Hacker News Reader

Comment by nwlieb | Hacker News Reader

Is this a 1-1 comparison? If the ARM compile is compiling to ARM binaries then there might be less work/optimizations since it is a newer architecture. Seems like a test with two variables that changed. Would be interesting to see them both cross-compile to their respective opposite archs.

View on HN

Maybe not, but A) it's close-- most of the work of compiling is not microarchitecture-level optimizations or emitting code, and B) if you're a developer, even if some of the advantage is being on an architecture that it's easier to emit code for... that's still a benefit you realize.

It's worth noting that cross-compiling is definitely harder in many ways, because you can't always evaluate constant expressions easily at compile-time in the same way your runtime code will, etc, too, and have to jump through hoops.

As someone who knows relatively little about this, I'm very curious why this is downvoted. It seems like a rebuttal would be enlightening.

Hm my experience was that compiling C on arm was always super fast compared to x86, because the latter had much more work to do.

This doesn't align with my experience. Clang is about the same, but GCC often seems much slower emitting cross-ARM code.

  jar% time x86_64-linux-gnu-gcc --std=c99 -O3 -c insgps14state.c -I inc -I ../../shared/api
  x86_64-linux-gnu-gcc --std=c99 -O3 -c insgps14state.c -I inc -I   0.97s user 0.02s system 99% cpu 0.992 total
  jar% time x86_64-linux-gnu-gcc --std=c99 -O3 -c insgps14state.c -I inc -I ../../shared/api
  x86_64-linux-gnu-gcc --std=c99 -O3 -c insgps14state.c -I inc -I   0.93s user 0.03s system 99% cpu 0.965 total
  jar% time x86_64-linux-gnu-gcc --std=c99 -O3 -c insgps14state.c -I inc -I ../../shared/api
  x86_64-linux-gnu-gcc --std=c99 -O3 -c insgps14state.c -I inc -I   0.94s user 0.01s system 99% cpu 0.947 total
  jar% time x86_64-linux-gnu-gcc --std=c99 -O3 -c insgps14state.c -I inc -I ../../shared/api
  x86_64-linux-gnu-gcc --std=c99 -O3 -c insgps14state.c -I inc -I   0.92s user 0.04s system 99% cpu 0.955 total

  jar% time arm-linux-gnueabihf-gcc --std=c99 -O3 -c insgps14state.c -I inc -I ../../shared/api
  arm-linux-gnueabihf-gcc --std=c99 -O3 -c insgps14state.c -I inc -I   1.43s user 0.03s system 99% cpu 1.458 total
  jar% time arm-linux-gnueabihf-gcc --std=c99 -O3 -c insgps14state.c -I inc -I ../../shared/api
  arm-linux-gnueabihf-gcc --std=c99 -O3 -c insgps14state.c -I inc -I   1.46s user 0.03s system 99% cpu 1.486 total
  jar% time arm-linux-gnueabihf-gcc --std=c99 -O3 -c insgps14state.c -I inc -I ../../shared/api
  arm-linux-gnueabihf-gcc --std=c99 -O3 -c insgps14state.c -I inc -I   1.55s user 0.04s system 99% cpu 1.587 total
  jar% time arm-linux-gnueabihf-gcc --std=c99 -O3 -c insgps14state.c -I inc -I ../../shared/api
  arm-linux-gnueabihf-gcc --std=c99 -O3 -c insgps14state.c -I inc -I   1.44s user 0.03s system 99% cpu 1.471 total

That’s interesting. I was not cross compiling so maybe the arm system I was using was just faster.

So cross compile for RISC-V, POWER, or something else would be fair?

Apple has been optimizing the compiler for a decade for iOS.

If everything else is the same, that seems like a solid reason to prefer the ARM architecture even setting aside 1:1 comparisons. Isn't faster compilation and execution the whole point of a faster processor?

The assertion is that compilation might be faster since there are fewer optimizations, and therefore runtime would be slower.

Reply on news.ycombinator.com