Timeline for Could there be a floating point CSPRNG?
Current License: CC BY-SA 4.0
14 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 4, 2021 at 6:14 | comment | added | fgrieu♦ | @forest: "do" does for me. | |
| Jan 4, 2021 at 1:21 | comment | added | forest | @fgrieu The existing TLD no longer works. The .se one is the only one I know. What other TLD does it have that works? | |
| Jan 3, 2021 at 21:30 | comment | added | fgrieu♦ | @forest: FYI, regarding edit of the TLD of sci-hub in several questions: (a) these have the side effect of bringing the question in the "hot" list, which is seen as undesirable by many. (b) Where I am (France), sci-hub in the TLD "se" is also ordered blocked by a justice decision ISPs bowed to (minimally so: the block is at the DNS level, and others TLD still work), thus your change does not help there. | |
| Jan 3, 2021 at 1:44 | history | edited | forest | CC BY-SA 4.0 |
fixed broken links
|
| Mar 19, 2019 at 16:28 | history | edited | Squeamish Ossifrage | CC BY-SA 4.0 |
Incorporate material from comments.
|
| Mar 18, 2019 at 18:40 | history | edited | Squeamish Ossifrage | CC BY-SA 4.0 |
More bold, more citations.
|
| Jul 11, 2018 at 0:27 | comment | added | Squeamish Ossifrage |
@PaulUszak Binary64 is an IEEE 754 floating-point format, and has nothing to do with the hardware's integer register size, address space size, etc. There's no binary8 defined in IEEE 754-2008 (though in principle there could be): only binary{16,32,64,128,256} and decimal{32,64,128}. Most computers use IEEE 754 binary64 arithmetic—this is what you usually get in C with double. Whether the hardware accelerates IEEE 754 binary64 arithmetic doesn't change what the semantics guarantees. Your 8-bit microcontrollers probably don't accelerate 64-bit integer addition either.
|
|
| Jul 10, 2018 at 23:43 | comment | added | Paul Uszak | I'd forgotten about this one. What about the same algorithm on binary32 (i386, 32 bit ARM), binary16 (Atmel, Teensy) and binary8 (STAMP, 6502)? IEEE 754 binary64 $\neq$ IEEE 754 binary8? Not sure about this answer. | |
| Aug 7, 2017 at 22:44 | comment | added | Squeamish Ossifrage | @PaulUszak: For every pair of floating-point numbers $x$ and $y$, $x \oplus y$ is the floating-point number nearest to $x + y$, or, in the case of ties, the one whose significand has zero in its least significant position. This is precisely defined for any pair of floating-point number you want, and predictably the same on all IEEE 754 machines, even if it's not appreciably easier to compute in my head than MD5("Paul Uszak"). See the functions $\operatorname{topleft}_{k,b} z$ and $\operatorname{bottomleft}_{k,b} z$ and Theorems 2.3, 2.4, 2.5 for examples of using this to mask off bits. | |
| Aug 7, 2017 at 22:24 | comment | added | Paul Uszak | I don't understand your predictable rounding error. What is the error prediction formula? For a very simple example, what is the total of 1 million 0.1s without programming it? And what would I do for a more complex example? | |
| Aug 7, 2017 at 22:09 | comment | added | Squeamish Ossifrage | @EllaRose: I added a couple references. Couldn't find any references specifically about the consequences for timing variability of arithmetic, but it's obvious—even aside from performing identity vs. value equality tests, since Python integers are boxed the value of the integer leaks into the address of the object you must load from memory, and it is different depending on whether it's a small integer or a large integer. Does that help? | |
| Aug 7, 2017 at 22:07 | history | edited | Squeamish Ossifrage | CC BY-SA 3.0 |
Add citations about leakiness of Python small integer cache. Move universal hashing citation to bottom and add mouseover text.
|
| Aug 7, 2017 at 21:18 | comment | added | Ella Rose | "...such as Python in which integer arithmetic is notoriously leaky due partly to the small integer cache (if for some reason you couldn't just write a C extension)." - Do you happen to have a link or more information on this? I'd love to read more about it. | |
| Aug 7, 2017 at 19:32 | history | answered | Squeamish Ossifrage | CC BY-SA 3.0 |