They have added 3 summoning methods from the last time i left. Wtf is a pendulum summon, links??/ what? i can somewhat understand synchos and xyz but card text right now on a card insane - its a frickign essay.
this is the first time i am making a IQfy thread not into a /misc/ one but a IQfy one but i kinda am interesting in some solo android games with cards like ui gi oh. i guess i will have to emulate the 5d one from 10 years ago. hearstone is getting stittier and shitter, mtg has basically no sane audience left. The japs are my only hope in gaming and they are just now starting to go moronic.
2 weeks ago
Anonymous
idk about android but I like playing legacy of the duelist, it's just fricking annoying sometimes with the trap activation prompts
2 weeks ago
Anonymous
will look it up. it looks good but not mobile. i guess i will play on the pc first.
what do we think, IQfy, was it a mistake? should sea of nodes have won instead? do we just accept that our fundamental compiler infrastructure is extremely complex and slow and always will be?
This. If you're any good at your job you only need a basic b***h 1 pass compiler. Optimizing compilers were a mistake that often break code.
optimizers make it so that you don't have to worry as much about the performance of thin abstractions (like factoring things into a separate function etc), as the cost is significantly diminished thanks to inlining+constant folding+common subexpression elimination
although i do agree that autovectorization is a waste of time, likewise for instruction selection and scheduling (I'd wager that they are not very useful on a modern arm or x64 machine, unfortunately i have not measured this yet)
even register allocation is not THAT important, I think a simple register allocator can get near 90% of optimal performance except for some pathological cases.
well, if we follow your definition, im in the process of *becoming one, then.
thence the gymbro memes
->cuz im still lifting
->im certainly not perfect yet, and prolly i will never be
if youre a real c programmer, following your definition
then i tip my hat
The hard part of a compiler is type checking. Optimizations are easy and fun in my opinion. I guess there is also the part where you need to learn every instruction in x86 but I wouldn't consider this part of writting a compiler
>there is no automatic OS porting interface as of 2024
still, LLVM is the greatest software ever written. After what I mentioned above gets written its over.
yugioh is not that important, frick off
They have added 3 summoning methods from the last time i left. Wtf is a pendulum summon, links??/ what? i can somewhat understand synchos and xyz but card text right now on a card insane - its a frickign essay.
same
this is the first time i am making a IQfy thread not into a /misc/ one but a IQfy one but i kinda am interesting in some solo android games with cards like ui gi oh. i guess i will have to emulate the 5d one from 10 years ago. hearstone is getting stittier and shitter, mtg has basically no sane audience left. The japs are my only hope in gaming and they are just now starting to go moronic.
idk about android but I like playing legacy of the duelist, it's just fricking annoying sometimes with the trap activation prompts
will look it up. it looks good but not mobile. i guess i will play on the pc first.
Blue eye white dragon is overrated.
i would argue that it depends on the point of view.
can you write llvm code? bc i cant...
C mentioned in 4chink post, must alloc my time.
We should ditch C and program in LLVM IR.
I don't care for hand-crafting all the phi nodes. That's literally stuff designed for code to generate.
what do we think, IQfy, was it a mistake? should sea of nodes have won instead? do we just accept that our fundamental compiler infrastructure is extremely complex and slow and always will be?
What do C programmers need with an optimizer?
This. If you're any good at your job you only need a basic b***h 1 pass compiler. Optimizing compilers were a mistake that often break code.
optimizers make it so that you don't have to worry as much about the performance of thin abstractions (like factoring things into a separate function etc), as the cost is significantly diminished thanks to inlining+constant folding+common subexpression elimination
although i do agree that autovectorization is a waste of time, likewise for instruction selection and scheduling (I'd wager that they are not very useful on a modern arm or x64 machine, unfortunately i have not measured this yet)
even register allocation is not THAT important, I think a simple register allocator can get near 90% of optimal performance except for some pathological cases.
idk about you, but im not THAT good.
i can write fast code with -O0
but -O3 always makes it faster
learn which UB you can use and which ones you cant?
i usually got a few UBs in my code,but it compiles just fine with -O3
>idk about you, but im not THAT good.
Are you sure you're a real C programmer?
well, if we follow your definition, im in the process of *becoming one, then.
thence the gymbro memes
->cuz im still lifting
->im certainly not perfect yet, and prolly i will never be
if youre a real c programmer, following your definition
then i tip my hat
You can't compile Clang without GCC. Checkmate atheist
only due to licensing autism
Yes LLVM is very important... For shitlangs whose makers are too incompetent to make their own compiler for.
The hard part of a compiler is type checking. Optimizations are easy and fun in my opinion. I guess there is also the part where you need to learn every instruction in x86 but I wouldn't consider this part of writting a compiler
>there is no automatic OS porting interface as of 2024
still, LLVM is the greatest software ever written. After what I mentioned above gets written its over.