4 ms·

Most popular modern languages aren't low level enough for developers to care about architectural differences unless they are trying to squeeze every last ounce

Most popular modern languages aren't low level enough for developers to care about architectural differences unless they are trying to squeeze every last ounce of performance. However, the x86 chip in your laptop and the x86 chip running on your deployment target, probably aren't similar enough to help you reach that goal. Also worth noting that ARM is becoming increasingly popular in the cloud services space, so that x86 advantage will deplete significantly with time. I have a M1 Macbook Pro with 8GB ram and 256 GB storage. I bought it mainly to play with, but I've found it surprisingly capable and I've used it for some moderate development workloads and even some gaming (Metro Exodus)! I've found it pretty fantastic for the most part, but I have to constantly make difficult decisions about exactly what I want to have running and stored on the device. For instance, running Docker rapidly consumes both memory and storage resources. I'd definitely recommend buying the Mac SKUs with 16GB memory and more storage, if you want to make it your primary workstation. I think 16GB is probably enough for most development workloads though, that's generally what I'm given for a work laptop an it's worked out okay.

eptcyka 5y ago
>However, the x86 chip in your laptop and the x86 chip running on your deployment target, probably aren't similar enough to help you reach that goal. Cross-compiling is still very buggy, code built for ARM on x86 can still just _not work_ correctly when deployed on ARM. In such cases, it really does help that my laptop is running x86. I hope running arm stuff will be similar.
- scoopertrooper 5y ago
  There is no theoretical limitation on the ability to cross-compile between architectures. There may be bugs in individual compilers or mistakes made in configuring them, but those sorts of problems are solvable. However, how often do you locally compile the binary you deploy? Every company for which I've worked has delegated that activity to a build server.
  - smoldesu 5y ago
    > how often do you locally compile the binary you deploy? Do people not debug software locally anymore? I feel like "95% of developers" are going to end up compiling and using a binary of some sort (especially if you're a developer on MacOS).

eptcyka 5y ago

>However, the x86 chip in your laptop and the x86 chip running on your deployment target, probably aren't similar enough to help you reach that goal. Cross-compiling is still very buggy, code built for ARM on x86 can still just _not work_ correctly when deployed on ARM. In such cases, it really does help that my laptop is running x86. I hope running arm stuff will be similar.

scoopertrooper 5y ago
There is no theoretical limitation on the ability to cross-compile between architectures. There may be bugs in individual compilers or mistakes made in configuring them, but those sorts of problems are solvable. However, how often do you locally compile the binary you deploy? Every company for which I've worked has delegated that activity to a build server.
- smoldesu 5y ago
  > how often do you locally compile the binary you deploy? Do people not debug software locally anymore? I feel like "95% of developers" are going to end up compiling and using a binary of some sort (especially if you're a developer on MacOS).