77 lines
3.2 KiB
Org Mode
77 lines
3.2 KiB
Org Mode
** Changelog
|
|
|
|
*** General
|
|
- [x] Change name (Thinking of using (Fat Address Translation) -> similar to GNU))
|
|
- [x] Avoid words which create options
|
|
- [x] Use citation with ```~/cite{foo}```
|
|
- [x] Clear distinction on Literature review and contribution done
|
|
- [x] Generalise capitalisation
|
|
- [x] Instead of "Leverage" -> "use"
|
|
- [ ] Remove whitespace from figures to be compressed
|
|
- [ ] Uniform font sizes in figures
|
|
- [ ] Symmetrical lines on the figure
|
|
- [x] TLB with acronym only once for the first instance in the paper
|
|
|
|
*** Abstract (Re-read required later on)
|
|
- [x] Reduce Abstract to a single paragraph'
|
|
- [x] "Mitigate gap, capacity" -> Expand on for instance on how is it measured.
|
|
- [x] Huge pages moved to the introduction.
|
|
- [ ] Robs proposed structure (This paper presents <NAME> user-space memory allocator. It uses capability-based addressing to <list properties>.
|
|
|
|
*** Introduction
|
|
- [x] "It becomes possible" expand more with baseline compressed bounds (Relook into this later on).
|
|
- [x] "This paper makes the following contribution" -. bullet points with marked sections.
|
|
- [x] Research questions removed.
|
|
- [x] "This implementation" Specify implementation
|
|
- [ ] Robs recommended Introduction structure (Expansion of the abstract, Name of the existing system, clearly describe limitation, Introduce memory allocator).
|
|
|
|
*** 2.1 Encoding ranges as bounds
|
|
- [x] Repetition "Tracking of memory ranges using information in pointers" -> remove this
|
|
- [x] Paraphrase last sentence (To be looked into later on again).
|
|
|
|
*** 2.2 128 bit compressed bounds
|
|
- [x] Cites Jemalloc paragraph.
|
|
- [x] Last sentence needs rewriting. (Needs re-check)
|
|
|
|
*** 2.3 Instrumenting block based allocator
|
|
- [x] Needs citation "Which increases overhead and adds complexity"
|
|
- [x] Elaborate on "this method"
|
|
- [x] "precise" memory management (Elaborate on this and prove which aspects of the other allocator are not precise). (Todo be looked later on).
|
|
|
|
*** Memory allocator design
|
|
- [x] Remove from heading "sample" and ":"
|
|
- [x] Remove future work paragraph
|
|
- [x] "unmapped" -> unmap
|
|
- [x] "Risks of memory leaks with metadata lookups or complex data structures" -> Elaborate on this (Commented until further notice).
|
|
|
|
*** Evaluation
|
|
- [x] Name of the project keeps the sentence shorter.
|
|
- [x] "Such as" -> Definitive and exhaustive.
|
|
- [x] Section reference on evaluation expanded.
|
|
- [x] Remove "Providing a solid foundation for meaningful comparisons"
|
|
- [x] Decide to use Macro or instead "real world C programs"
|
|
|
|
- [x] Experiment section typo heading
|
|
- [x] Fix Latex benchmarks structure
|
|
- [x] Change from "Diagram" to "Figure"
|
|
|
|
- [x] Result para 2 -> too wordy
|
|
- [x] "aimed" first use of past tense (Recommendation: avoid)
|
|
- [x] Rewrite usability (Changed to Analysis)
|
|
|
|
*** Related work
|
|
- [x] Related work lifted up after introduction based on the reflection of the feedback.
|
|
|
|
*** Conclusion
|
|
- [x] Rewrite
|
|
|
|
*** Experiment work
|
|
- [x] 3 line switch in as mmap in Jemalloc for with sample C program.
|
|
- [x] Ported Libc ```shm_large_page``` to a kernel module to ensure multiple large scale allocators can use our implementation.
|
|
- [ ] Testing the Ported Libc kernel module (Micro benchmark for TLB stress test)
|
|
- [ ] Port this over to Jemalloc
|
|
|
|
|
|
|
|
|