Timeline for Heaps: Why is there a tradeoff between amount of space occupied (fragmentation), and speed at which operations are carried out?
Current License: CC BY-SA 3.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 1, 2018 at 19:21 | history | tweeted | twitter.com/StackSoftEng/status/991397212784680960 | ||
| May 1, 2018 at 14:43 | answer | added | Peter Green | timeline score: 2 | |
| Mar 4, 2013 at 2:44 | history | edited | user28988 |
edited tags
|
|
| Nov 21, 2011 at 4:36 | vote | accept | Dark Templar | ||
| Oct 11, 2011 at 2:05 | comment | added | Dark Templar | @delnan: still, isn't that hole better than automatically creating holes by making everything 8-byte aligned? What I mean is this: alloc 4 byte @ 0, alloc 4 byte @ 8, alloc 4 byte @ 16, etc... we already have way more fragmentation when everything is 8-byte aligned, than in the counter-example you gave... | |
| Oct 10, 2011 at 21:25 | answer | added | Pubby | timeline score: 13 | |
| Oct 10, 2011 at 21:19 | comment | added | justin | Related reading: en.wikipedia.org/wiki/Memory_alignment | |
| Oct 10, 2011 at 21:15 | comment | added | justin |
doubles are often 4 and 8 byte aligned -- it varies by platform/arch.
|
|
| Oct 10, 2011 at 20:56 | comment | added | user7043 |
How would this avoid fragmentation? alloc 16 byte @ 0; alloc 16 byte @ 32; free 16 byte @ 0; alloc 12 byte @ 0; and you got a hole in which only a 4-byte allocation can fill.
|
|
| S Oct 10, 2011 at 20:38 | history | suggested | yoozer8 | CC BY-SA 3.0 |
typos/formatting
|
| Oct 10, 2011 at 20:35 | review | Suggested edits | |||
| S Oct 10, 2011 at 20:38 | |||||
| Oct 10, 2011 at 20:31 | history | asked | Dark Templar | CC BY-SA 3.0 |