What books would you recommend to learn more about cache locality, branch prediction, and other optimizations that'll help me understand performance at a low level?I'm looking for books on performance for concurrency/parallelism or low level networking too.
>>101235044bump for interest
>>101235321Fuck off, OP posted a vtumor. He made his bed so now he should be forced to lie in it.
>>101235373ahim so far removed from normalspace i didnt even recognizei guess herbs then
>>101235044literally think for one second on your own, if you get caught in the trap of having to read a book for everything you are never going to move forwardsI made so many assumptions you have no idea, did many of them turn out to be wrong? yeah, did i feel humiliateed? yes, learning to program is literally a humiliation ritual. Was i right other times? yes, and it saved me from making retarded threads and spending weeks and weeks procrastinating on some shitty pdf.
>>101235425if you dont know where to start its a mountain to climb.>>101235044i heard this guy is da bombhttps://www.agner.org/optimize/#manualsbut like the other anon said i procrastinate on starting readingalso i learned in an unstructured way so i know a good portion of that already which demotivates me even morebut yeah, that might be a starting pointyou can learn opencl or cuda too.i learned opencl.having to manage memory locality manually was an eye-opener for me.and the whole semi-asynchronous execution model is a good mental exercise too
>>101235425I'm working on low latency systems that aren't easy to experiment with. I find books give me a nice overview of subject areas which I can dig into deeper as needed.>>101235500I'm not a fan of straight up reading books either. I get my motivation from thinking about how I can apply what I learned to a project.
>>101235044Software Developer Manual of your target CPU architecture, if you care about cutting edge.University course materials are always 10 years behind and textbooks are 20 years behind.
>>101235500i read programming from the ground up, realized i wasted my time, reverse engineered how an entire kernel works by thinking and then lurked more to hear and learn about optimizations in processors.I guess its 2024...
>>101235557this.
>>101235373>knowing vtumors by heart and instantly recognizing themsir, I'm going to need you to go back...
>>101235555>how I can apply what I learned to a project.apparently its the best method of learning.bootcamps, paying sites, even teachers in the classical curriculum apply this method. or try to.>>101235704based.with practice, i learned bits here and there until an image emerged, then clarified with timeits a mental model that works at my decently advanced levelbut i know theres stuff i dont knowi will have to read the book at some point to make sure i have no holes in my knowledge anymore + learn the proper wording