ALLVM: Forthcoming Project to Ship All Software As LLVM IR

[dagger]

Will it work on z/OS? I can keep dreaming...

[zamadatix]

There are many language implementations other than Java and C# that already target the Java and .NET VMs.

Yes, but then you are constrained to being inside the Java and .NET VMs along with their respective limitations on GC and JIT. This can run native with no GC, JIT, or other runtime tasks - You could give someone 5 of your projects composed of 3 languages each and for that person to have full native optimization like they had your entire build environment for each language (which may or may not have all dependencies available on their environment) they instead only need ALLVM and the files you gave them.

[GrayShade]

Supposedly, the LLVM IR is not really a good representation for shipping code.

[pal666]

This is a pretty cool idea. Write in any language you want and let the end user choose if they want to compile it, interpret it or interpret it with with a JIT compiler.

they do not for interpretation, but for compilation with optimization for user's cpu. like source-based distros but without need for development libraries

[pal666]

Because you can do link-time-optimizations. Meaning you optimize natively to your microcode.

you could also do link-time optimizations on developer machine. on user machine you could do optimization for user machine's cpu

[pal666]

The concept behind the idea is nothing new - Java, MSIL, etc. but using LLVM's IR, you can use any language you want, you're not constrained to Java, C# or whatever.

with any byte code you could use any language you want with compiler targeting this bytecode

[pal666]

i guess gcc -flto -fno-fat-lto-objects is better suited for this

[dragorth]

How would this handle moving a hard disk from one cpu to another? What benefit does this have for the user?

[peppercats]

Because you can do link-time-optimizations.

…so can the people compiling the packages.

Meaning you optimize natively to your microcode.

Distros moving away from 2003 would be more beneficial than this. How many people aren't using at least a CPU with at least SSE 4.2 instructions available right now? Steam hardware survey for Linux says 84% of surveyed computers have SSE 4.2. Instead of targeting the bare minimum x86-64 CPU when compiling packages, move the minimum up a few years from 2002. SSE4.2 is almost a decade old.

[Tomin]

Distros moving away from 2003 would be more beneficial than this. How many people aren't using at least a CPU with at least SSE 4.2 instructions available right now? Steam hardware survey for Linux says 84% of surveyed computers have SSE 4.2. Instead of targeting the bare minimum x86-64 CPU when compiling packages, move the minimum up a few years from 2002. SSE4.2 is almost a decade old.

My desktops are Core 2 Duo and Phenom II X4, neither of them has SSE4.2. Latter doesn't even have SSSE3. Maybe I should upgrade, but the performance has been all right and for now all the software I've used has been working fine. What kind of programs should use those SSE4.2 instructions?