The optimizations for subfcx introduced in #13852 also apply to subfx.
Rather than duplicating the logic, we merge the handlers, like we did
in #10120 for x86.
Equivalent to a negation, no need to materialize the zero.
Before:
0x52800015 mov w21, #0x0 ; =0
0x6b1802b6 subs w22, w21, w24
After:
0x6b1803f6 negs w22, w24
This case can be handled as a move. It also generates a constant carry
flag.
Before:
0x52800013 mov w19, #0x0 ; =0
0x6b1302b3 subs w19, w21, w19
After:
0x2a1503f3 mov w19, w21
Number of instructions stays the same, but we remove the false
dependency on the input registers.
Before:
0x7a1b037a sbcs w26, w27, w27
After:
0x5a9f23fa csetm w26, lo
Another instance where we needlessly materialized constant zero in a
register. We can just write the carry flag directly.
Before:
0x5280001a mov w26, #0x0 ; =0
0x1a1f035a adc w26, w26, wzr
After:
0x1a9f37fa cset w26, hs
Fix some common anti-patterns with these data structures.
- You can dereference the iterator returned by `find` to access the
underlying value directly, without an extra `operator[]`/`at`.
- Rather than checking for an element before insertion/deletion, you can
just do the operation and if needed check the return value to
determine if the insertion/deletion succeeded.
You can encode a shifted 12-bit immediate in a SUB instruction on ARM64.
We exploit this to avoid materializing the immediate.
This approach saves an instruction if it does not need to be
materialized in a register afterwards. Otherwise, we just materialize
it later and the total number of instructions stays the same.
Before:
0x52a00218 mov w24, #0x100000 ; =1048576
0xcb180379 sub x25, x27, x24
After:
0xd1440379 sub x25, x27, #0x100, lsl #12 ; =0x100000
You can encode a 12-bit immediate in a SUB instruction on ARM64. We can
exploit this to avoid materializing the immediate.
This approach saves an instruction if it does not need to be
materialized in a register afterwards. Otherwise, we just materialize
it later and the total number of instructions stays the same.
Before:
0x5280003a mov w26, #0x1 ; =1
0xcb1a033b sub x27, x25, x26
After:
0xd100073b sub x27, x25, #0x1
While we cannot always avoid materializing immediates, we can still
inspect the most significant bit and potentially skip sign extension.
This can sometimes save an instruction.
Before:
0x5280003a mov w26, #0x1 ; =1
0x93407f5b sxtw x27, w26
0xcb38c37b sub x27, x27, w24, sxtw
After:
0x5280003a mov w26, #0x1 ; =1
0xcb38c35b sub x27, x26, w24, sxtw
Before:
0x52a20018 mov w24, #0x10000000 ; =268435456
0x93407f79 sxtw x25, w27
0xcb38c339 sub x25, x25, w24, sxtw
After:
0x52a20018 mov w24, #0x10000000 ; =268435456
0x93407f79 sxtw x25, w27
0xcb180339 sub x25, x25, x24
You can encode a shifted 12-bit immediate in an ADD instruction on
ARM64. If the negated constant fits in this range, we can exploit this
to avoid materializing the immediate.
This approach saves an instruction if it does not need to be
materialized in a register afterwards. Otherwise, we just materialize
it later and the total number of instructions stays the same.
Before:
0x52bff01a mov w26, #-0x800000 ; =-8388608
0x93407f1b sxtw x27, w24
0xcb3ac37b sub x27, x27, w26, sxtw
After:
0x93407f1b sxtw x27, w24
0x9160037b add x27, x27, #0x800, lsl #12 ; =0x800000
You can encode a 12-bit immediate in an ADD instruction on ARM64. If the
negated constant fits in this range, we can exploit this to avoid
materializing the immediate.
This approach saves an instruction if it does not need to be
materialized in a register afterwards. Otherwise, we just materialize
it later and the total number of instructions stays the same.
Before:
0x12800019 mov w25, #-0x1 ; =-1
0x93407f5b sxtw x27, w26
0xcb39c37b sub x27, x27, w25, sxtw
After:
0x93407f5b sxtw x27, w26
0x9100077b add x27, x27, #0x1
You can encode a shifted 12-bit immediate in a SUB instruction on ARM64.
Constants in this range do not need to be sign extended, so we can
exploit this to avoid materializing the immediate.
This approach saves an instruction if it does not need to be
materialized in a register afterwards. Otherwise, we just materialize
it later and the total number of instructions stays the same.
Before:
0x52a00099 mov w25, #0x40000 ; =262144
0x93407f7a sxtw x26, w27
0xcb39c35a sub x26, x26, w25, sxtw
After:
0x93407f7a sxtw x26, w27
0xd141035a sub x26, x26, #0x40, lsl #12 ; =0x40000
You can encode a 12-bit immediate in a SUB instruction on ARM64.
Constants in this range do not need to be sign extended, so we can
exploit this to avoid materializing the immediate.
This approach saves an instruction if it does not need to be
materialized in a register afterwards. Otherwise, we just materialize
it later and the total number of instructions stays the same.
Before:
0x52800416 mov w22, #0x20 ; =32
0x93407f78 sxtw x24, w27
0xcb36c318 sub x24, x24, w22, sxtw
After:
0x93407f78 sxtw x24, w27
0xd1008318 sub x24, x24, #0x20
The computed value is only used when the register is equal to zero, so
we can fully precompute it and materialize the constant instead. In
other words, we change from
```
return reg == 0 ? (reg | 1ULL << 63) : reg;
```
to
```
return reg == 0 ? 1ULL << 63 : reg;
```
The number of instructions remains the same, but we eliminate an
unnecessary dependency on the register value.
Before:
0xb241037a orr x26, x27, #0x8000000000000000
0xeb1f037f cmp x27, xzr
0x9a9a137b csel x27, x27, x26, ne
After:
0xd2f0001a mov x26, #-0x8000000000000000 ; =-9223372036854775808
0xeb1f037f cmp x27, xzr
0x9a9a137b csel x27, x27, x26, ne
When the input register and carry flags are known, we can always
precompute the result.
We still materialize the immediate when the condition register
needs to be updated, but this seems to be a general problem. I might
look into that one day, but for now this'll do.
- ConstantFalse
Before:
0x52800119 mov w25, #0x8 ; =8
0x2a1903fa mov w26, w25
After:
N/A
- ConstantTrue
Before:
0x52800119 mov w25, #0x8 ; =8
0x1100073a add w26, w25, #0x1
After:
N/A
Same optimization we did for subfex. Skip loading the carry flag into a
temporary register first when we're dealing with zero.
Before:
0x394bd3b8 ldrb w24, [x29, #0x2f4]
0x2a1803f9 mov w25, w24
After:
0x394bd3b9 ldrb w25, [x29, #0x2f4]
When both the input register and the carry flag are constants, the
result can be precomputed.
Before:
0x52800016 mov w22, #0x0 ; =0
0x2a3603f6 mvn w22, w22
After:
The result is either -1 or 0 depending on the state of the carry flag.
This can be done with a csetm instruction.
Before:
0x1280001a mov w26, #-0x1 ; =-1
0x1a1f035a adc w26, w26, wzr
After:
0x5a9f23fa csetm w26, lo
When the immediate is zero, we can load the carry flag from memory
directly to the destination register.
Before:
0x394bd3b8 ldrb w24, [x29, #0x2f4]
0x2a1803f9 mov w25, w24
After:
0x394bd3b9 ldrb w25, [x29, #0x2f4]
Prefer BLENDVPD over VBLENDVPD if the latter doesn't save any
instructions.
VBLENDVPD allows separate source and destination registers, which can
eliminate a MOVAPD/MOVSD. However, on Intel since Skylake, VBLENDVPD
takes additional uops to execute compared to BLENDVPD (according to
https://uops.info). On AMD and older Intel microarchitectures there is no
difference.
The value being stored must be loaded into a register. In the case of an
immediate value, this means it must be materialized. The value is
eventually byteswapped before performing the store.
This can be simplified for the value 0 for two reasons:
- ARM64 has a dedicated zero register, so does not need to be
materialized.
- Byteswapping zero is still zero, so we can skip this step.
We could skip byteswapping for other values by immediately materializing
the byteswapped value in a register, but the benefits are not so clear
there (if the value needs to be materialized anyway, it is better to do
it up front).
Before:
0x5280001b mov w27, #0x0 ; =0
0xb9404fba ldr w26, [x29, #0x4c]
0x12881862 mov w2, #-0x40c4 ; =-16580
0x0b020342 add w2, w26, w2
0x5ac00b61 rev w1, w27
0xb8226b81 str w1, [x28, x2]
After:
0xb9404fbb ldr w27, [x29, #0x4c]
0x12881862 mov w2, #-0x40c4 ; =-16580
0x0b020362 add w2, w27, w2
0xb8226b9f str wzr, [x28, x2]
MUL and SUB can be combined in one instruction.
Before:
0x1b1a7c01 mul w1, w0, w26
0x4b010318 sub w24, w24, w1
After:
0x1b1ae018 msub w24, w0, w26, w24
No need to materialize the immediate if it is zero, we can just use WZR.
Before:
mov w27, #0x0 ; =0
str w27, [x29, #0x1178]
After:
str wzr, [x29, #0x1178]
When the immediate value is zero, we can do a negation. On ARM64 the NEG
/NEGS instructions are just an alias for SUB/SUBS with a hardcoded WZR.
Before:
```
ldr w22, [x29, #0x28]
mov w21, #0x0 ; =0
subs w22, w21, w22
```
After:
```
ldr w22, [x29, #0x28]
negs w22, w22
```
Sintendo