Support AVX/AVX2/FMA/F16C on x86-32 Linux

[luke-jr]

This works:

--- a/Makefile
+++ b/Makefile
@@ -53,7 +53,7 @@ endif
 # Architecture specific
 # TODO: probably these flags need to be tweaked on some architectures
 #       feel free to update the Makefile for your architecture and send a pull request or issue
-ifeq ($(UNAME_M),x86_64)
+ifeq ($(UNAME_M),i686)
        ifeq ($(UNAME_S),Darwin)
                CFLAGS += -mfma -mf16c
                AVX1_M := $(shell sysctl machdep.cpu.features)

autotools would probably do it better...

[JohnnyOpcode]

I'm wondering if the Intel compilers in OneAPI can help squeeze even more performance. They work across Windows, Linux and Intel-based MacOS.

[luke-jr]

They're presumably not open source, so I at least wouldn't trust them...