If automatically generated code is better than hand-written code, you suck at handwriting code

I disagree. You'd be surprised how the compiler optimizes

If automatically generated code is better than hand-written code, you suck at handwriting code

If your hand generated code is better than the one made by the compiler, you are coding in the wrong place

and if you're going to claim that plain asm does everything just as well no it doesn't

so is overly using stackframes and not using the correct opcodes :D

"premature optimization is the root of all evil" Measure it, if it is to slow, improve it. There are so many things on modern CPU's you don't even think of

and if you're going to claim that plain asm does everything just as well no it doesn't

this one?

I think you were the one that said we should not insult anyone

i'm not insulting him

But his claim is a bit sketchy in my eyes

@florianbAT you'll get another ot warning soon

Ok this gone #ot quickly

But his claim is a bit sketchy in my eyes

Then show me the readability of an asm multiple binary search that is as fast as, say, one made using C++20 coroutines

@florianbAT you'll get another ot warning soon

the latest one ( ͡° ͜ʖ ͡°)

But his claim is a bit sketchy in my eyes

your "argument" (that picture) totally isn't sketchy at all

it makes perefect sense

Then show me the readability of an asm multiple binary search that is as fast as, say, one made using C++20 coroutines

all he said is that asm cant do everything as well as C++ also, asm can be readable if you use it well

all he said is that asm cant do everything as well as C++ also, asm can be readable if you use it well

readable, but then loop is slower than manually decrementing registers

Then show me the readability of an asm multiple binary search that is as fast as, say, one made using C++20 coroutines

Refering to the "Nano-Coroutines" ? :)

also i haven't said what you claimed i said

readable, but then loop is slower than manually decrementing registers

i didnt imply that you should use loop

i didnt imply that you should use loop

well the superior solution is very readable

just gimme admin and you can warn me all you want friends

just gimme admin and you can warn me all you want friends

/warn doesnt think like we want our members to think

Refering to the "Nano-Coroutines" ? :)

Yep, would love to see maxi try that in pure x86_64 asm

...

Oh yes.

yay

/warn doesnt think like we want our members to think

don’t you feel like you just stabbed your teammate?

don’t you feel like you just stabbed your teammate?

He is German, I guess he says "TEIMMATE", there is no I in team but what about teim

Honestly, he received 6 warns... time for him to go

he already went through 7

/warn doesnt think like we want our members to think

he already went through 7

Wait what

@BinaryByter are you moron

😕

Wait what

got another chance

i regret giving him that chance

He is German, I guess he says "TEIMMATE", there is no I in team but what about teim

in germany, we just shoot

yeah and 100 another chances while Maxi can burn their butts thru this all and by going to ot again /ot /closed

in germany, we just shoot

but thats for OT or NSFW

Hi rose can you rectify this? But don’t say that use Void main()

Hi rose can you rectify this? But don’t say that use Void main()

Can someone tell please ??

Hi rose can you rectify this? But don’t say that use Void main()

Try casting your a to double before returning it

Hi rose can you rectify this? But don’t say that use Void main()

Your function returns an int, so the result in truncated

Your function returns an int, so the result in truncated

How to rectify

How to rectify

Coz i tried thya

That

Try casting your a to double before returning it

How ?

How to rectify

you can't return either int or double based on the input

Change it into double

Or better, to x

Change it into double

Oh th

Also I don't think that's the best use case for templates

Or better, to x

Yeah but I wanna different data type for both so i

Yeah but I wanna different data type for both so i

Double will be the safest, will suit for them all

Also I don't think that's the best use case for templates

I am learning templets topic so i asked

Also I don't think that's the best use case for templates

Why not

Double will be the safest, will suit for them all

Unless you're gonna have chars or strings there

Double will be the safest, will suit for them all

Kk thanks both of you. I posted multiple places but didn’t get anywhere

Why not

Because this template basically offers no benefits. It relies on the existance of the operator < and that the result is convertible to an int. A template would make sense to make sure both types are the same

Because this template basically offers no benefits. It relies on the existance of the operator < and that the result is convertible to an int. A template would make sense to make sure both types are the same

Yeah sure, but as he said, it's just for learning

Yeah sure, but as he said, it's just for learning

Yeah sure - nothing wrong about that :)

Because this template basically offers no benefits. It relies on the existance of the operator < and that the result is convertible to an int. A template would make sense to make sure both types are the same

And the int was a mistake

THE mistake

Because this template basically offers no benefits. It relies on the existance of the operator < and that the result is convertible to an int. A template would make sense to make sure both types are the same

Is there any other thing you know in that except conversing?

And the int was a mistake

but then double would be also wrong in my opinion..

but then double would be also wrong in my opinion..

For this specific case it'll be enough

For this specific case it'll be enough

double might loose precision

but then double would be also wrong in my opinion..

I think so coz i used double and int here but it may be possible to use char and int Then will the double work ??

double might loose precision

Why is that

I think so coz i used double and int here but it may be possible to use char and int Then will the double work ??

No because the char can't be returned as double

Not in the way you expect it to

Why is that

Because double is a floating point number. The range of floating point number is way bigger (given the same amount of bits) so there must be a precision loss

No because the char can't be returned as double

Then whats the solution here

Because double is a floating point number. The range of floating point number is way bigger (given the same amount of bits) so there must be a precision loss

😍

Because double is a floating point number. The range of floating point number is way bigger (given the same amount of bits) so there must be a precision loss

If it's way bigger, shouldn't it be more accurate??

Then whats the solution here

There's no solution, what will you do if you have chars? Check if 'a' is bigger than 3? What are you getting from it?

Because double is a floating point number. The range of floating point number is way bigger (given the same amount of bits) so there must be a precision loss

I mean, if int should be returned, it will be returned as a double: 5 as 5.0. And if it's a double it'll be returned as is: 3.14 as 3.14. Where's the loss? *~*

If it's way bigger, shouldn't it be more accurate??

The difference (absolute) obviously increases, the relative error stays roughly the same. This explains how they are encoded https://en.wikipedia.org/wiki/Floating-point_arithmetic Try this to see the difference #include <iostream> template <class A, class B> double check(A && a, B && b) { if (a > b) { return a; } else { return b; } } int main() { std::cout << static_cast<long long>( check(2.0, 100000000000000013LL)) << '\n'; std::cout << 100000000000000013LL; }