
Quizzes 10% May be held during any class, usually at the start of class. A minimum of 5 one-page quizzes will be given. No make-up/retake option is offered if you miss a quiz. Lowest 3 scores will not be counted.
Labs 10% See deliverables column above.
Project work 60% March 16 (15%), March 31 (20%), April 22 (25%)

Week 1

Tuesday (Jan 13)

Introduction to the Problems

Porting and Portability
  • Most software is written in a high-level language which can be compiled into machine code for a specific computer architecture. In many cases, this code can be compiled for multiple architectures. However, there is a lot of existing code that contains some architecture-specific code fragments written in Assembly Language (or, in some cases, machine-specific high-level code).
  • Reasons for writing code in Assembly Langauge include:
    • Performance
    • Atomic Operations
    • Direct access to hardware features, e.g., CPUID registers
  • Most of the historical reasons for including assembler are no longer valid. Modern compilers can out-perform most hand-optimized assembly code, atomic operations can be handled by libraries or compiler intrinsics, and most hardware access should be performed through the operating system or appropriate libraries.
  • A new architecture has appeared: Aarch64, which is part of ARMv8. This is the first new computer architecture to appear in several years (at least, the first mainstream computer architecture).
  • At this point, most key open source software (the software typically present in a Linux distribution such as Ubuntu or Fedora, for example) now runs on AArch64. However, it may not run as well as on older architectures (such as x86_64).
Benchmarking and Profiling

Benchmarking involves testing software performance under controlled conditions so that the performance can be compared to other software, the same software operating on other types of computers, or so that the impact of a change to the software can be gauged.

Profiling is the process of analyzing software performance on finer scale, determining resource usage per program part (typically per function/method). This can identify software bottlenecks and potential targets for optimization.

Optimization

Optimization is the process of evaluating different ways that software can be written or built and selecting the option that has the best performance tradeoffs.

Optimization may involve substituting software algorithms, altering the sequence of operations, using architecture-specific code, or altering the build process. It is important to ensure that the optimized software produces correct results and does not cause an unacceptable performance regression for other use-cases, system configurations, operating systems, or architectures.

The definition of "performance" varies according to the target system and the operating goals. For example, in some contexts, low memory or storage usage is important; in other cases, fast operation; and in other cases, low CPU utilization or long battery life may be the most important factor. It is often possible to trade off performance in one area for another; using a lookup table, for example, can reduce CPU utilization and improve battery life in some algorithms, in return for increased memory consumption.

Most advanced compilers perform some level of optimization, and the options selected for compilation can have a significant effect on the trade-offs made by the compiler, affecting memory usage, execution speed, executable size, power consumption, and debuggability.

Build Process

Building software is a complex task that many developers gloss over. The simple act of compiling a program invokes a process with five or more stages, including pre-proccessing, compiling, optimizing, assembling, and linking. However, a complex software system will have hundreds or even thousands of source files, as well as dozens or hundreds of build configuration options, auto configuration scripts (cmake, autotools), build scripts (such as Makefiles) to coordinate the process, test suites, and more.

The build process varies significantly between software packages. Most software distribution projects (including Linux distributions such as Ubuntu and Fedora) use a packaging system that further wraps the build process in a standardized script format, so that different software packages can be built using a consistent process.

In order to get consistent and comparable benchmark results, you need to ensure that the software is being built in a consistent way. Altering the build process is one way of optimizing software.

Note that the build time for a complex package can range up to hours or even days!

Course Projects

In this course, you will:

  1. Test the performance of all or part of the LAMP stack.
  2. Determine how the performance of one specific part of the stack could be improved, either specifically on AArch64 systems (without detriment to other systems) or on all systems.
  3. Prepare a fix/patch for the software to implement that performance improvement.
  4. Prove that your changes do not cause any unacceptable performance regressions (Note: there might be performance regressions which are considered acceptable!).
  5. Upstream your Code - Submit your code to the upstream (originating) software project so that it can be incorporated into future versions of the software. This will involve going through a code review to ensure that your code is compatible with and acceptable to the upstream community.

General Course Information

  • Course resources are linked from the CDOT wiki, starting at http://zenit.senecac.on.ca/wiki/index.php/SPO600 (Quick find: This page will usually be Google's top result for a search on "SPO600").
  • Coursework is submitted by blogging.
  • Quizzes will be short (1 page) and will be held without announcement at any time. Your lowest three quiz scores will not be counted, so do not worry if you miss one or two.
  • Course marks (see Weekly Schedule for dates):
    • 60% - Project Deliverables
    • 20% - Communication (Blog and Wiki writing)
    • 20% - Labs and Quizzes (10% labs - completed/not completed; 10% for quizzes - lowest 3 scores not counted)
  • All classes will be held in an Active Learning Classroom -- you are encouraged to bring your own laptop to class. If you do not have a laptop, consider signing one out of the Learning Commons for class, or using a smartphone with an HDMI adapter.
  • For more course information, refer to the SPO600 Weekly Schedule (this page), the Course Outline, and SPO600 Course Policies.

Discussion of how open source communities work

Thursday (Jan 15)

  • Benchmarking
  • Profiling

Week 1 Deliverables

  1. Set up your SPO600 Communication Tools - in particular, set up a blog and add it to Planet CDOT (via the Planet CDOT Feed List).
  2. Add yourself to the Winter 2015 SPO600 Participants page (leave the projects columns blank).
  3. Generate a pair of keys for SSH and email the public key to your professor.
  4. Sign and return the Open Source Professional Option Student Agreement.
  5. Optional but recommended: Set up a personal Fedora system.

Week 2

Tuesday (Jan 20)

Idea.png
Bring Your Laptop
Classes are held in a Active Learning Classroom. If you have a laptop or other device with a VGA or HDMI output (such as a smartphone!) please bring it. You'll need either a local linux environment or an SSH client -- which is built-in to Linux, Mac, and Chromebook systems, and readily available for Windows, Android, and iOS devices.

Status Check

  • Introductions around the Room
    • Name
    • Program
    • Interest in the course
    • Results of gzip benchmarking and most interesting thing you observed
  • Check the server accounts that were set up with your SSH keys
  • Sheets from Last Week
    • Open Source Student Agreement
    • Survey (Optional)

Working with the Code

  • Getting Code
    • In a tarball
    • From git
      • Git basics
    • Working with other version control systems
  • Building the Code
    • Make
    • Configuration tools (autotools, cmake)
    • The compiler toolchain
      • Preprocessor
      • Compiler
      • Assembler
      • Linker
    • Debug vs. Non-debug/Stripped binaries

Looking at How Distributions Package the Code

  • Using fedpkg

Resources

Thursday (Jan 22)

Week 2 Deliverables


Week 3

Tuesday (Jan 27)

Profiling

  • Profiling with gprof
    • Build with profiling enabled (use the option -pg with both gcc and ld)
    • Run the profile-enabled executable
    • Analyze the data in the gmon.out file
      • gprof nameOfBinary # Displays text profile including call graph
      • gprof nameOfBinary | gprof2dot | dot | display - # Displays visualization of call graph
  • Other profiling tools
    • OProf, SystemTap, and others
  • SPO600 Profiling Lab (Lab 3)

Thursday (Jan 29)

Week 3 Deliverables


Week 4

Tuesday (Feb 3)

Thursday (Feb 5)

Week 4 Deliverables

Week 5

Week 5 Deliverables

Week 6

Tuesday (Feb 17)

Thursday (Feb 19)

Week 6 Deliverables

  • Blog about your presentation.

Week 7

Tuesday (Feb 24)

Thursday (Feb 26)

Week 7 Deliverables

  • Blog your conclusion to the assembly language lab (Lab 5) - See the Deliverables section in the lab instructions for details on what to include in your blog post.

Week 8

Tuesday (Mar 10)

  • Project discussion
    • Discussion of project ideas and fine-tuning of project plans
    • Next steps
    • Addressing Problem Areas
  • Accessing and Reading Reference Documentation
    • AArch64 ISA

Thursday (Mar 12)

  • Project presentations
    • Provide a very short (2 minute) overview of your project. Include:
      • Which piece of software you are working on
      • What area of that software needs optimization/performance tuning
      • How you are going to perform the optimization/tuning (algorithm replacement, platform-specific code, removing platform-specific code, build options, and so forth)
      • Plans for the next step
      • Engagement with the upstream community
  • Use this project presentation to tell the class what you're working on and incorporate feedback into your blog post

Week 8 Deliverables

Week 9

Tuesday (Mar 17)

  • Upstreaming
    • Git Discussion
    • Creating a Patch
    • Various open source community workflows

Thursday (Mar 19)

  • Spinlocks
    • The need for Atomics in Spinlocks
    • Using Intrinsics for Atomic Operations
  • Tail Call Optimization (TCO)
    • gcc will perform TCO at -O2 and higher
    • Important to ensure that gcc recognizes code patterns that permit TCO to be applied

Week 9 Deliverables

  • Blog at least once (and ideally more than once) about your Project
    • Frequent shorter posts are better than rare long ones
    • Aim to make steady progress on your project, a bit each day
    • Blog about your evolving project plan and the steps you're making along the way
    • Keep the work short, and focus on narrowing the scope of work as early as possible
    • Keep talking to the community

Week 10

Tuesday (Mar 24)

  • Discussion of GCC Intrinsics
    • Where to find documentation: GCC Manual
    • __atomic vs __sync intrinsic families
  • Examination of Build Files
    • Differential analysis of successful and unsuccessfull build

Thursday (Mar 26)

  • Project Presentations - Stage 2
    • At this point, you should have a proposed patch that you're working to get upstream

Week 10 Deliverables

  • Blog posts about your project
    • Include information about your patch(es)
    • Include links to discussion about the patch(es) with the community (e.g., link to bug/issue tracker entries or email archives)
    • If you don't have patches and/or have not pushed them upstream yet, describe the state of your project and your plan to get changes upstream.
  • Blogs will be marked Monday for Stage 2 project work (20%)

Week 11

Tuesday (Mar 31)

Your professor will be away. You are welcome to use the classroom for a project hacking time and collaboration with your colleagues.

Thursday (Apr 2)

  • Single Instruction / Multiple Data (SIMD)
    • Brief look at x86_64 & AArch64 implementations of SIMD

Week 11 Deliverables

  • Blog about your ongoing project work. Please post at least 1-2 entries per week.

Week 12

Tuesday (Apr 7)

  • Discussion & Hacking Session

Thursday (Apr 9)

  • Vectorization
    • Using Intrinsics
    • Using Auto-Vectorization

Week 12 Deliverables

  • Blog about your ongoing project work. Please post at least 1-2 entries per week.
  • You should have patches (code, makefiles, documentation, or tests) in the upstream review process at or before this point.
  • Blog about vectorization - specifically:
    • What -O level and/or options are needed to turn on the auto-vectorizer in GCC
    • At least three limitations -- conditions under which the vectorization will not be performed
    • The significance of alignment and non-overlapping pointers for vectorization

Week 13

Tuesday (Apr 14)

  • Informal Project Presentations - Phase 3

Thursday (Apr 16)

  • Wrap-up Session

Week 13 Deliverables

  • Blog about your ongoing project work. Your patches should be through the upstream review process.
  • Final date for posting about your project: April 22.