Do Precompiled M1 binaries work on Apple M2 Pro machines?

[chmanoj]

Hi @DevinTDHa

I'm trying to run the introductory Tensorflow example from https://www.tensorflow.org/jvm/install#example_program and use the precompiled binaries from the release 0.4.0-SNAPSHOT. When I run the command mvn compile exec:java, the JRE crashes with SIGILL error (Full log below). I'm running this example on a 2023 Apple M2 Pro with macOS Ventura 13.5.

I wanted to find out if these precompiled M1 binaries work for M2 processors? Or do we need to recompile them on M2 hardware separately?

Thanks!

---- Detailed log ----

[INFO] Compiling 1 source file to tensorflow-test/target/classes
[INFO] 
[INFO] --- exec:3.1.0:java (default-cli) @ hellotensorflow ---
#
# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGILL (0x4) at pc=0x000000012bc419f5, pid=14525, tid=0x000000000000a11b
#
# JRE version: Java(TM) SE Runtime Environment (8.0_311-b11) (build 1.8.0_311-b11)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (25.311-b11 mixed mode bsd-amd64 compressed oops)
# Problematic frame:
# C  [libtensorflow_framework.2.dylib+0x1f9f5]  tensorflow::monitoring::MetricDef<(tensorflow::monitoring::MetricKind)1, long long, 2>::MetricDef<char [11], char [7]>(absl::lts_20210324::string_view, absl::lts_20210324::string_view, char const (&) [11], char const (&) [7])+0x125
#
# Failed to write core dump. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited" before starting Java again
#
# An error report file with more information is saved as:
# tensorflow-test/hs_err_pid14525.log
#
# If you would like to submit a bug report, please visit:
#   http://bugreport.java.com/bugreport/crash.jsp
# The crash happened outside the Java Virtual Machine in native code.
# See problematic frame for where to report the bug.
#
[1]    14525 abort      mvn compile exec:java

[DevinTDHa]

Hi @chmanoj,

I have seen this error before, usually it has something to do with the other apple silicon derivatives (in your case M2 Pro). I'm afraid it's not compatible with this binary.

I don't have a machine to test it, you might be able to compile your own binaries.

You can try this branch but I can't guarantee that it works:

https://github.com/DevinTDHa/tensorflow-java/tree/apple-silicon-0.4.2-shaded