This is patch #1 of 3 that adds the support that can be used in developing GCC
support for potential future PowerPC processors. With all 3 patches, the tuning
for the 'future' processor is the same as power10 and power11. It may be in the
future this tuning will change as any future PowerPC processor evolves.
Patch #2 will change the tuning support in power10.md to treat -mtune=future the
same as -mtune=power10.
Patch #3 will add tests for -mcpu=future.
These changes are being added so that hardware designers can evaluate potential
new features to be added to the PowerPC processors in the future. It may be
these features will be incorporated into real hardware using a different name in
the future. Or it may be these features will not be incoporated into actual
PowerPC hardware in the future.
I have rewritten these patches to make it easier in the future to add new
processors that scheduled like power10 and power11 systems, or remove the
'future' cpu from being scheduled like a power10 if we add new tuning
characteristics for potential future processors.
I added a new macro (CASE_PROCESSOR_POWER10_TUNING) that expands to a set of
PROCESSOR_<xxx> cases for processors that should be tuned like a power10
processor. In this patch, power10, power11, and future are selected.
I also added a new inline function (power10_tuning_p) that returns true if the
processor is to be scheduled like a power10.
I have modified the various ASM_CPU_SPEC macros to pass -mfuture to the
assembler if -mcpu=future wa used.
I have updated config.guess to allow the user to configure the GCC compiler
using the --with-cpu=future option.
I have tested these patches on both big endian and little endian PowerPC
servers, with no regressions. Can I check these patchs into the trunk?
2025-06-25 Michael Meissner <meiss...@linux.ibm.com>
gcc/
* config.guess (powerpc*-*-*): Add support for using --with-cpu=future.
* config/rs6000/aix71.h (ASM_CPU_SPEC): Pass -mfuture to the assembler
if -mcpu=future was used on the command line.
* config/rs6000/aix72.h (ASM_CPU_SPEC): Likewise.
* config/rs6000/aix73.h (ASM_CPU_SPEC): Likewise.
* config/rs6000/driver-rs6000.c (asm_names): Likewise.
* config/rs6000/rs6000-c.cc (rs6000_target_modify_macros): If
-mcpu=future, define the macro _ARCH_FUTURE.
* config/rs6000/rs6000-cpus.def (FUTURE_MASKS_SERVER): New macro.
(POWERPC_MASKS): Add support for the internal -mfuture that is set by
the -mcpu=future option.
(future cpu): Add support for -mcpu=future.
* config/rs6000/rs6000-opts.h (enum processor_type): Add
PROCESSOR_FUTURE.
(CASE_PROCESSOR_POWER10_TUNING): New macro.
(power10_tuning_p): New inline helper function.
* config/rs6000/rs6000-string.cc (expand_compare_loop): Convert
PROCESSOR_POWER10 and PROCESSOR_POWER11 cases into
CASE_PROCESSOR_POWER10_TUNING to allow -mcpu=future as well.
* config/rs6000/rs6000-tables.opt: Regenerate the table, adding the
future cpu to the enumeration.
* config/rs6000/rs6000.cc (rs6000_option_override_internal): Change
tests against PROCESSOR_POWER10 and PROCESSOR_POWER11 to include
PROCESSOR_FUTURE as well.
(rs6000_reassociation_width): Likewise.
(rs6000_adjust_cost): Likewise.
(rs6000_issue_rate): Likewise.
(rs6000_sched_reorder): Likewise.
(rs6000_sched_reorder2): Likewise.
(rs6000_register_move_cost): Likewise.
* config/rs6000/rs6000.md (cpu attribute): Add future cpu.
* config/rs6000/rs6000.h (ASM_CPU_SPEC): Pass -mfuture to the assembler
if -mcpu=future was used on the command line.
* config/rs6000/rs6000.opt (-mfuture): New internal ISA bit for
-mcpu=future.
---
gcc/config.gcc | 4 +--
gcc/config/rs6000/aix71.h | 1 +
gcc/config/rs6000/aix72.h | 1 +
gcc/config/rs6000/aix73.h | 1 +
gcc/config/rs6000/driver-rs6000.cc | 2 ++
gcc/config/rs6000/rs6000-c.cc | 2 ++
gcc/config/rs6000/rs6000-cpus.def | 5 ++++
gcc/config/rs6000/rs6000-opts.h | 22 ++++++++++++++++
gcc/config/rs6000/rs6000-string.cc | 3 +--
gcc/config/rs6000/rs6000-tables.opt | 11 +++++---
gcc/config/rs6000/rs6000.cc | 39 ++++++++++-------------------
gcc/config/rs6000/rs6000.h | 1 +
gcc/config/rs6000/rs6000.md | 2 +-
gcc/config/rs6000/rs6000.opt | 6 +++++
14 files changed, 65 insertions(+), 35 deletions(-)
diff --git a/gcc/config.gcc b/gcc/config.gcc
index a6f6efec4e1..d9318f257e7 100644
--- a/gcc/config.gcc
+++ b/gcc/config.gcc
@@ -533,7 +533,7 @@ powerpc*-*-*)
extra_headers="${extra_headers} ppu_intrinsics.h spu2vmx.h vec_types.h
si2vmx.h"
extra_headers="${extra_headers} amo.h"
case x$with_cpu in
-
xpowerpc64|xdefault64|x6[23]0|x970|xG5|xpower[3456789]|xpower1[01]|xpower6x|xrs64a|xcell|xa2|xe500mc64|xe5500|xe6500)
+
xpowerpc64|xdefault64|x6[23]0|x970|xG5|xpower[3456789]|xpower1[01]|xpower6x|xrs64a|xcell|xa2|xe500mc64|xe5500|xe6500|xfuture)
cpu_is_64bit=yes
;;
esac
@@ -5694,7 +5694,7 @@ case "${target}" in
tm_defines="${tm_defines} CONFIG_PPC405CR"
eval "with_$which=405"
;;
- "" | common | native \
+ "" | common | native | future \
| power[3456789] | power1[01] | power5+ | power6x \
| powerpc | powerpc64 | powerpc64le \
| rs64 \
diff --git a/gcc/config/rs6000/aix71.h b/gcc/config/rs6000/aix71.h
index 2b21dd7cd1e..77651f5ea30 100644
--- a/gcc/config/rs6000/aix71.h
+++ b/gcc/config/rs6000/aix71.h
@@ -79,6 +79,7 @@ do {
\
#undef ASM_CPU_SPEC
#define ASM_CPU_SPEC \
"%{mcpu=native: %(asm_cpu_native); \
+ mcpu=future: -mfuture; \
mcpu=power11: -mpwr11; \
mcpu=power10: -mpwr10; \
mcpu=power9: -mpwr9; \
diff --git a/gcc/config/rs6000/aix72.h b/gcc/config/rs6000/aix72.h
index 53c0bde5ad4..652f60c7f49 100644
--- a/gcc/config/rs6000/aix72.h
+++ b/gcc/config/rs6000/aix72.h
@@ -79,6 +79,7 @@ do {
\
#undef ASM_CPU_SPEC
#define ASM_CPU_SPEC \
"%{mcpu=native: %(asm_cpu_native); \
+ mcpu=future: -mfuture; \
mcpu=power11: -mpwr11; \
mcpu=power10: -mpwr10; \
mcpu=power9: -mpwr9; \
diff --git a/gcc/config/rs6000/aix73.h b/gcc/config/rs6000/aix73.h
index c7639368a26..3c66ac1d917 100644
--- a/gcc/config/rs6000/aix73.h
+++ b/gcc/config/rs6000/aix73.h
@@ -79,6 +79,7 @@ do {
\
#undef ASM_CPU_SPEC
#define ASM_CPU_SPEC \
"%{mcpu=native: %(asm_cpu_native); \
+ mcpu=future: -mfuture; \
mcpu=power11: -mpwr11; \
mcpu=power10: -mpwr10; \
mcpu=power9: -mpwr9; \
diff --git a/gcc/config/rs6000/driver-rs6000.cc
b/gcc/config/rs6000/driver-rs6000.cc
index 2ab675e6d04..d20332d76f1 100644
--- a/gcc/config/rs6000/driver-rs6000.cc
+++ b/gcc/config/rs6000/driver-rs6000.cc
@@ -452,6 +452,7 @@ static const struct asm_name asm_names[] = {
{ "power9", "-mpwr9" },
{ "power10", "-mpwr10" },
{ "power11", "-mpwr11" },
+ { "future", "-mfuture" },
{ "powerpc", "-mppc" },
{ "rs64", "-mppc" },
{ "603", "-m603" },
@@ -481,6 +482,7 @@ static const struct asm_name asm_names[] = {
{ "power9", "-mpower9" },
{ "power10", "-mpower10" },
{ "power11", "-mpower11" },
+ { "future", "-mfuture" },
{ "a2", "-ma2" },
{ "powerpc", "-mppc" },
{ "powerpc64", "-mppc64" },
diff --git a/gcc/config/rs6000/rs6000-c.cc b/gcc/config/rs6000/rs6000-c.cc
index d3b0a566821..6757a2477ad 100644
--- a/gcc/config/rs6000/rs6000-c.cc
+++ b/gcc/config/rs6000/rs6000-c.cc
@@ -437,6 +437,8 @@ rs6000_target_modify_macros (bool define_p, HOST_WIDE_INT
flags)
rs6000_define_or_undefine_macro (define_p, "_ARCH_PWR10");
if ((flags & OPTION_MASK_POWER11) != 0)
rs6000_define_or_undefine_macro (define_p, "_ARCH_PWR11");
+ if ((flags & OPTION_MASK_FUTURE) != 0)
+ rs6000_define_or_undefine_macro (define_p, "_ARCH_FUTURE");
if ((flags & OPTION_MASK_SOFT_FLOAT) != 0)
rs6000_define_or_undefine_macro (define_p, "_SOFT_FLOAT");
if ((flags & OPTION_MASK_RECIP_PRECISION) != 0)
diff --git a/gcc/config/rs6000/rs6000-cpus.def
b/gcc/config/rs6000/rs6000-cpus.def
index 4a1037616d7..228d0b5e7b5 100644
--- a/gcc/config/rs6000/rs6000-cpus.def
+++ b/gcc/config/rs6000/rs6000-cpus.def
@@ -83,6 +83,9 @@
#define POWER11_MASKS_SERVER (ISA_3_1_MASKS_SERVER \
| OPTION_MASK_POWER11)
+#define FUTURE_MASKS_SERVER (POWER11_MASKS_SERVER \
+ | OPTION_MASK_FUTURE)
+
/* Flags that need to be turned off if -mno-vsx. */
#define OTHER_VSX_VECTOR_MASKS (OPTION_MASK_EFFICIENT_UNALIGNED_VSX \
| OPTION_MASK_FLOAT128_KEYWORD \
@@ -119,6 +122,7 @@
| OPTION_MASK_FLOAT128_HW \
| OPTION_MASK_FLOAT128_KEYWORD \
| OPTION_MASK_FPRND \
+ | OPTION_MASK_FUTURE \
| OPTION_MASK_POWER10 \
| OPTION_MASK_POWER11 \
| OPTION_MASK_P10_FUSION \
@@ -249,6 +253,7 @@ RS6000_CPU ("power9", PROCESSOR_POWER9, MASK_POWERPC64 |
ISA_3_0_MASKS_SERVER
| OPTION_MASK_HTM)
RS6000_CPU ("power10", PROCESSOR_POWER10, MASK_POWERPC64 |
ISA_3_1_MASKS_SERVER)
RS6000_CPU ("power11", PROCESSOR_POWER11, MASK_POWERPC64 |
POWER11_MASKS_SERVER)
+RS6000_CPU ("future", PROCESSOR_FUTURE, MASK_POWERPC64 | FUTURE_MASKS_SERVER)
RS6000_CPU ("powerpc", PROCESSOR_POWERPC, 0)
RS6000_CPU ("powerpc64", PROCESSOR_POWERPC64, OPTION_MASK_PPC_GFXOPT
| MASK_POWERPC64)
diff --git a/gcc/config/rs6000/rs6000-opts.h b/gcc/config/rs6000/rs6000-opts.h
index c31d2975f04..0ec2016a4d2 100644
--- a/gcc/config/rs6000/rs6000-opts.h
+++ b/gcc/config/rs6000/rs6000-opts.h
@@ -63,6 +63,7 @@ enum processor_type
PROCESSOR_POWER9,
PROCESSOR_POWER10,
PROCESSOR_POWER11,
+ PROCESSOR_FUTURE,
PROCESSOR_RS64A,
PROCESSOR_MPCCORE,
@@ -72,6 +73,27 @@ enum processor_type
};
+/* Processors that tune the same as a power10 in terms of tuning that will be
+ added to switch statements. */
+#define CASE_PROCESSOR_POWER10_TUNING \
+ case PROCESSOR_POWER10:
\
+ case PROCESSOR_POWER11: \
+ case PROCESSOR_FUTURE
+
+/* Return true if the processor uses the same tuning as power10. */
+static inline bool
+power10_tuning_p (enum processor_type processor)
+{
+ switch (processor)
+ {
+ CASE_PROCESSOR_POWER10_TUNING:
+ return true;
+
+ default:
+ return false;
+ }
+}
+
/* Types of costly dependences. */
enum rs6000_dependence_cost
{
diff --git a/gcc/config/rs6000/rs6000-string.cc
b/gcc/config/rs6000/rs6000-string.cc
index 3d2911ca08a..16107c909f3 100644
--- a/gcc/config/rs6000/rs6000-string.cc
+++ b/gcc/config/rs6000/rs6000-string.cc
@@ -963,8 +963,7 @@ expand_compare_loop (rtx operands[])
max_bytes = 64;
break;
case PROCESSOR_POWER9:
- case PROCESSOR_POWER10:
- case PROCESSOR_POWER11:
+ CASE_PROCESSOR_POWER10_TUNING:
if (bytes_is_const)
max_bytes = 191;
else
diff --git a/gcc/config/rs6000/rs6000-tables.opt
b/gcc/config/rs6000/rs6000-tables.opt
index f5bbed5ea74..518324e7aea 100644
--- a/gcc/config/rs6000/rs6000-tables.opt
+++ b/gcc/config/rs6000/rs6000-tables.opt
@@ -189,14 +189,17 @@ EnumValue
Enum(rs6000_cpu_opt_value) String(power11) Value(53)
EnumValue
-Enum(rs6000_cpu_opt_value) String(powerpc) Value(54)
+Enum(rs6000_cpu_opt_value) String(future) Value(54)
EnumValue
-Enum(rs6000_cpu_opt_value) String(powerpc64) Value(55)
+Enum(rs6000_cpu_opt_value) String(powerpc) Value(55)
EnumValue
-Enum(rs6000_cpu_opt_value) String(powerpc64le) Value(56)
+Enum(rs6000_cpu_opt_value) String(powerpc64) Value(56)
EnumValue
-Enum(rs6000_cpu_opt_value) String(rs64) Value(57)
+Enum(rs6000_cpu_opt_value) String(powerpc64le) Value(57)
+
+EnumValue
+Enum(rs6000_cpu_opt_value) String(rs64) Value(58)
diff --git a/gcc/config/rs6000/rs6000.cc b/gcc/config/rs6000/rs6000.cc
index 7ee26e52b13..ec789f86932 100644
--- a/gcc/config/rs6000/rs6000.cc
+++ b/gcc/config/rs6000/rs6000.cc
@@ -4387,8 +4387,7 @@ rs6000_option_override_internal (bool global_init_p)
generating power10 instructions. */
if (!(rs6000_isa_flags_explicit & OPTION_MASK_P10_FUSION))
{
- if (rs6000_tune == PROCESSOR_POWER10
- || rs6000_tune == PROCESSOR_POWER11)
+ if (power10_tuning_p (rs6000_tune))
rs6000_isa_flags |= OPTION_MASK_P10_FUSION;
else
rs6000_isa_flags &= ~OPTION_MASK_P10_FUSION;
@@ -4416,8 +4415,7 @@ rs6000_option_override_internal (bool global_init_p)
&& rs6000_tune != PROCESSOR_POWER7
&& rs6000_tune != PROCESSOR_POWER8
&& rs6000_tune != PROCESSOR_POWER9
- && rs6000_tune != PROCESSOR_POWER10
- && rs6000_tune != PROCESSOR_POWER11
+ && !power10_tuning_p (rs6000_tune)
&& rs6000_tune != PROCESSOR_PPCA2
&& rs6000_tune != PROCESSOR_CELL
&& rs6000_tune != PROCESSOR_PPC476);
@@ -4431,8 +4429,7 @@ rs6000_option_override_internal (bool global_init_p)
|| rs6000_tune == PROCESSOR_POWER7
|| rs6000_tune == PROCESSOR_POWER8
|| rs6000_tune == PROCESSOR_POWER9
- || rs6000_tune == PROCESSOR_POWER10
- || rs6000_tune == PROCESSOR_POWER11
+ || power10_tuning_p (rs6000_tune)
|| rs6000_tune == PROCESSOR_PPCE500MC
|| rs6000_tune == PROCESSOR_PPCE500MC64
|| rs6000_tune == PROCESSOR_PPCE5500
@@ -4731,8 +4728,7 @@ rs6000_option_override_internal (bool global_init_p)
rs6000_cost = &power9_cost;
break;
- case PROCESSOR_POWER10:
- case PROCESSOR_POWER11:
+ CASE_PROCESSOR_POWER10_TUNING:
rs6000_cost = &power10_cost;
break;
@@ -10153,8 +10149,7 @@ rs6000_reassociation_width (unsigned int opc
ATTRIBUTE_UNUSED,
{
case PROCESSOR_POWER8:
case PROCESSOR_POWER9:
- case PROCESSOR_POWER10:
- case PROCESSOR_POWER11:
+ CASE_PROCESSOR_POWER10_TUNING:
if (DECIMAL_FLOAT_MODE_P (mode))
return 1;
if (VECTOR_MODE_P (mode))
@@ -18196,8 +18191,7 @@ rs6000_adjust_cost (rtx_insn *insn, int dep_type,
rtx_insn *dep_insn, int cost,
/* Separate a load from a narrower, dependent store. */
if ((rs6000_sched_groups || rs6000_tune == PROCESSOR_POWER9
- || rs6000_tune == PROCESSOR_POWER10
- || rs6000_tune == PROCESSOR_POWER11)
+ || power10_tuning_p (rs6000_tune))
&& GET_CODE (PATTERN (insn)) == SET
&& GET_CODE (PATTERN (dep_insn)) == SET
&& MEM_P (XEXP (PATTERN (insn), 1))
@@ -18235,8 +18229,7 @@ rs6000_adjust_cost (rtx_insn *insn, int dep_type,
rtx_insn *dep_insn, int cost,
|| rs6000_tune == PROCESSOR_POWER7
|| rs6000_tune == PROCESSOR_POWER8
|| rs6000_tune == PROCESSOR_POWER9
- || rs6000_tune == PROCESSOR_POWER10
- || rs6000_tune == PROCESSOR_POWER11
+ || power10_tuning_p (rs6000_tune)
|| rs6000_tune == PROCESSOR_CELL)
&& recog_memoized (dep_insn)
&& (INSN_CODE (dep_insn) >= 0))
@@ -18810,8 +18803,7 @@ rs6000_issue_rate (void)
return 7;
case PROCESSOR_POWER9:
return 6;
- case PROCESSOR_POWER10:
- case PROCESSOR_POWER11:
+ CASE_PROCESSOR_POWER10_TUNING:
return 8;
default:
return 1;
@@ -19527,10 +19519,8 @@ rs6000_sched_reorder (FILE *dump ATTRIBUTE_UNUSED, int
sched_verbose,
if (rs6000_tune == PROCESSOR_POWER6)
load_store_pendulum = 0;
- /* Do Power10/Power11 dependent reordering. */
- if (last_scheduled_insn
- && (rs6000_tune == PROCESSOR_POWER10
- || rs6000_tune == PROCESSOR_POWER11))
+ /* Do Power10 and similar processors dependent reordering. */
+ if (last_scheduled_insn && power10_tuning_p (rs6000_tune))
power10_sched_reorder (ready, n_ready - 1);
return rs6000_issue_rate ();
@@ -19554,10 +19544,8 @@ rs6000_sched_reorder2 (FILE *dump, int sched_verbose,
rtx_insn **ready,
&& recog_memoized (last_scheduled_insn) >= 0)
return power9_sched_reorder2 (ready, *pn_ready - 1);
- /* Do Power10/Power11 dependent reordering. */
- if (last_scheduled_insn
- && (rs6000_tune == PROCESSOR_POWER10
- || rs6000_tune == PROCESSOR_POWER11))
+ /* Do Power10 and similar processors dependent reordering. */
+ if (last_scheduled_insn && power10_tuning_p (rs6000_tune))
return power10_sched_reorder (ready, *pn_ready - 1);
return cached_can_issue_more;
@@ -22779,8 +22767,7 @@ rs6000_register_move_cost (machine_mode mode,
allocation a move within the same class might turn
out to be a nop. */
if (rs6000_tune == PROCESSOR_POWER9
- || rs6000_tune == PROCESSOR_POWER10
- || rs6000_tune == PROCESSOR_POWER11)
+ || power10_tuning_p (rs6000_tune))
ret = 3 * hard_regno_nregs (FIRST_GPR_REGNO, mode);
else
ret = 4 * hard_regno_nregs (FIRST_GPR_REGNO, mode);
diff --git a/gcc/config/rs6000/rs6000.h b/gcc/config/rs6000/rs6000.h
index db6112a09e1..76c6cd1ab8e 100644
--- a/gcc/config/rs6000/rs6000.h
+++ b/gcc/config/rs6000/rs6000.h
@@ -101,6 +101,7 @@
you make changes here, make them also there. */
#define ASM_CPU_SPEC \
"%{mcpu=native: %(asm_cpu_native); \
+ mcpu=future: -mfuture; \
mcpu=power11: -mpower11; \
mcpu=power10: -mpower10; \
mcpu=power9: -mpower9; \
diff --git a/gcc/config/rs6000/rs6000.md b/gcc/config/rs6000/rs6000.md
index 9c718ca2a22..605047f99a8 100644
--- a/gcc/config/rs6000/rs6000.md
+++ b/gcc/config/rs6000/rs6000.md
@@ -364,7 +364,7 @@ (define_attr "cpu"
ppc750,ppc7400,ppc7450,
ppc403,ppc405,ppc440,ppc476,
ppc8540,ppc8548,ppce300c2,ppce300c3,ppce500mc,ppce500mc64,ppce5500,ppce6500,
- power4,power5,power6,power7,power8,power9,power10,power11,
+ power4,power5,power6,power7,power8,power9,power10,power11,future,
rs64a,mpccore,cell,ppca2,titan"
(const (symbol_ref "(enum attr_cpu) rs6000_tune")))
diff --git a/gcc/config/rs6000/rs6000.opt b/gcc/config/rs6000/rs6000.opt
index 88cf16ca581..7e1af7120bc 100644
--- a/gcc/config/rs6000/rs6000.opt
+++ b/gcc/config/rs6000/rs6000.opt
@@ -634,6 +634,12 @@ mieee128-constant
Target Var(TARGET_IEEE128_CONSTANT) Init(1) Save
Generate (do not generate) code that uses the LXVKQ instruction.
+;; Users should not use -mfuture, but we need to use a bit to identify when
+;; the user changes the default cpu via #pragma GCC target("cpu=future")
+;; and then resets it later.
+mfuture
+Target Undocumented Mask(FUTURE) Var(rs6000_isa_flags) WarnRemoved
+
; Documented parameters
-param=rs6000-vect-unroll-limit=
--
2.49.0
--
Michael Meissner, IBM
PO Box 98, Ayer, Massachusetts, USA, 01432
email: meiss...@linux.ibm.com
