AES ARM slow on M1

[Sam Van Gucht]

Hi everyone,

We noticed AES GCM in significant slower on M1 compare to an Intel Mac. Encrypting a 100mb files takes ~130ms on Intel Mac & while on M1 it takes ~510ms.

Are the ARMv8 Cryptographic instructions not supported?

Best regards,

Sam

[Jeffrey Walton]

The Clang compiler provided by Apple is crippled. It does not provide the intrinsics needed for the multiplies.

The fix is to use inline assembly. Apple compilers support the mnemonics.

I don't have access to an Apple ARMv8 and M1 machines any longer. I donated the one I bought to the GCC Compile Farm. They loaded Debian on it.

Jeff

[Uri Blumenthal]

If you have the code with  inline assembly for MacOS M1, I can test it. 

On Apr 25, 2023, at 22:39, Jeffrey Walton <nolo...@gmail.com> wrote:



--
You received this message because you are subscribed to the Google Groups "Crypto++ Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cryptopp-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cryptopp-users/56d9db2e-ab96-4259-817a-c767141a80b7n%40googlegroups.com.