I am designing some double-double and quad-double arithmetic libraries for C++11. Part of the code is based off of the QD and DDFUN library from https://www.davidhbailey.com/dhbsoftware/.
struct Float64x2 {
double hi;
double lo;
};
struct Float64x4 {
double val[4];
};
Dekker-float ty...
I'm pretty new to x64 programming and tried to recreated parts of head for study (targeting linux).
Are there any stylistic issues?
The head function uses a buffer in the bss section for IO, but I've also seen people allocate buffers on the stack for similar purposes.
Is there a benefit to alloca...
You are trying to optimize code that already works. Maybe try Code Review with a complete example of your code and a more thorough explanation of what you're trying to accomplish. — Ignatius Reilly27 secs ago
I have a Polars dataframe in which the cells contain a sequence of single digits as a string of characters, and I want to find the number of differences between the single-digits elements of the string. For example:
df = pl.DataFrame({"pop_1": ["100","0021"],"pop_2":["11002","0000",]})
| pop_1 |...
I have the following code to amend two rows of "test_base.csv" with the entries of the arrays "a_temp" and "b_temp," saving the result into "result.csv." .csv files are here. It is working, but I'd like it to be faster. The total number of columns to modify is dim*dim, and dim = 100 takes about 6...
Your design code does not follow good coding practices. If you are interested in getting feedback on it, consider posting on Code Review. Make sure to read the policy first. — toolic10 secs ago
Class<T> is a bad idea. It's a two-way lie. It cannot represent arbitrary T - often code that needs to take in a typeref does not know if the type ref you want to take in is generics-free, and Class<T> cannot represent any type that contains generics. But, it's even worse - j.l.Classcan represent primitives; int.class is a thing. But generics cannot. So, it's just a dumb fucking idea, and Class<T> as a parameter is pretty much an automatic veto in code review from me. There are always better answers. In rare cases, it's supertypetokens like here. In most others, factories. — rzwitserloot5 secs ago
@NFeruch-FreePalestine The reasons for how we got into this messy situation are not very important. Yes a developer made some mistakes, what we want to do is catch those mistakes before we merge the PR, so that more junior devs can catch their own mistakes right as they create/modify the PR, without needing the code reviewer to manually check the branchs are ok. The feature branch wasn't resurected, the dev just made some mistakes while rebasing/pulling I guess, and pushed the result w/o checking. — frozenkoijust now
This a cnn in Lua, the reason I’m not using torch, or any other external library, is because they take up too much space, and I can’t import them to external services like Roblox. The code works fine! But I’m looking to see if it can be further optimized.
local unpack = table.unpack
local floor, ...
I am designing some double-double and quad-double arithmetic libraries for C++11. Part of the code is based off of the QD and DDFUN library from https://www.davidhbailey.com/dhbsoftware/.
struct Float64x2 {
double hi;
double lo;
};
struct Float64x4 {
double val[4];
};
Dekker-float ty...
