https://git.ruby-lang.org/ruby.git/commit/?id=511b55bcef

From 511b55bcefc81c036294dc9a544d14bd342acd3b Mon Sep 17 00:00:00 2001
From: AGSaidi <AGSaidi@u...>
Date: Thu, 13 Aug 2020 12:15:54 -0500
Subject: Enable arm64 optimizations that exist for power/x86 (#3393)

* Enable unaligned accesses on arm64

64-bit Arm platforms support unaligned accesses.

Running the string benchmarks this change improves performance
by an average of 1.04x, min .96x, max 1.21x, median 1.01x

* arm64 enable gc optimizations

Similar to x86 and powerpc optimizations.

|       |compare-ruby|built-ruby|
|:------|-----------:|---------:|
|hash1  |       0.225|     0.237|
|       |           -|     1.05x|
|hash2  |       0.110|     0.110|
|       |       1.00x|         -|

* vm_exec.c: improve performance for arm64

|                               |compare-ruby|built-ruby|
|:------------------------------|-----------:|---------:|
|vm_array                       |     26.501M|   27.959M|
|                               |           -|     1.06x|
|vm_attr_ivar                   |     21.606M|   31.429M|
|                               |           -|     1.45x|
|vm_attr_ivar_set               |     21.178M|   26.113M|
|                               |           -|     1.23x|
|vm_backtrace                   |       6.621|     6.668|
|                               |           -|     1.01x|
|vm_bigarray                    |     26.205M|   29.958M|
|                               |           -|     1.14x|
|vm_bighash                     |    504.155k|  479.306k|
|                               |       1.05x|         -|
|vm_block                       |     16.692M|   21.315M|
|                               |           -|     1.28x|
|block_handler_type_iseq        |       5.083|     7.004|
|                               |           -|     1.38x|

diff --git a/gc.c b/gc.c
index 22972df..788f06f 100644
--- a/gc.c
+++ b/gc.c
@@ -1115,6 +1115,19 @@ tick(void) https://github.com/ruby/ruby/blob/trunk/gc.c#L1115
     return val;
 }
 
+#elif defined(__aarch64__) &&  defined(__GNUC__)
+typedef unsigned long tick_t;
+#define PRItick "lu"
+
+static __inline__ tick_t
+tick(void)
+{
+    unsigned long val;
+    __asm__ __volatile__ ("mrs %0, cntvct_el0", : "=r" (val));
+    return val;
+}
+
+
 #elif defined(_WIN32) && defined(_MSC_VER)
 #include <intrin.h>
 typedef unsigned __int64 tick_t;
diff --git a/gc.h b/gc.h
index 6568079..47a4ca1 100644
--- a/gc.h
+++ b/gc.h
@@ -8,6 +8,8 @@ https://github.com/ruby/ruby/blob/trunk/gc.h#L8
 #define SET_MACHINE_STACK_END(p) __asm__ __volatile__ ("movl\t%%esp, %0" : "=r" (*(p)))
 #elif defined(__powerpc64__) && defined(__GNUC__)
 #define SET_MACHINE_STACK_END(p) __asm__ __volatile__ ("mr\t%0, %%r1" : "=r" (*(p)))
+#elif defined(__aarch64__) && defined(__GNUC__)
+#define SET_MACHINE_STACK_END(p) __asm__ __volatile__ ("mov\t%0, sp" : "=r" (*(p)))
 #else
 NOINLINE(void rb_gc_set_stack_end(VALUE **stack_end_p));
 #define SET_MACHINE_STACK_END(p) rb_gc_set_stack_end(p)
diff --git a/include/ruby/internal/config.h b/include/ruby/internal/config.h
index 49f673e..0193275 100644
--- a/include/ruby/internal/config.h
+++ b/include/ruby/internal/config.h
@@ -103,6 +103,8 @@ https://github.com/ruby/ruby/blob/trunk/include/ruby/internal/config.h#L103
 # define UNALIGNED_WORD_ACCESS 1
 #elif defined(__powerpc64__)
 # define UNALIGNED_WORD_ACCESS 1
+#elif defined(__aarch64__)
+# define UNALIGNED_WORD_ACCESS 1
 #elif defined(__mc68020__)
 # define UNALIGNED_WORD_ACCESS 1
 #else
diff --git a/regint.h b/regint.h
index a2f5bbb..0740429 100644
--- a/regint.h
+++ b/regint.h
@@ -52,7 +52,7 @@ https://github.com/ruby/ruby/blob/trunk/regint.h#L52
 #ifndef UNALIGNED_WORD_ACCESS
 # if defined(__i386) || defined(__i386__) || defined(_M_IX86) || \
      defined(__x86_64) || defined(__x86_64__) || defined(_M_AMD64) || \
-     defined(__powerpc64__) || \
+     defined(__powerpc64__) || defined(__aarch64__) || \
      defined(__mc68020__)
 #  define UNALIGNED_WORD_ACCESS 1
 # else
diff --git a/siphash.c b/siphash.c
index 153d2c6..ddf8ee2 100644
--- a/siphash.c
+++ b/siphash.c
@@ -30,7 +30,7 @@ https://github.com/ruby/ruby/blob/trunk/siphash.c#L30
 #ifndef UNALIGNED_WORD_ACCESS
 # if defined(__i386) || defined(__i386__) || defined(_M_IX86) || \
      defined(__x86_64) || defined(__x86_64__) || defined(_M_AMD64) || \
-     defined(__powerpc64__) || \
+     defined(__powerpc64__) || defined(__aarch64__) || \
      defined(__mc68020__)
 #   define UNALIGNED_WORD_ACCESS 1
 # endif
diff --git a/st.c b/st.c
index c11535e..8be466b 100644
--- a/st.c
+++ b/st.c
@@ -1662,7 +1662,7 @@ st_values_check(st_table *tab, st_data_t *values, st_index_t size, https://github.com/ruby/ruby/blob/trunk/st.c#L1662
 #ifndef UNALIGNED_WORD_ACCESS
 # if defined(__i386) || defined(__i386__) || defined(_M_IX86) || \
      defined(__x86_64) || defined(__x86_64__) || defined(_M_AMD64) || \
-     defined(__powerpc64__) || \
+     defined(__powerpc64__) || defined(__aarch64__) || \
      defined(__mc68020__)
 #   define UNALIGNED_WORD_ACCESS 1
 # endif
diff --git a/vm_exec.c b/vm_exec.c
index ce2e053..7aa56f6 100644
--- a/vm_exec.c
+++ b/vm_exec.c
@@ -57,6 +57,9 @@ static void vm_insns_counter_count_insn(int insn) {} https://github.com/ruby/ruby/blob/trunk/vm_exec.c#L57
 #elif defined(__GNUC__) && defined(__powerpc64__)
 #define DECL_SC_REG(type, r, reg) register type reg_##r __asm__("r" reg)
 
+#elif defined(__GNUC__) && defined(__aarch64__)
+#define DECL_SC_REG(type, r, reg) register type reg_##r __asm__("x" reg)
+
 #else
 #define DECL_SC_REG(type, r, reg) register type reg_##r
 #endif
@@ -93,6 +96,11 @@ vm_exec_core(rb_execution_context_t *ec, VALUE initial) https://github.com/ruby/ruby/blob/trunk/vm_exec.c#L96
     DECL_SC_REG(rb_control_frame_t *, cfp, "15");
 #define USE_MACHINE_REGS 1
 
+#elif defined(__GNUC__) && defined(__aarch64__)
+    DECL_SC_REG(const VALUE *, pc, "19");
+    DECL_SC_REG(rb_control_frame_t *, cfp, "20");
+#define USE_MACHINE_REGS 1
+
 #else
     register rb_control_frame_t *reg_cfp;
     const VALUE *reg_pc;
-- 
cgit v0.10.2


--
ML: ruby-changes@q...
Info: http://www.atdot.net/~ko1/quickml/