|
|
|
|
|
by MobiusHorizons
174 days ago
|
|
> Conclusion: Coderive doesn't just make loops faster—it redefines what's computationally possible on commodity hardware. I mean this as kindly as possible, but please don’t say things like this if you want to be taken seriously. Computer languages cannot possibly change what is possible on a given machine for the simple reason that whatever they are doing had to previously be possible in assembly on the same machine. I don’t mean to overly discourage you. Lazy execution can be very useful, but it’s also not clearly new or hard to get in other languages (although it would require different syntax than an idiomatic for loop most of the time). It may help to try to pick an example where the lazy execution is actually exercised. Preferably one that would be hard for an optimizing compiler to optimize. I would also not recommend claiming iteration if you also claim 50ms, since that’s clearly impossible regardless of memory consumption, so you have to optimize away or defer the work in some way (at which point iteration is no longer occurring). For these examples, I think you would just express the code as a function taking i instead of pre-populating the array. This doesn’t seem hard at least for the provided examples, and has the benefit that it can be opted into when appropriate. |
|
Before Coderive: To explore 1 trillion cases, you'd need:
· A cluster of machines · Distributed computing framework (Spark/Hadoop) · More time for setup
With Coderive:
```java results := [1 to 1T] // Conceptually 1 trillion for i in results { results[i] = analyzeCase(i) } // Check interesting cases immediately ```
It's not about computing faster than physics allows. It's about thinking and exploring with ease without infrastructure constraints.