Timeline for F'up: Arbitrary large unsigned integers
Current License: CC BY-SA 3.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 7, 2017 at 7:28 | comment | added | Felix Palmen |
@mvds: the typedefis already there in the newest revision :) Thanks for the further idea, this indeed sounds interesting!
|
|
| Jun 7, 2017 at 1:20 | comment | added | mvds |
It could be interesting to see if you can use SIMD instructions and/or registers, if performance is a concern. From my little experience in this field, long ago, I remember this: you can either hand-craft C code with SIMD intrinsics, or provide gentle nudges (in code) and flags to the compiler so that it starts using them. Such a nudge would be to make any array a multiple of 8 elements, and make that explicit in your code, e.g. so that a loop can process elements in groups of 8. (i.e. for ( int i=0;i<count&~7;i++ ) { } signals to the compiler that the loop always runs a multiple of 8 times)
|
|
| Jun 7, 2017 at 1:08 | comment | added | mvds | Why not typedef the "base" type and make it so that you can easily profile which type works best on your current hardware? | |
| Jun 6, 2017 at 20:54 | answer | added | chux | timeline score: 1 | |
| Jun 6, 2017 at 8:54 | answer | added | JS1 | timeline score: 1 | |
| Jun 4, 2017 at 16:51 | history | asked | Felix Palmen | CC BY-SA 3.0 |