[1/2,ARM] Refactor costs calculation for MEM.

Message ID 1487696064-3233-2-git-send-email-charles.baylis@linaro.org
State New
Headers show
Series
  • PR61551 addressing mode costs
Related show

Commit Message

Charles Baylis Feb. 21, 2017, 4:54 p.m.
From: Charles Baylis <charles.baylis@linaro.org>


This patch moves the calculation of costs for MEM into a
separate function, and reforms the calculation into two
parts. Firstly any additional cost of the addressing mode
is calculated, and then the cost of the memory access itself
is added.

In this patch, the calculation of the cost of the addressing
mode is left as a placeholder, to be added in a subsequent
patch.

gcc/ChangeLog:

<date>  Charles Baylis  <charles.baylis@linaro.org>

        * config/arm/arm.c (arm_mem_costs): New function.
        (arm_rtx_costs_internal): Use arm_mem_costs.

Change-Id: I99e93406ea39ee31f71c7bf428ad3e127b7a618e
---
 gcc/config/arm/arm.c | 66 +++++++++++++++++++++++++++++++++-------------------
 1 file changed, 42 insertions(+), 24 deletions(-)

-- 
2.7.4

Comments

Bernhard Reutner-Fischer Feb. 23, 2017, 7:25 a.m. | #1
On 21 February 2017 17:54:23 CET, charles.baylis@linaro.org wrote:
>From: Charles Baylis <charles.baylis@linaro.org>

>


>+/* Convert fron bytes to ints.  */


s/fron/from/

>+#define ARM_NUM_INTS(X) (((X) + UNITS_PER_WORD - 1) / UNITS_PER_WORD)

>+

>+

> /* RTX costs.  Make an estimate of the cost of executing the operation

>    X, which is contained with an operation with code OUTER_CODE.

>    SPEED_P indicates whether the cost desired is the performance cost,


s/contained with /contained within /
while you are at it.

thanks,
Richard Earnshaw (lists) June 9, 2017, 1:59 p.m. | #2
On 21/02/17 16:54, charles.baylis@linaro.org wrote:
> From: Charles Baylis <charles.baylis@linaro.org>

> 

> This patch moves the calculation of costs for MEM into a

> separate function, and reforms the calculation into two

> parts. Firstly any additional cost of the addressing mode

> is calculated, and then the cost of the memory access itself

> is added.

> 

> In this patch, the calculation of the cost of the addressing

> mode is left as a placeholder, to be added in a subsequent

> patch.

> 

> gcc/ChangeLog:

> 

> <date>  Charles Baylis  <charles.baylis@linaro.org>

> 

>         * config/arm/arm.c (arm_mem_costs): New function.

>         (arm_rtx_costs_internal): Use arm_mem_costs.


I like the idea of this patch, but it needs further work...

Comments inline.

R.

> 

> Change-Id: I99e93406ea39ee31f71c7bf428ad3e127b7a618e

> ---

>  gcc/config/arm/arm.c | 66 +++++++++++++++++++++++++++++++++-------------------

>  1 file changed, 42 insertions(+), 24 deletions(-)

> 

> diff --git a/gcc/config/arm/arm.c b/gcc/config/arm/arm.c

> index 6cae178..7f002f1 100644

> --- a/gcc/config/arm/arm.c

> +++ b/gcc/config/arm/arm.c

> @@ -9072,6 +9072,47 @@ arm_unspec_cost (rtx x, enum rtx_code /* outer_code */, bool speed_p, int *cost)

>  	  }								\

>  	while (0);

>  

> +/* Helper function for arm_rtx_costs_internal. Calculates the cost of a MEM,

> +   considering the costs of the addressing mode and memory access

> +   separately.  */

> +static bool

> +arm_mem_costs (rtx x, const struct cpu_cost_table *extra_cost,

> +	       int *cost, bool speed_p)

> +{

> +  machine_mode mode = GET_MODE (x);

> +  if (flag_pic

> +      && GET_CODE (XEXP (x, 0)) == PLUS

> +      && will_be_in_index_register (XEXP (XEXP (x, 0), 1)))

> +    /* This will be split into two instructions.  Add the cost of the

> +       additional instruction here.  The cost of the memory access is computed

> +       below.  See arm.md:calculate_pic_address.  */

> +    *cost = COSTS_N_INSNS (1);

> +  else

> +    *cost = 0;

> +

> +  /* Calculate cost of the addressing mode.  */

> +  if (speed_p)

> +  {


This patch needs to be reformatted in the GNU style (indentation of
braces, braces and else clauses on separate lines etc).

> +    /* TODO: Add table-driven costs for addressing modes.  */


You need to sort out the comment.  What's missing here?

> +  }

> +

> +  /* cost of memory access */

> +  if (speed_p)

> +  {

> +    /* data transfer is transfer size divided by bus width.  */

> +    int bus_width = arm_arch7 ? 8 : 4;


Basing bus width on the architecture is a bit too simplistic.  Instead
this should be a parameter that comes from the CPU cost tables, based on
the current tune target.

> +    *cost += COSTS_N_INSNS((GET_MODE_SIZE (mode) + bus_width - 1) / bus_width);


Use CEIL (from system.h)

> +    *cost += extra_cost->ldst.load;

> +  } else {

> +    *cost += COSTS_N_INSNS (1);

> +  }

> +

> +  return true;

> +}

> +/* Convert fron bytes to ints.  */

> +#define ARM_NUM_INTS(X) (((X) + UNITS_PER_WORD - 1) / UNITS_PER_WORD)

> +

> +

>  /* RTX costs.  Make an estimate of the cost of executing the operation

>     X, which is contained with an operation with code OUTER_CODE.

>     SPEED_P indicates whether the cost desired is the performance cost,

> @@ -9152,30 +9193,7 @@ arm_rtx_costs_internal (rtx x, enum rtx_code code, enum rtx_code outer_code,

>        return false;

>  

>      case MEM:

> -      /* A memory access costs 1 insn if the mode is small, or the address is

> -	 a single register, otherwise it costs one insn per word.  */

> -      if (REG_P (XEXP (x, 0)))

> -	*cost = COSTS_N_INSNS (1);

> -      else if (flag_pic

> -	       && GET_CODE (XEXP (x, 0)) == PLUS

> -	       && will_be_in_index_register (XEXP (XEXP (x, 0), 1)))

> -	/* This will be split into two instructions.

> -	   See arm.md:calculate_pic_address.  */

> -	*cost = COSTS_N_INSNS (2);

> -      else

> -	*cost = COSTS_N_INSNS (ARM_NUM_REGS (mode));

> -

> -      /* For speed optimizations, add the costs of the address and

> -	 accessing memory.  */

> -      if (speed_p)

> -#ifdef NOT_YET

> -	*cost += (extra_cost->ldst.load

> -		  + arm_address_cost (XEXP (x, 0), mode,

> -				      ADDR_SPACE_GENERIC, speed_p));

> -#else

> -        *cost += extra_cost->ldst.load;

> -#endif

> -      return true;

> +      return arm_mem_costs (x, extra_cost, cost, speed_p);

>  

>      case PARALLEL:

>      {

>
Charles Baylis Aug. 25, 2017, 5:43 p.m. | #3
On 9 June 2017 at 14:59, Richard Earnshaw (lists)
<Richard.Earnshaw@arm.com> wrote:
> On 21/02/17 16:54, charles.baylis@linaro.org wrote:

>> From: Charles Baylis <charles.baylis@linaro.org>

>>

>> This patch moves the calculation of costs for MEM into a

>> separate function, and reforms the calculation into two

>> parts. Firstly any additional cost of the addressing mode

>> is calculated, and then the cost of the memory access itself

>> is added.

>>

>> In this patch, the calculation of the cost of the addressing

>> mode is left as a placeholder, to be added in a subsequent

>> patch.

>>

>> gcc/ChangeLog:

>>

>> <date>  Charles Baylis  <charles.baylis@linaro.org>

>>

>>         * config/arm/arm.c (arm_mem_costs): New function.

>>         (arm_rtx_costs_internal): Use arm_mem_costs.

>

> I like the idea of this patch, but it needs further work...

>

> Comments inline.

>

> R.

>

>>

>> Change-Id: I99e93406ea39ee31f71c7bf428ad3e127b7a618e

>> ---

>>  gcc/config/arm/arm.c | 66 +++++++++++++++++++++++++++++++++-------------------

>>  1 file changed, 42 insertions(+), 24 deletions(-)

>>

>> diff --git a/gcc/config/arm/arm.c b/gcc/config/arm/arm.c

>> index 6cae178..7f002f1 100644

>> --- a/gcc/config/arm/arm.c

>> +++ b/gcc/config/arm/arm.c

>> @@ -9072,6 +9072,47 @@ arm_unspec_cost (rtx x, enum rtx_code /* outer_code */, bool speed_p, int *cost)

>>         }                                                             \

>>       while (0);

>>

>> +/* Helper function for arm_rtx_costs_internal. Calculates the cost of a MEM,

>> +   considering the costs of the addressing mode and memory access

>> +   separately.  */

>> +static bool

>> +arm_mem_costs (rtx x, const struct cpu_cost_table *extra_cost,

>> +            int *cost, bool speed_p)

>> +{

>> +  machine_mode mode = GET_MODE (x);

>> +  if (flag_pic

>> +      && GET_CODE (XEXP (x, 0)) == PLUS

>> +      && will_be_in_index_register (XEXP (XEXP (x, 0), 1)))

>> +    /* This will be split into two instructions.  Add the cost of the

>> +       additional instruction here.  The cost of the memory access is computed

>> +       below.  See arm.md:calculate_pic_address.  */

>> +    *cost = COSTS_N_INSNS (1);

>> +  else

>> +    *cost = 0;

>> +

>> +  /* Calculate cost of the addressing mode.  */

>> +  if (speed_p)

>> +  {

>

> This patch needs to be reformatted in the GNU style (indentation of

> braces, braces and else clauses on separate lines etc).


Done.

>> +    /* TODO: Add table-driven costs for addressing modes.  */

>

> You need to sort out the comment.  What's missing here?


What's missing is patch 2... I've updated the comment for clarity.

>> +  }

>> +

>> +  /* cost of memory access */

>> +  if (speed_p)

>> +  {

>> +    /* data transfer is transfer size divided by bus width.  */

>> +    int bus_width = arm_arch7 ? 8 : 4;

>

> Basing bus width on the architecture is a bit too simplistic.  Instead

> this should be a parameter that comes from the CPU cost tables, based on

> the current tune target.


This was actually Ramana's suggestion, so I've left it as-is in this
patch. If necessary, I think it's better to move this to a table in a
separate patch, as I'll need to guess the correct bus width for a
number of CPUs and will probably get some wrong.

>> +    *cost += COSTS_N_INSNS((GET_MODE_SIZE (mode) + bus_width - 1) / bus_width);

>

> Use CEIL (from system.h)


Done.

Updated patch attached.
From 18629835ba12fdfa693e2f9492a5fc23d95ef165 Mon Sep 17 00:00:00 2001
From: Charles Baylis <charles.baylis@linaro.org>
Date: Wed, 8 Feb 2017 16:52:10 +0000
Subject: [PATCH 1/3] [ARM] Refactor costs calculation for MEM.

This patch moves the calculation of costs for MEM into a
separate function, and reforms the calculation into two
parts. Firstly any additional cost of the addressing mode
is calculated, and then the cost of the memory access itself
is added.

In this patch, the calculation of the cost of the addressing
mode is left as a placeholder, to be added in a subsequent
patch.

gcc/ChangeLog:

<date>  Charles Baylis  <charles.baylis@linaro.org>

        * config/arm/arm.c (arm_mem_costs): New function.
        (arm_rtx_costs_internal): Use arm_mem_costs.

Change-Id: I99e93406ea39ee31f71c7bf428ad3e127b7a618e
---
 gcc/config/arm/arm.c | 67 ++++++++++++++++++++++++++++++++--------------------
 1 file changed, 42 insertions(+), 25 deletions(-)

diff --git a/gcc/config/arm/arm.c b/gcc/config/arm/arm.c
index fa3e2fa..13cd421 100644
--- a/gcc/config/arm/arm.c
+++ b/gcc/config/arm/arm.c
@@ -9198,8 +9198,48 @@ arm_unspec_cost (rtx x, enum rtx_code /* outer_code */, bool speed_p, int *cost)
 	  }								\
 	while (0);
 
+/* Helper function for arm_rtx_costs_internal.  Calculates the cost of a MEM,
+   considering the costs of the addressing mode and memory access
+   separately.  */
+static bool
+arm_mem_costs (rtx x, const struct cpu_cost_table *extra_cost,
+	       int *cost, bool speed_p)
+{
+  machine_mode mode = GET_MODE (x);
+  if (flag_pic
+      && GET_CODE (XEXP (x, 0)) == PLUS
+      && will_be_in_index_register (XEXP (XEXP (x, 0), 1)))
+    /* This will be split into two instructions.  Add the cost of the
+       additional instruction here.  The cost of the memory access is computed
+       below.  See arm.md:calculate_pic_address.  */
+    *cost = COSTS_N_INSNS (1);
+  else
+    *cost = 0;
+
+  /* Calculate cost of the addressing mode.  */
+  if (speed_p)
+    {
+      /* TODO: Add table-driven costs for addressing modes.  (See patch 2) */
+    }
+
+  /* Calculate cost of memory access.  */
+  if (speed_p)
+    {
+      /* data transfer is transfer size divided by bus width.  */
+      int bus_width = arm_arch7 ? 8 : 4;
+      *cost += CEIL (GET_MODE_SIZE (mode), bus_width);
+      *cost += extra_cost->ldst.load;
+    }
+  else
+    {
+      *cost += COSTS_N_INSNS (1);
+    }
+
+  return true;
+}
+
 /* RTX costs.  Make an estimate of the cost of executing the operation
-   X, which is contained with an operation with code OUTER_CODE.
+   X, which is contained within an operation with code OUTER_CODE.
    SPEED_P indicates whether the cost desired is the performance cost,
    or the size cost.  The estimate is stored in COST and the return
    value is TRUE if the cost calculation is final, or FALSE if the
@@ -9278,30 +9318,7 @@ arm_rtx_costs_internal (rtx x, enum rtx_code code, enum rtx_code outer_code,
       return false;
 
     case MEM:
-      /* A memory access costs 1 insn if the mode is small, or the address is
-	 a single register, otherwise it costs one insn per word.  */
-      if (REG_P (XEXP (x, 0)))
-	*cost = COSTS_N_INSNS (1);
-      else if (flag_pic
-	       && GET_CODE (XEXP (x, 0)) == PLUS
-	       && will_be_in_index_register (XEXP (XEXP (x, 0), 1)))
-	/* This will be split into two instructions.
-	   See arm.md:calculate_pic_address.  */
-	*cost = COSTS_N_INSNS (2);
-      else
-	*cost = COSTS_N_INSNS (ARM_NUM_REGS (mode));
-
-      /* For speed optimizations, add the costs of the address and
-	 accessing memory.  */
-      if (speed_p)
-#ifdef NOT_YET
-	*cost += (extra_cost->ldst.load
-		  + arm_address_cost (XEXP (x, 0), mode,
-				      ADDR_SPACE_GENERIC, speed_p));
-#else
-        *cost += extra_cost->ldst.load;
-#endif
-      return true;
+      return arm_mem_costs (x, extra_cost, cost, speed_p);
 
     case PARALLEL:
     {

Patch

diff --git a/gcc/config/arm/arm.c b/gcc/config/arm/arm.c
index 6cae178..7f002f1 100644
--- a/gcc/config/arm/arm.c
+++ b/gcc/config/arm/arm.c
@@ -9072,6 +9072,47 @@  arm_unspec_cost (rtx x, enum rtx_code /* outer_code */, bool speed_p, int *cost)
 	  }								\
 	while (0);
 
+/* Helper function for arm_rtx_costs_internal. Calculates the cost of a MEM,
+   considering the costs of the addressing mode and memory access
+   separately.  */
+static bool
+arm_mem_costs (rtx x, const struct cpu_cost_table *extra_cost,
+	       int *cost, bool speed_p)
+{
+  machine_mode mode = GET_MODE (x);
+  if (flag_pic
+      && GET_CODE (XEXP (x, 0)) == PLUS
+      && will_be_in_index_register (XEXP (XEXP (x, 0), 1)))
+    /* This will be split into two instructions.  Add the cost of the
+       additional instruction here.  The cost of the memory access is computed
+       below.  See arm.md:calculate_pic_address.  */
+    *cost = COSTS_N_INSNS (1);
+  else
+    *cost = 0;
+
+  /* Calculate cost of the addressing mode.  */
+  if (speed_p)
+  {
+    /* TODO: Add table-driven costs for addressing modes.  */
+  }
+
+  /* cost of memory access */
+  if (speed_p)
+  {
+    /* data transfer is transfer size divided by bus width.  */
+    int bus_width = arm_arch7 ? 8 : 4;
+    *cost += COSTS_N_INSNS((GET_MODE_SIZE (mode) + bus_width - 1) / bus_width);
+    *cost += extra_cost->ldst.load;
+  } else {
+    *cost += COSTS_N_INSNS (1);
+  }
+
+  return true;
+}
+/* Convert fron bytes to ints.  */
+#define ARM_NUM_INTS(X) (((X) + UNITS_PER_WORD - 1) / UNITS_PER_WORD)
+
+
 /* RTX costs.  Make an estimate of the cost of executing the operation
    X, which is contained with an operation with code OUTER_CODE.
    SPEED_P indicates whether the cost desired is the performance cost,
@@ -9152,30 +9193,7 @@  arm_rtx_costs_internal (rtx x, enum rtx_code code, enum rtx_code outer_code,
       return false;
 
     case MEM:
-      /* A memory access costs 1 insn if the mode is small, or the address is
-	 a single register, otherwise it costs one insn per word.  */
-      if (REG_P (XEXP (x, 0)))
-	*cost = COSTS_N_INSNS (1);
-      else if (flag_pic
-	       && GET_CODE (XEXP (x, 0)) == PLUS
-	       && will_be_in_index_register (XEXP (XEXP (x, 0), 1)))
-	/* This will be split into two instructions.
-	   See arm.md:calculate_pic_address.  */
-	*cost = COSTS_N_INSNS (2);
-      else
-	*cost = COSTS_N_INSNS (ARM_NUM_REGS (mode));
-
-      /* For speed optimizations, add the costs of the address and
-	 accessing memory.  */
-      if (speed_p)
-#ifdef NOT_YET
-	*cost += (extra_cost->ldst.load
-		  + arm_address_cost (XEXP (x, 0), mode,
-				      ADDR_SPACE_GENERIC, speed_p));
-#else
-        *cost += extra_cost->ldst.load;
-#endif
-      return true;
+      return arm_mem_costs (x, extra_cost, cost, speed_p);
 
     case PARALLEL:
     {