The load and test instructions compare the given value
against zero and will produce a condition code indicating
one of the following scenarios:
0: Result is zero
1: Result is less than zero
2: Result is greater than zero
3: Result is not a number (NaN)
The instruction can be used to simplify floating point comparisons
against zero, which can enable further optimizations.
This CL also reduces the size of .text section of math.test binary by around
0.7 KB (in hexadecimal, from 1358f0 to 135620).
Change-Id: I33cb714f0c6feebac7a1c46dfcc735e7daceff9c
Reviewed-on: https://go-review.googlesource.com/c/go/+/209159
Reviewed-by: Michael Munday <mike.munday@ibm.com>
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This reverts commit CL 222782.
Reason for revert: Reverting to see if 386 errors go away
Update #37881
Change-Id: I74f287404c52414db1b6ff1649effa4ed9e5cc0c
Reviewed-on: https://go-review.googlesource.com/c/go/+/225218
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Rolling back portions of CL 222782 to see if that helps
issue #37881 any.
Update #37881
Change-Id: I9cc3ff8c469fa5e4b22daec715d04148033f46f7
Reviewed-on: https://go-review.googlesource.com/c/go/+/224837
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
This commit extends the -spectre flag to cmd/asm and adds
a new Spectre mitigation mode "ret", which enables the use
of retpolines.
Retpolines prevent speculation about the target of an indirect
jump or call and are described in more detail here:
https://support.google.com/faqs/answer/7625886
Change-Id: I4f2cb982fa94e44d91e49bd98974fd125619c93a
Reviewed-on: https://go-review.googlesource.com/c/go/+/222661
Reviewed-by: Keith Randall <khr@golang.org>
This commit adds a new cmd/compile flag -spectre,
which accepts a comma-separated list of possible
Spectre mitigations to apply, or the empty string (none),
or "all". The only known mitigation right now is "index",
which uses conditional moves to ensure that x86-64 CPUs
do not speculate past index bounds checks.
Speculating past index bounds checks may be problematic
on systems running privileged servers that accept requests
from untrusted users who can execute their own programs
on the same machine. (And some more constraints that
make it even more unlikely in practice.)
The cases this protects against are analogous to the ones
Microsoft explains in the "Array out of bounds load/store feeding ..."
sections here:
https://docs.microsoft.com/en-us/cpp/security/developer-guidance-speculative-execution?view=vs-2019#array-out-of-bounds-load-feeding-an-indirect-branch
Change-Id: Ib7532d7e12466b17e04c4e2075c2a456dc98f610
Reviewed-on: https://go-review.googlesource.com/c/go/+/222660
Reviewed-by: Keith Randall <khr@golang.org>
Use a separate compiler pass to introduce complicated x86 addressing
modes. Loads in the normal architecture rules (for x86 and all other
platforms) can have constant offsets (AuxInt values) and symbols (Aux
values), but no more.
The complex addressing modes (x+y, x+2*y, etc.) are introduced in a
separate pass that combines loads with LEAQx ops.
Organizing rewrites this way simplifies the number of rewrites
required, as there are lots of different rule orderings that have to
be specified to ensure these complex addressing modes are always found
if they are possible.
Update #36468
Change-Id: I5b4bf7b03a1e731d6dfeb9ef19b376175f3b4b44
Reviewed-on: https://go-review.googlesource.com/c/go/+/217097
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
There are still two places in src/runtime/string.go that use
staticbytes, so we cannot delete it just yet.
There is a new codegen test to verify that the index calculation
is constant-folded, at least on amd64. ppc64, mips[64] and s390x
cannot currently do that.
There is also a new runtime benchmark to ensure that this does not
slow down performance (tested against parent commit):
name old time/op new time/op delta
ConvT2EByteSized/bool-4 1.07ns ± 1% 1.07ns ± 1% ~ (p=0.060 n=14+15)
ConvT2EByteSized/uint8-4 1.06ns ± 1% 1.07ns ± 1% ~ (p=0.095 n=14+15)
Updates #37612
Change-Id: I5ec30738edaa48cda78dfab4a78e24a32fa7fd6a
Reviewed-on: https://go-review.googlesource.com/c/go/+/221957
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
Trying this CL again, with a fixed test that allows platforms
to disagree on the exact behavior of converting NaNs.
We store 32-bit floating point constants in a 64-bit field, by
converting that 32-bit float to 64-bit float to store it, and convert
it back to use it.
That works for *almost* all floating-point constants. The exception is
signaling NaNs. The round trip described above means we can't represent
a 32-bit signaling NaN, because conversions strip the signaling bit.
To fix this issue, just forbid NaNs as floating-point constants in SSA
form. This shouldn't affect any real-world code, as people seldom
constant-propagate NaNs (except in test code).
Additionally, NaNs are somewhat underspecified (which of the many NaNs
do you get when dividing 0/0?), so when cross-compiling there's a
danger of using the compiler machine's NaN regime for some math, and
the target machine's NaN regime for other math. Better to use the
target machine's NaN regime always.
Update #36400
Change-Id: Idf203b688a15abceabbd66ba290d4e9f63619ecb
Reviewed-on: https://go-review.googlesource.com/c/go/+/221790
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
Ensure that any comparison between two values has the same argument
order. This helps ensure that they can be eliminated during the
lowered CSE pass which will be particularly important if we eliminate
the Greater and Geq ops (see #37316).
Example:
CMP R0, R1
BLT L1
CMP R1, R0 // different order, cannot eliminate
BEQ L2
CMP R0, R1
BLT L1
CMP R0, R1 // same order, can eliminate
BEQ L2
This does have some drawbacks. Notably comparisons might 'flip'
direction in the assembly output after even small changes to the
code or compiler. It should help make optimizations more reliable
however.
compilecmp master -> HEAD
master (218f4572f5): text/template: make reflect.Value indirections more robust
HEAD (f1661fef3e): cmd/compile: canonicalize comparison argument order
platform: linux/amd64
file before after Δ %
api 6063927 6068023 +4096 +0.068%
asm 5191757 5183565 -8192 -0.158%
cgo 4893518 4901710 +8192 +0.167%
cover 5330345 5326249 -4096 -0.077%
fix 3417778 3421874 +4096 +0.120%
pprof 14889456 14885360 -4096 -0.028%
test2json 2848138 2844042 -4096 -0.144%
trace 11746239 11733951 -12288 -0.105%
total 132739173 132722789 -16384 -0.012%
Change-Id: I11736b3fe2a4553f6fc65018f475e88217fa22f9
Reviewed-on: https://go-review.googlesource.com/c/go/+/220425
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This reverts CL 213477.
Reason for revert: tests are failing on linux-mips*-rtrk builders.
Change-Id: I8168f7450890233f1bd7e53930b73693c26d4dc0
Reviewed-on: https://go-review.googlesource.com/c/go/+/220897
Run-TryBot: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
We store 32-bit floating point constants in a 64-bit field, by
converting that 32-bit float to 64-bit float to store it, and convert
it back to use it.
That works for *almost* all floating-point constants. The exception is
signaling NaNs. The round trip described above means we can't represent
a 32-bit signaling NaN, because conversions strip the signaling bit.
To fix this issue, just forbid NaNs as floating-point constants in SSA
form. This shouldn't affect any real-world code, as people seldom
constant-propagate NaNs (except in test code).
Additionally, NaNs are somewhat underspecified (which of the many NaNs
do you get when dividing 0/0?), so when cross-compiling there's a
danger of using the compiler machine's NaN regime for some math, and
the target machine's NaN regime for other math. Better to use the
target machine's NaN regime always.
This has been a bug since 1.10, and there's an easy workaround
(declare a global varaible containing the signaling NaN pattern, and
use that as the argument to math.Float32frombits) so we'll fix it in
1.15.
Fixes#36400
Update #36399
Change-Id: Icf155e743281560eda2eed953d19a829552ccfda
Reviewed-on: https://go-review.googlesource.com/c/go/+/213477
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
The address calculations in the example end up doing x << 4 + y + 0.
Before this CL we use a SHLQ+LEAQ. Since the constant offset is 0,
we can use SHLQ+ADDQ instead.
Change-Id: Ia048c4fdbb3a42121c7e1ab707961062e8247fca
Reviewed-on: https://go-review.googlesource.com/c/go/+/209959
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This API was added for #25819, where it was discussed as math.FMA.
The commit adding it used math.Fma, presumably for consistency
with the rest of the unusual names in package math
(Sincos, Acosh, Erfcinv, Float32bits, etc).
I believe that using an idiomatic Go name is more important here
than consistency with these other names, most of which are historical
baggage from C's standard library.
Early additions like Float32frombits happened before "uppercase for export"
(so they were originally like "float32frombits") and they were not properly
reconsidered when we uppercased the symbols to export them.
That's a mistake we live with.
The names of functions we have added since then, and even a few
that were legacy, are more properly Go-cased, such as IsNaN, IsInf,
and RoundToEven, rather than Isnan, Isinf, and Roundtoeven.
And also constants like MaxFloat32.
For new API, we should keep using proper Go-cased symbols
instead of minimally-upper-cased-C symbols.
So math.FMA, not math.Fma.
This API has not yet been released, so this change does not break
the compatibility promise.
This CL also modifies cmd/compile, since the compiler knows
the name of the function. I could have stopped at changing the
string constants, but it seemed to make more sense to use a
consistent casing everywhere.
Change-Id: I0f6f3407f41e99bfa8239467345c33945088896e
Reviewed-on: https://go-review.googlesource.com/c/go/+/205317
Run-TryBot: Russ Cox <rsc@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
This change introduces an arm intrinsic that generates the FMULAD
instruction for the fused-multiply-add operation on systems that
support it. System support is detected via cpu.ARM.HasVFPv4. A rewrite
rule translates the generic intrinsic to FMULAD.
Updates #25819.
Change-Id: I8459e5dd1cdbdca35f88a78dbeb7d387f1e20efa
Reviewed-on: https://go-review.googlesource.com/c/go/+/142117
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
To permit ssa-level optimization, this change introduces an amd64 intrinsic
that generates the VFMADD231SD instruction for the fused-multiply-add
operation on systems that support it. System support is detected via
cpu.X86.HasFMA. A rewrite rule can then translate the generic ssa intrinsic
("Fma") to VFMADD231SD.
The benchmark compares the software implementation (old) with the intrinsic
(new).
name old time/op new time/op delta
Fma-4 27.2ns ± 1% 1.0ns ± 9% -96.48% (p=0.008 n=5+5)
Updates #25819.
Change-Id: I966655e5f96817a5d06dff5942418a3915b09584
Reviewed-on: https://go-review.googlesource.com/c/go/+/137156
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
In order to make math.FMA a compiler intrinsic for ISAs like ARM64,
PPC64[le], and S390X, a generic 3-argument opcode "Fma" is provided and
rewritten as
ARM64: (Fma x y z) -> (FMADDD z x y)
PPC64: (Fma x y z) -> (FMADD x y z)
S390X: (Fma x y z) -> (FMADD z x y)
Updates #25819.
Change-Id: Ie5bc628311e6feeb28ddf9adaa6e702c8c291efa
Reviewed-on: https://go-review.googlesource.com/c/go/+/131959
Run-TryBot: Akhil Indurti <aindurti@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
When a subsequent load/store of a ptr makes the nil check of that pointer
unnecessary, if their lines differ, change the line of the load/store
to that of the nilcheck, and attempt to rehome the load/store position
instead.
This fix makes profiling less accurate in order to make panics more
informative.
Fixes#33724
Change-Id: Ib9afaac12fe0d0320aea1bf493617facc34034b3
Reviewed-on: https://go-review.googlesource.com/c/go/+/200197
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
For commuting ops, check whether the second argument is dead before
checking if the first argument is rematerializeable. Reusing the register
holding a dead value is always best.
Fixes#33580
Change-Id: I7372cfc03d514e6774d2d9cc727a3e6bf6ce2657
Reviewed-on: https://go-review.googlesource.com/c/go/+/199559
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
Add a bunch of extra tests and benchmarks for defer, in preparation for new
low-cost (open-coded) implementation of defers (see #34481),
- New file defer_test.go that tests a bunch more unusual defer scenarios,
including things that might have problems for open-coded defers.
- Additions to callers_test.go actually verifying what the stack trace looks like
for various panic or panic-recover scenarios.
- Additions to crash_test.go testing several more crash scenarios involving
recursive panics.
- New benchmark in runtime_test.go measuring speed of panic-recover
- New CGo benchmark in cgo_test.go calling from Go to C back to Go that
shows defer overhead
Updates #34481
Change-Id: I423523f3e05fc0229d4277dd00073289a5526188
Reviewed-on: https://go-review.googlesource.com/c/go/+/197017
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
On modern 64bit CPUs a SHR, SHL or AND instruction take 1 cycle to execute.
A pair of shifts that operate on the same register will take 2 cycles
and needs to wait for the input register value to be available.
Large constants used to mask the high bits of a register with an AND
instruction can not be encoded as an immediate in the AND instruction
on amd64 and therefore need to be loaded into a register with a MOV
instruction.
However that MOV instruction is not dependent on the output register and
on many CPUs does not compete with the AND or shift instructions for
execution ports.
Using a pair of shifts to mask high bits instead of an AND to mask high
bits of a register has a shorter encoding and uses one less general
purpose register but is slower due to taking one clock cycle longer
if there is no register pressure that would make the AND variant need to
generate a spill.
For example the instructions emitted for (x & 1 << 63) before this CL are:
48c1ea3f SHRQ $0x3f, DX
48c1e23f SHLQ $0x3f, DX
after this CL the instructions are the same as GCC and LLVM use:
48b80000000000000080 MOVQ $0x8000000000000000, AX
4821d0 ANDQ DX, AX
Some platforms such as arm64 already have SSA optimization rules to fuse
two shift instructions back into an AND.
Removing the general rule to rewrite AND to SHR+SHL speeds up this benchmark:
var GlobalU uint
func BenchmarkAndHighBits(b *testing.B) {
x := uint(0)
for i := 0; i < b.N; i++ {
x &= 1 << 63
}
GlobalU = x
}
amd64/darwin on Intel(R) Core(TM) i7-3520M CPU @ 2.90GHz:
name old time/op new time/op delta
AndHighBits-4 0.61ns ± 6% 0.42ns ± 6% -31.42% (p=0.000 n=25+25):
'go run run.go -all_codegen -v codegen' passes with following adjustments:
ARM64: The BFXIL pattern ((x << lc) >> rc | y & ac) needed adjustment
since ORshiftRL generation fusing '>> rc' and '|' interferes
with matching ((x << lc) >> rc) to generate UBFX. Previously
ORshiftLL was created first using the shifts generated for (y & ac).
S390X: Add rules for abs and copysign to match use of AND instead of SHIFTs.
Updates #33826
Updates #32781
Change-Id: I5a59f6239660d53c029cd22dfb44ddf39f93a56c
Reviewed-on: https://go-review.googlesource.com/c/go/+/196810
Run-TryBot: Martin Möhrmann <moehrmann@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
On modern 64bit CPUs a SHR, SHL or AND instruction take 1 cycle to execute.
A pair of shifts that operate on the same register will take 2 cycles
and needs to wait for the input register value to be available.
Large constants used to mask the high bits of a register with an AND
instruction can not be encoded as an immediate in the AND instruction
on amd64 and therefore need to be loaded into a register with a MOV
instruction.
However that MOV instruction is not dependent on the output register and
on many CPUs does not compete with the AND or shift instructions for
execution ports.
Using a pair of shifts to mask high bits instead of an AND to mask high
bits of a register has a shorter encoding and uses one less general
purpose register but is slower due to taking one clock cycle longer
if there is no register pressure that would make the AND variant need to
generate a spill.
For example the instructions emitted for (x & 1 << 63) before this CL are:
48c1ea3f SHRQ $0x3f, DX
48c1e23f SHLQ $0x3f, DX
after this CL the instructions are the same as GCC and LLVM use:
48b80000000000000080 MOVQ $0x8000000000000000, AX
4821d0 ANDQ DX, AX
Some platforms such as arm64 already have SSA optimization rules to fuse
two shift instructions back into an AND.
Removing the general rule to rewrite AND to SHR+SHL speeds up this benchmark:
var GlobalU uint
func BenchmarkAndHighBits(b *testing.B) {
x := uint(0)
for i := 0; i < b.N; i++ {
x &= 1 << 63
}
GlobalU = x
}
amd64/darwin on Intel(R) Core(TM) i7-3520M CPU @ 2.90GHz:
name old time/op new time/op delta
AndHighBits-4 0.61ns ± 6% 0.42ns ± 6% -31.42% (p=0.000 n=25+25):
'go run run.go -all_codegen -v codegen' passes with following adjustments:
ARM64: The BFXIL pattern ((x << lc) >> rc | y & ac) needed adjustment
since ORshiftRL generation fusing '>> rc' and '|' interferes
with matching ((x << lc) >> rc) to generate UBFX. Previously
ORshiftLL was created first using the shifts generated for (y & ac).
S390X: Add rules for abs and copysign to match use of AND instead of SHIFTs.
Updates #33826
Updates #32781
Change-Id: I43227da76b625de03fbc51117162b23b9c678cdb
Reviewed-on: https://go-review.googlesource.com/c/go/+/194297
Run-TryBot: Martin Möhrmann <martisch@uos.de>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
i32.eqz instructions don't appear unless needed in if conditions anymore
after CL 195204. I forgot to run the codegen tests while submitting the CL.
Thanks to @martisch for catching it.
Fixes#34442
Change-Id: I177b064b389be48e39d564849714d7a8839be13e
Reviewed-on: https://go-review.googlesource.com/c/go/+/196580
Run-TryBot: Agniva De Sarker <agniva.quicksilver@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Martin Möhrmann <moehrmann@google.com>
When compiling expression switches, we try to optimize runs of
constants into binary searches. The ordering used isn't visible to the
application, so it's unimportant as long as we're consistent between
sorting and searching.
For strings, it's much cheaper to compare string lengths than strings
themselves, so instead of ordering strings by "si <= sj", we currently
order them by "len(si) < len(sj) || len(si) == len(sj) && si <= sj"
(i.e., the lexicographical ordering on the 2-tuple (len(s), s)).
However, it's also somewhat cheaper to compare strings for equality
(i.e., ==) than for ordering (i.e., <=). And if there were two or
three string constants of the same length in a switch statement, we
might unnecessarily emit ordering comparisons.
For example, given:
switch s {
case "", "1", "2", "3": // ordered by length then content
goto L
}
we currently compile this as:
if len(s) < 1 || len(s) == 1 && s <= "1" {
if s == "" { goto L }
else if s == "1" { goto L }
} else {
if s == "2" { goto L }
else if s == "3" { goto L }
}
This CL switches to using a 2-level binary search---first on len(s),
then on s itself---so that string ordering comparisons are only needed
when there are 4 or more strings of the same length. (4 being the
cut-off for when using binary search is actually worthwhile.)
So the above switch instead now compiles to:
if len(s) == 0 {
if s == "" { goto L }
} else if len(s) == 1 {
if s == "1" { goto L }
else if s == "2" { goto L }
else if s == "3" { goto L }
}
which is better optimized by walk and SSA. (Notably, because there are
only two distinct lengths and no more than three strings of any
particular length, this example ends up falling back to simply using
linear search.)
Test case by khr@ from CL 195138.
Fixes#33934.
Change-Id: I8eeebcaf7e26343223be5f443d6a97a0daf84f07
Reviewed-on: https://go-review.googlesource.com/c/go/+/195340
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
golang.org/cl/109517 optimized the compiler to avoid the allocation for make in
append(x, make([]T, y)...). This was only implemented for the case that y has type int.
This change extends the optimization to trigger for all integer types where the value
is known at compile time to fit into an int.
name old time/op new time/op delta
ExtendInt-12 106ns ± 4% 106ns ± 0% ~ (p=0.351 n=10+6)
ExtendUint64-12 1.03µs ± 5% 0.10µs ± 4% -90.01% (p=0.000 n=9+10)
name old alloc/op new alloc/op delta
ExtendInt-12 0.00B 0.00B ~ (all equal)
ExtendUint64-12 13.6kB ± 0% 0.0kB -100.00% (p=0.000 n=10+10)
name old allocs/op new allocs/op delta
ExtendInt-12 0.00 0.00 ~ (all equal)
ExtendUint64-12 1.00 ± 0% 0.00 -100.00% (p=0.000 n=10+10)
Updates #29785
Change-Id: Ief7760097c285abd591712da98c5b02bc3961fcd
Reviewed-on: https://go-review.googlesource.com/c/go/+/182559
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This reverts CL 192101.
Reason for revert: The same paragraph was added 2 weeks ago
(look a few lines above)
Change-Id: I05efb2631d7b4966f66493f178f2a649c715a3cc
Reviewed-on: https://go-review.googlesource.com/c/go/+/195637
Reviewed-by: Cherry Zhang <cherryyz@google.com>
It is useful to know about the -all_codegen option for running
codegen tests for all platforms. I was puzzling that some codegen
test was not failing on my local machine or on trybot, until I
found this option.
Change-Id: I062cf4d73f6a6c9ebc2258195779d2dab21bc36d
Reviewed-on: https://go-review.googlesource.com/c/go/+/192101
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This change adds an intrinsic for Mul64 on s390x. To achieve that,
a new assembly instruction, MLGR, is introduced in s390x/asmz.go. This assembly
instruction directly uses an existing instruction on Z and supports multiplication
of two 64 bit unsigned integer and stores the result in two separate registers.
In this case, we require the multiplcand to be stored in register R3 and
the output result (the high and low 64 bit of the product) to be stored in
R2 and R3 respectively.
A test case is also added.
Benchmark:
name old time/op new time/op delta
Mul-18 11.1ns ± 0% 1.4ns ± 0% -87.39% (p=0.002 n=8+10)
Mul32-18 2.07ns ± 0% 2.07ns ± 0% ~ (all equal)
Mul64-18 11.1ns ± 1% 1.4ns ± 0% -87.42% (p=0.000 n=10+10)
Change-Id: Ieca6ad1f61fff9a48a31d50bbd3f3c6d9e6675c1
Reviewed-on: https://go-review.googlesource.com/c/go/+/194572
Reviewed-by: Michael Munday <mike.munday@ibm.com>
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This CL gets rid of the MOVDreg and MOVDnop SSA operations on
s390x. They were originally inserted to help avoid situations
where a sign/zero extension was elided but a spill invalidated
the optimization. It's not really clear we need to do this though
(amd64 doesn't have these ops for example) so long as we are
careful when removing sign/zero extensions. Also, the MOVDreg
technique doesn't work if the register is spilled before the
MOVDreg op (I haven't seen that in practice).
Removing these ops reduces the complexity of the rules and also
allows us to unblock optimizations. For example, the compiler can
now merge the loads in binary.{Big,Little}Endian.PutUint16 which
it wasn't able to do before. This CL reduces the size of the .text
section in the go tool by about 4.7KB (0.09%).
Change-Id: Icaddae7f2e4f9b2debb6fabae845adb3f73b41db
Reviewed-on: https://go-review.googlesource.com/c/go/+/173897
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
On modern 64bit CPUs a SHR, SHL or AND instruction take 1 cycle to execute.
A pair of shifts that operate on the same register will take 2 cycles
and needs to wait for the input register value to be available.
Large constants used to mask the high bits of a register with an AND
instruction can not be encoded as an immediate in the AND instruction
on amd64 and therefore need to be loaded into a register with a MOV
instruction.
However that MOV instruction is not dependent on the output register and
on many CPUs does not compete with the AND or shift instructions for
execution ports.
Using a pair of shifts to mask high bits instead of an AND to mask high
bits of a register has a shorter encoding and uses one less general
purpose register but is slower due to taking one clock cycle longer
if there is no register pressure that would make the AND variant need to
generate a spill.
For example the instructions emitted for (x & 1 << 63) before this CL are:
48c1ea3f SHRQ $0x3f, DX
48c1e23f SHLQ $0x3f, DX
after this CL the instructions are the same as GCC and LLVM use:
48b80000000000000080 MOVQ $0x8000000000000000, AX
4821d0 ANDQ DX, AX
Some platforms such as arm64 already have SSA optimization rules to fuse
two shift instructions back into an AND.
Removing the general rule to rewrite AND to SHR+SHL speeds up this benchmark:
var GlobalU uint
func BenchmarkAndHighBits(b *testing.B) {
x := uint(0)
for i := 0; i < b.N; i++ {
x &= 1 << 63
}
GlobalU = x
}
amd64/darwin on Intel(R) Core(TM) i7-3520M CPU @ 2.90GHz:
name old time/op new time/op delta
AndHighBits-4 0.61ns ± 6% 0.42ns ± 6% -31.42% (p=0.000 n=25+25):
Updates #33826
Updates #32781
Change-Id: I862d3587446410c447b9a7265196b57f85358633
Reviewed-on: https://go-review.googlesource.com/c/go/+/191780
Run-TryBot: Martin Möhrmann <moehrmann@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
For performance reasons (avoiding costly cross-compilations) CL 177577
changed the codegen test harness to only run the tests for the
machine's GOARCH by default.
This change updates the codegen README accordingly, explaining what
all.bash does run by default and how to perform the tests for all
architectures.
Fixes#33924
Change-Id: I43328d878c3e449ebfda46f7e69963a44a511d40
Reviewed-on: https://go-review.googlesource.com/c/go/+/192619
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
This CL eliminates unnecessary pairs of I32WrapI64 and
I64ExtendI32U generated by the WASM backend for IF
statements. And it makes the total size of pkg/js_wasm/
decreases about 490KB.
Change-Id: I16b0abb686c4e30d5624323166ec2d0ec57dbe2d
Reviewed-on: https://go-review.googlesource.com/c/go/+/191758
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Richard Musiol <neelance@gmail.com>
This CL reverts CL 192097 and fixes the issue in CL 189277.
Change-Id: Icd271262e1f5019a8e01c91f91c12c1261eeb02b
Reviewed-on: https://go-review.googlesource.com/c/go/+/192519
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
The syntax of a shifted operation does not have a "$" sign for
the shift amount. Remove it.
Change-Id: I50782fe942b640076f48c2fafea4d3175be8ff99
Reviewed-on: https://go-review.googlesource.com/c/go/+/192100
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Extend the optimization introduced in CL 141118 to the wasm architecture.
And for reference, the rules trigger 212 times while building std and cmd
$GOOS=js GOARCH=wasm gotip build std cmd
$grep -E "Wasm.rules:44(1|2|3|4)" rulelog | wc -l
212
Updates #26498
Change-Id: I153684a2b98589ae812b42268da08b65679e09d1
Reviewed-on: https://go-review.googlesource.com/c/go/+/185477
Run-TryBot: Agniva De Sarker <agniva.quicksilver@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Richard Musiol <neelance@gmail.com>
Use the shiftIsBounded function to generate more efficient
Shift instructions.
Updates #25167
Change-Id: Id350f8462dc3a7ed3bfed0bcbea2860b8f40048a
Reviewed-on: https://go-review.googlesource.com/c/go/+/182558
Run-TryBot: Agniva De Sarker <agniva.quicksilver@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Richard Musiol <neelance@gmail.com>
This CL optimizes math.bits.TrailingZeros16 on 386 with
a pair of BSFL and ORL instrcutions.
The case TrailingZeros16-4 of the benchmark test in
math/bits shows big improvement.
name old time/op new time/op delta
TrailingZeros16-4 1.55ns ± 1% 0.87ns ± 1% -43.87% (p=0.000 n=50+49)
Change-Id: Ia899975b0e46f45dcd20223b713ed632bc32740b
Reviewed-on: https://go-review.googlesource.com/c/go/+/189277
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This CL performs the branchelim optimization on WASM with its
select instruction. And the total size of pkg/js_wasm decreased
about 80KB by this optimization.
Change-Id: I868eb146120a1cac5c4609c8e9ddb07e4da8a1d9
Reviewed-on: https://go-review.googlesource.com/c/go/+/190957
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Richard Musiol <neelance@gmail.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
The assembly output for x & c == c, where c is power of 2:
MOVQ "".set+8(SP), AX
ANDQ $8, AX
CMPQ AX, $8
SETEQ "".~r2+24(SP)
With optimization using bitset:
MOVQ "".set+8(SP), AX
BTL $3, AX
SETCS "".~r2+24(SP)
output less than 1 instruction.
However, there is no speed improvement:
name old time/op new time/op delta
AllBitSet-8 0.35ns ± 0% 0.35ns ± 0% ~ (all equal)
Fixes#31904
Change-Id: I5dca4e410bf45716ed2145e3473979ec997e35d4
Reviewed-on: https://go-review.googlesource.com/c/go/+/175957
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This reverts CL 180761
Reason for revert: Reinstate the stack-allocated defer CL.
There was nothing wrong with the CL proper, but stack allocation of defers exposed two other issues.
Issue #32477: Fix has been submitted as CL 181258.
Issue #32498: Possible fix is CL 181377 (not submitted yet).
Change-Id: I32b3365d5026600069291b068bbba6cb15295eb3
Reviewed-on: https://go-review.googlesource.com/c/go/+/181378
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
This reverts commit fff4f599fe.
Reason for revert: Seems to still have issues around GC.
Fixes#32452
Change-Id: Ibe7af629f9ad6a3d5312acd7b066123f484da7f0
Reviewed-on: https://go-review.googlesource.com/c/go/+/180761
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
When a defer is executed at most once in a function body,
we can allocate the defer record for it on the stack instead
of on the heap.
This should make defers like this (which are very common) faster.
This optimization applies to 363 out of the 370 static defer sites
in the cmd/go binary.
name old time/op new time/op delta
Defer-4 52.2ns ± 5% 36.2ns ± 3% -30.70% (p=0.000 n=10+10)
Fixes#6980
Update #14939
Change-Id: I697109dd7aeef9e97a9eeba2ef65ff53d3ee1004
Reviewed-on: https://go-review.googlesource.com/c/go/+/171758
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
The CL 164718 mistyped the comparison flags. The rules for floating
point comparison should be GreaterThanF and GreaterEqualF. Fortunately,
the wrong optimizations were overwritten by other integer rules, so the
issue won't cause failure but just some performance impact.
The fixed CL optimizes the floating point test as follows.
source code: func foo(f float64) bool { return f > 4 || f < -4}
previous version: "FCMPD", "CSET\tGT", "CBZ"
fixed version: "FCMPD", BLE"
Add the test case.
Change-Id: Iea954fdbb8272b2d642dae0f816dc77286e6e1fa
Reviewed-on: https://go-review.googlesource.com/c/go/+/177121
Reviewed-by: Ben Shi <powerman1st@163.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This CL adds intrinsics for the 64-bit addition and subtraction
functions in math/bits. These intrinsics use the condition code
to propagate the carry or borrow bit.
To make the carry chains more efficient I've removed the
'clobberFlags' property from most of the load and store
operations. Originally these ops did clobber flags when using
offsets that didn't fit in a signed 20-bit integer, however
that is no longer true.
As with other platforms the intrinsics are faster when executed
in a chain rather than a loop because currently we need to spill
and restore the carry bit between each loop iteration. We may
be able to reduce the need to do this on s390x (e.g. by using
compare-and-branch instructions that do not clobber flags) in the
future.
name old time/op new time/op delta
Add64 1.21ns ± 2% 2.03ns ± 2% +67.18% (p=0.000 n=7+10)
Add64multiple 2.98ns ± 3% 1.03ns ± 0% -65.39% (p=0.000 n=10+9)
Sub64 1.23ns ± 4% 2.03ns ± 1% +64.85% (p=0.000 n=10+10)
Sub64multiple 3.73ns ± 4% 1.04ns ± 1% -72.28% (p=0.000 n=10+8)
Change-Id: I913bbd5e19e6b95bef52f5bc4f14d6fe40119083
Reviewed-on: https://go-review.googlesource.com/c/go/+/174303
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This enables more of the testcases in memcombine for ppc64le,
and adds more detail to some existing.
Change-Id: Ic522a1175bed682b546909c96f9ea758f8db247c
Reviewed-on: https://go-review.googlesource.com/c/go/+/174737
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
"Division by invariant integers using multiplication" paper
by Granlund and Montgomery contains a method for directly computing
divisibility (x%c == 0 for c constant) by means of the modular inverse.
The method is further elaborated in "Hacker's Delight" by Warren Section 10-17
This general rule can compute divisibilty by one multiplication, and add
and a compare for odd divisors and an additional rotate for even divisors.
To apply the divisibility rule, we must take into account
the rules to rewrite x%c = x-((x/c)*c) and (x/c) for c constant on the first
optimization pass "opt". This complicates the matching as we want to match
only in the cases where the result of (x/c) is not also needed.
So, we must match on the expanded form of (x/c) in the expression x == c*(x/c)
in the "late opt" pass after common subexpresion elimination.
Note, that if there is an intermediate opt pass introduced in the future we
could simplify these rules by delaying the magic division rewrite to "late opt"
and matching directly on (x/c) in the intermediate opt pass.
On amd64, the divisibility check is 30-45% faster.
name old time/op new time/op delta`
DivisiblePow2constI64-4 0.83ns ± 1% 0.82ns ± 0% ~ (p=0.079 n=5+4)
DivisibleconstI64-4 2.68ns ± 1% 1.87ns ± 0% -30.33% (p=0.000 n=5+4)
DivisibleWDivconstI64-4 2.69ns ± 1% 2.71ns ± 3% ~ (p=1.000 n=5+5)
DivisiblePow2constI32-4 1.15ns ± 1% 1.15ns ± 0% ~ (p=0.238 n=5+4)
DivisibleconstI32-4 2.24ns ± 1% 1.20ns ± 0% -46.48% (p=0.016 n=5+4)
DivisibleWDivconstI32-4 2.27ns ± 1% 2.27ns ± 1% ~ (p=0.683 n=5+5)
DivisiblePow2constI16-4 0.81ns ± 1% 0.82ns ± 1% ~ (p=0.135 n=5+5)
DivisibleconstI16-4 2.11ns ± 2% 1.20ns ± 1% -42.99% (p=0.008 n=5+5)
DivisibleWDivconstI16-4 2.23ns ± 0% 2.27ns ± 2% +1.79% (p=0.029 n=4+4)
DivisiblePow2constI8-4 0.81ns ± 1% 0.81ns ± 1% ~ (p=0.286 n=5+5)
DivisibleconstI8-4 2.13ns ± 3% 1.19ns ± 1% -43.84% (p=0.008 n=5+5)
DivisibleWDivconstI8-4 2.23ns ± 1% 2.25ns ± 1% ~ (p=0.183 n=5+5)
Fixes#30282Fixes#15806
Change-Id: Id20d78263a4fdfe0509229ae4dfa2fede83fc1d0
Reviewed-on: https://go-review.googlesource.com/c/go/+/173998
Run-TryBot: Brian Kessler <brian.m.kessler@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This change creates an intrinsic for Add64 for ppc64x and adds a
testcase for it.
name old time/op new time/op delta
Add64-160 1.90ns ±40% 2.29ns ± 0% ~ (p=0.119 n=5+5)
Add64multiple-160 6.69ns ± 2% 2.45ns ± 4% -63.47% (p=0.016 n=4+5)
Change-Id: I9abe6fb023fdf62eea3c9b46a1820f60bb0a7f97
Reviewed-on: https://go-review.googlesource.com/c/go/+/173758
Reviewed-by: Lynn Boger <laboger@linux.vnet.ibm.com>
Run-TryBot: Carlos Eduardo Seo <cseo@linux.vnet.ibm.com>
"Division by invariant integers using multiplication" paper
by Granlund and Montgomery contains a method for directly computing
divisibility (x%c == 0 for c constant) by means of the modular inverse.
The method is further elaborated in "Hacker's Delight" by Warren Section 10-17
This general rule can compute divisibilty by one multiplication and a compare
for odd divisors and an additional rotate for even divisors.
To apply the divisibility rule, we must take into account
the rules to rewrite x%c = x-((x/c)*c) and (x/c) for c constant on the first
optimization pass "opt". This complicates the matching as we want to match
only in the cases where the result of (x/c) is not also available.
So, we must match on the expanded form of (x/c) in the expression x == c*(x/c)
in the "late opt" pass after common subexpresion elimination.
Note, that if there is an intermediate opt pass introduced in the future we
could simplify these rules by delaying the magic division rewrite to "late opt"
and matching directly on (x/c) in the intermediate opt pass.
Additional rules to lower the generic RotateLeft* ops were also applied.
On amd64, the divisibility check is 25-50% faster.
name old time/op new time/op delta
DivconstI64-4 2.08ns ± 0% 2.08ns ± 1% ~ (p=0.881 n=5+5)
DivisibleconstI64-4 2.67ns ± 0% 2.67ns ± 1% ~ (p=1.000 n=5+5)
DivisibleWDivconstI64-4 2.67ns ± 0% 2.67ns ± 0% ~ (p=0.683 n=5+5)
DivconstU64-4 2.08ns ± 1% 2.08ns ± 1% ~ (p=1.000 n=5+5)
DivisibleconstU64-4 2.77ns ± 1% 1.55ns ± 2% -43.90% (p=0.008 n=5+5)
DivisibleWDivconstU64-4 2.99ns ± 1% 2.99ns ± 1% ~ (p=1.000 n=5+5)
DivconstI32-4 1.53ns ± 2% 1.53ns ± 0% ~ (p=1.000 n=5+5)
DivisibleconstI32-4 2.23ns ± 0% 2.25ns ± 3% ~ (p=0.167 n=5+5)
DivisibleWDivconstI32-4 2.27ns ± 1% 2.27ns ± 1% ~ (p=0.429 n=5+5)
DivconstU32-4 1.78ns ± 0% 1.78ns ± 1% ~ (p=1.000 n=4+5)
DivisibleconstU32-4 2.52ns ± 2% 1.26ns ± 0% -49.96% (p=0.000 n=5+4)
DivisibleWDivconstU32-4 2.63ns ± 0% 2.85ns ±10% +8.29% (p=0.016 n=4+5)
DivconstI16-4 1.54ns ± 0% 1.54ns ± 0% ~ (p=0.333 n=4+5)
DivisibleconstI16-4 2.10ns ± 0% 2.10ns ± 1% ~ (p=0.571 n=4+5)
DivisibleWDivconstI16-4 2.22ns ± 0% 2.23ns ± 1% ~ (p=0.556 n=4+5)
DivconstU16-4 1.09ns ± 0% 1.01ns ± 1% -7.74% (p=0.000 n=4+5)
DivisibleconstU16-4 1.83ns ± 0% 1.26ns ± 0% -31.52% (p=0.008 n=5+5)
DivisibleWDivconstU16-4 1.88ns ± 0% 1.89ns ± 1% ~ (p=0.365 n=5+5)
DivconstI8-4 1.54ns ± 1% 1.54ns ± 1% ~ (p=1.000 n=5+5)
DivisibleconstI8-4 2.10ns ± 0% 2.11ns ± 0% ~ (p=0.238 n=5+4)
DivisibleWDivconstI8-4 2.22ns ± 0% 2.23ns ± 2% ~ (p=0.762 n=5+5)
DivconstU8-4 0.92ns ± 1% 0.94ns ± 1% +2.65% (p=0.008 n=5+5)
DivisibleconstU8-4 1.66ns ± 0% 1.26ns ± 1% -24.28% (p=0.008 n=5+5)
DivisibleWDivconstU8-4 1.79ns ± 0% 1.80ns ± 1% ~ (p=0.079 n=4+5)
A follow-up change will address the signed division case.
Updates #30282
Change-Id: I7e995f167179aa5c76bb10fbcbeb49c520943403
Reviewed-on: https://go-review.googlesource.com/c/go/+/168037
Run-TryBot: Brian Kessler <brian.m.kessler@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
For powers of two (c=1<<k), the divisibility check x%c == 0 can be made
just by checking the trailing zeroes via a mask x&(c-1) == 0 even for signed
integers. This avoids division fix-ups when just divisibility check is needed.
To apply this rule, we match on the fixed-up version of the division. This is
neccessary because the mod and division rewrite rules are already applied
during the initial opt pass.
The speed up on amd64 due to elimination of unneccessary fix-up code is ~55%:
name old time/op new time/op delta
DivconstI64-4 2.08ns ± 0% 2.09ns ± 1% ~ (p=0.730 n=5+5)
DivisiblePow2constI64-4 1.78ns ± 1% 0.81ns ± 1% -54.66% (p=0.008 n=5+5)
DivconstU64-4 2.08ns ± 0% 2.08ns ± 0% ~ (p=0.683 n=5+5)
DivconstI32-4 1.53ns ± 0% 1.53ns ± 1% ~ (p=0.968 n=4+5)
DivisiblePow2constI32-4 1.79ns ± 1% 0.81ns ± 1% -54.97% (p=0.008 n=5+5)
DivconstU32-4 1.78ns ± 1% 1.80ns ± 2% ~ (p=0.206 n=5+5)
DivconstI16-4 1.54ns ± 2% 1.54ns ± 0% ~ (p=0.238 n=5+4)
DivisiblePow2constI16-4 1.78ns ± 0% 0.81ns ± 1% -54.72% (p=0.000 n=4+5)
DivconstU16-4 1.00ns ± 5% 1.01ns ± 1% ~ (p=0.119 n=5+5)
DivconstI8-4 1.54ns ± 0% 1.54ns ± 2% ~ (p=0.571 n=4+5)
DivisiblePow2constI8-4 1.78ns ± 0% 0.82ns ± 8% -53.71% (p=0.008 n=5+5)
DivconstU8-4 0.93ns ± 1% 0.93ns ± 1% ~ (p=0.643 n=5+5)
A follow-up CL will address the general case of x%c == 0 for signed integers.
Updates #15806
Change-Id: Iabadbbe369b6e0998c8ce85d038ebc236142e42a
Reviewed-on: https://go-review.googlesource.com/c/go/+/173557
Run-TryBot: Brian Kessler <brian.m.kessler@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This reverts CL 168038 (git 68819fb6d2)
Reason for revert: Doesn't work on 32 bit archs.
Change-Id: Idec9098060dc65bc2f774c5383f0477f8eb63a3d
Reviewed-on: https://go-review.googlesource.com/c/go/+/173442
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
For powers of two (c=1<<k), the divisibility check x%c == 0 can be made
just by checking the trailing zeroes via a mask x&(c-1)==0 even for signed
integers. This avoids division fixups when just divisibility check is needed.
To apply this rule the generic divisibility rule for A%B = A-(A/B*B) is disabled
on the "opt" pass, but this does not affect generated code as this rule is applied
later.
The speed up on amd64 due to elimination of unneccessary fixup code is ~55%:
name old time/op new time/op delta
DivconstI64-4 2.08ns ± 0% 2.07ns ± 0% ~ (p=0.079 n=5+5)
DivisiblePow2constI64-4 1.78ns ± 1% 0.81ns ± 1% -54.55% (p=0.008 n=5+5)
DivconstU64-4 2.08ns ± 0% 2.08ns ± 0% ~ (p=1.000 n=5+5)
DivconstI32-4 1.53ns ± 0% 1.53ns ± 0% ~ (all equal)
DivisiblePow2constI32-4 1.79ns ± 1% 0.81ns ± 4% -54.75% (p=0.008 n=5+5)
DivconstU32-4 1.78ns ± 1% 1.78ns ± 1% ~ (p=1.000 n=5+5)
DivconstI16-4 1.54ns ± 2% 1.53ns ± 0% ~ (p=0.333 n=5+4)
DivisiblePow2constI16-4 1.78ns ± 0% 0.79ns ± 1% -55.39% (p=0.000 n=4+5)
DivconstU16-4 1.00ns ± 5% 0.99ns ± 1% ~ (p=0.730 n=5+5)
DivconstI8-4 1.54ns ± 0% 1.53ns ± 0% ~ (p=0.714 n=4+5)
DivisiblePow2constI8-4 1.78ns ± 0% 0.80ns ± 0% -55.06% (p=0.000 n=5+4)
DivconstU8-4 0.93ns ± 1% 0.95ns ± 1% +1.72% (p=0.024 n=5+5)
A follow-up CL will address the general case of x%c == 0 for signed integers.
Updates #15806
Change-Id: I0d284863774b1bc8c4ce87443bbaec6103e14ef4
Reviewed-on: https://go-review.googlesource.com/c/go/+/168038
Reviewed-by: Keith Randall <khr@golang.org>
In 31618, we end up comparing the is-stmt-ness of positions
to repurpose real instructions as inline marks. If the is-stmt-ness
doesn't match, we end up not being able to remove the inline mark.
Always use statement-full positions to do the matching, so we
always find a match if there is one.
Also always use positions that are statements for inline marks.
Fixes#31618
Change-Id: Idaf39bdb32fa45238d5cd52973cadf4504f947d5
Reviewed-on: https://go-review.googlesource.com/c/go/+/173324
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: David Chase <drchase@google.com>
These new calls should not prevent NOSPLIT promotion, like the old ones.
These new calls should not prevent racefuncenter/exit removal.
(The latter was already true, as the new calls are not yet lowered
to StaticCalls at the point where racefuncenter/exit removal is done.)
Add tests to make sure we don't regress (again).
Fixes#31219
Change-Id: I3fb6b17cdd32c425829f1e2498defa813a5a9ace
Reviewed-on: https://go-review.googlesource.com/c/go/+/170639
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ilya Tocar <ilya.tocar@intel.com>
This CL instrinsifies Add64 with arm64 instruction sequence ADDS, ADCS
and ADC, and optimzes the case of carry chains.The CL also changes the
test code so that the intrinsic implementation can be tested.
Benchmarks:
name old time/op new time/op delta
Add-224 2.500000ns +- 0% 2.090000ns +- 4% -16.40% (p=0.000 n=9+10)
Add32-224 2.500000ns +- 0% 2.500000ns +- 0% ~ (all equal)
Add64-224 2.500000ns +- 0% 1.577778ns +- 2% -36.89% (p=0.000 n=10+9)
Add64multiple-224 6.000000ns +- 0% 2.000000ns +- 0% -66.67% (p=0.000 n=10+10)
Change-Id: I6ee91c9a85c16cc72ade5fd94868c579f16c7615
Reviewed-on: https://go-review.googlesource.com/c/go/+/159017
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
'SUBQ $-0x80, r' is shorter to encode than 'ADDQ $0x80, r',
and functionally equivalent. Use it instead.
Shaves off a few bytes here and there:
file before after Δ %
compile 25935856 25927664 -8192 -0.032%
nm 4251840 4247744 -4096 -0.096%
Change-Id: Ia9e02ea38cbded6a52a613b92e3a914f878d931e
Reviewed-on: https://go-review.googlesource.com/c/go/+/168344
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
A few examples (for accessing a slice of length 3):
s[-1] runtime error: index out of range [-1]
s[3] runtime error: index out of range [3] with length 3
s[-1:0] runtime error: slice bounds out of range [-1:]
s[3:0] runtime error: slice bounds out of range [3:0]
s[3:-1] runtime error: slice bounds out of range [:-1]
s[3:4] runtime error: slice bounds out of range [:4] with capacity 3
s[0:3:4] runtime error: slice bounds out of range [::4] with capacity 3
Note that in cases where there are multiple things wrong with the
indexes (e.g. s[3:-1]), we report one of those errors kind of
arbitrarily, currently the rightmost one.
An exhaustive set of examples is in issue30116[u].out in the CL.
The message text has the same prefix as the old message text. That
leads to slightly awkward phrasing but hopefully minimizes the chance
that code depending on the error text will break.
Increases the size of the go binary by 0.5% (amd64). The panic functions
take arguments in registers in order to keep the size of the compiled code
as small as possible.
Fixes#30116
Change-Id: Idb99a827b7888822ca34c240eca87b7e44a04fdd
Reviewed-on: https://go-review.googlesource.com/c/go/+/161477
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
We know that a & 31 is non-negative for all a, signed or not.
We can avoid checking that and needing to write out an
unreachable call to panicshift.
Change-Id: I32f32fb2c950d2b2b35ac5c0e99b7b2dbd47f917
Reviewed-on: https://go-review.googlesource.com/c/go/+/167499
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Keith Randall <khr@golang.org>
Two tests (load_le_byte8_uint64_inv and load_be_byte8_uint64)
pass but the generated code isn't actually correct.
The test regexp provides a false negative, as it matches the
MOVQ (SP), BP instruction in the epilogue.
Combined loads never worked for these cases - the test was added in error
as part of a batch and not noticed because of the above false match.
Normalize the amd64/386 tests to always negative match on narrower
loads and OR.
Change-Id: I256861924774d39db0e65723866c81df5ab5076f
Reviewed-on: https://go-review.googlesource.com/c/go/+/166837
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Current compiler reverses operands to work around NaN in
"less than" and "less equal than" comparisons. But if we
want to use "FCMPD/FCMPS $(0.0), Fn" to do some optimization,
the workaround way does not work. Because assembler does
not support instruction "FCMPD/FCMPS Fn, $(0.0)".
This CL sets condition flags for floating-point comparisons
to resolve this problem.
Change-Id: Ia48076a1da95da64596d6e68304018cb301ebe33
Reviewed-on: https://go-review.googlesource.com/c/go/+/164718
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This CL adds two rules to turn patterns like ((x<<8) | (x>>8)) (the type of
x is uint16, "|" can also be "+" or "^") to a REV16 instruction on arm v6+.
This optimization rule can be used for math/bits.ReverseBytes16.
Benchmarks on arm v6:
name old time/op new time/op delta
ReverseBytes-32 2.86ns ± 0% 2.86ns ± 0% ~ (all equal)
ReverseBytes16-32 2.86ns ± 0% 2.86ns ± 0% ~ (all equal)
ReverseBytes32-32 1.29ns ± 0% 1.29ns ± 0% ~ (all equal)
ReverseBytes64-32 1.43ns ± 0% 1.43ns ± 0% ~ (all equal)
Change-Id: I819e633c9a9d308f8e476fb0c82d73fb73dd019f
Reviewed-on: https://go-review.googlesource.com/c/go/+/159019
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Emit &runtime.zerobase instead of a call to newobject for
allocations of zero sized objects in walk.go.
Fixes#29446
Change-Id: I11b67981d55009726a17c2e582c12ce0c258682e
Reviewed-on: https://go-review.googlesource.com/c/155840
Run-TryBot: Iskander Sharipov <quasilyte@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
var a []int = ...
p := &a[0]
_ = *p
We don't need to nil check on the 3rd line. If the bounds check on the 2nd
line passes, we know p is non-nil.
We rely on the fact that any cap>0 slice has a non-nil pointer as its
pointer to the backing array. This is true for all safely-constructed slices,
and I don't see any reason why someone would violate this rule using unsafe.
R=go1.13
Fixes#30366
Change-Id: I3ed764fcb72cfe1fbf963d8c1a82e24e3b6dead7
Reviewed-on: https://go-review.googlesource.com/c/163740
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
If someone takes a pointer to a zero-sized stack variable, it can
be incorrectly interpreted as a pointer to the next object in the
stack frame. To avoid this, add some padding after zero-sized variables.
We only need to pad if the next variable in memory (which is the
previous variable in the order in which we allocate variables to the
stack frame) has pointers. If the next variable has no pointers, it
won't hurt to have a pointer to it.
Because we allocate all pointer-containing variables before all
non-pointer-containing variables, we should only have to pad once per
frame.
Fixes#24993
Change-Id: Ife561cdfdf964fdbf69af03ae6ba97d004e6193c
Reviewed-on: https://go-review.googlesource.com/c/155698
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This CL adds several test cases of arithmetic operations for
386/amd64/arm/arm64.
Change-Id: I362687c06249f31091458a1d8c45fc4d006b616a
Reviewed-on: https://go-review.googlesource.com/c/151897
Run-TryBot: Ben Shi <powerman1st@163.com>
Reviewed-by: Keith Randall <khr@golang.org>
We want to issue loads as soon as possible, especially when they
are going to miss in the cache. Using a conditional move (CMOV) here:
i := ...
if cond {
i++
}
... = a[i]
means that we have to wait for cond to be computed before the load
is issued. Without a CMOV, if the branch is predicted correctly the
load can be issued in parallel with computing cond.
Even if the branch is predicted incorrectly, maybe the speculative
load is close to the real load, and we get a prefetch for free.
In the worst case, when the prediction is wrong and the address is
way off, we only lose by the time difference between the CMOV
latency (~2 cycles) and the mispredict restart latency (~15 cycles).
We only squash CMOVs that affect load addresses. Results of CMOVs
that are used for other things (store addresses, store values) we
use as before.
Fixes#26306
Change-Id: I82ca14b664bf05e1d45e58de8c4d9c775a127ca1
Reviewed-on: https://go-review.googlesource.com/c/145717
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
Note that the intrinsic implementation panics separately for overflow and
divide by zero, which matches the behavior of the pure go implementation.
There is a modest performance improvement after intrinsic implementation.
name old time/op new time/op delta
Div-4 53.0ns ± 1% 47.0ns ± 0% -11.28% (p=0.008 n=5+5)
Div32-4 18.4ns ± 0% 18.5ns ± 1% ~ (p=0.444 n=5+5)
Div64-4 53.3ns ± 0% 47.5ns ± 4% -10.77% (p=0.008 n=5+5)
Updates #28273
Change-Id: Ic1688ecc0964acace2e91bf44ef16f5fb6b6bc82
Reviewed-on: https://go-review.googlesource.com/c/144378
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
The current support_XXX variables are specific for the
amd64 and 386 platforms.
Prefix processor capability variables by architecture to have a
consistent naming scheme and avoid reuse of the existing
variables for new platforms.
This also aligns naming of runtime variables closer with internal/cpu
processor capability variable names.
Change-Id: I3eabb29a03874678851376185d3a62e73c1aff1d
Reviewed-on: https://go-review.googlesource.com/c/91435
Run-TryBot: Martin Möhrmann <martisch@uos.de>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This change makes use of the cc versions of the AND, OR, XOR
instructions, omitting the need for a CMP instruction.
In many test programs and in the go binary, this reduces the
size of 20-30 functions by at least 1 instruction, many in
runtime.
Testcase added to test/codegen/comparisons.go
Change-Id: I6cc1ca8b80b065d7390749c625bc9784b0039adb
Reviewed-on: https://go-review.googlesource.com/c/143059
Reviewed-by: Carlos Eduardo Seo <cseo@linux.vnet.ibm.com>
Reviewed-by: Michael Munday <mike.munday@ibm.com>
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
For moves >8,<16 bytes, do a move using non-overlapping loads/stores
if it would require no more instructions.
This helps a bit with the case when the move is from a static
constant, because then the code to materialize the value being moved
is smaller.
Change-Id: Ie47a5a7c654afeb4973142b0a9922faea13c9b54
Reviewed-on: https://go-review.googlesource.com/c/146019
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This CL fixes several typos and adds two more cases
to arithmetic test.
Change-Id: I086560162ea351e2166866e444e2317da36c1729
Reviewed-on: https://go-review.googlesource.com/c/145210
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Unlike normal load+op opcodes, the load+compare opcode does
not clobber its non-load argument. Allow the load+compare merge
to happen even if the non-load argument is used elsewhere.
Noticed when investigating issue #28417.
Change-Id: Ibc48d1f2e06ae76034c59f453815d263e8ec7288
Reviewed-on: https://go-review.googlesource.com/c/145097
Reviewed-by: Ainar Garipov <gugl.zadolbal@gmail.com>
Reviewed-by: Ben Shi <powerman1st@163.com>
prove is able to find 94 occurrences in std cmd where a divisor
can't have the value -1. The change removes
the extraneous fix-up code for these cases.
Fixes#25239
Change-Id: Ic184de971f47cc57c702eb72805b8e291c14035d
Reviewed-on: https://go-review.googlesource.com/c/130215
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This CL add 3 rules to combine byte-store to word-store on386 and
amd64.
Change-Id: Iffd9cda42f1961680c81def4edc773ad58f211b3
Reviewed-on: https://go-review.googlesource.com/c/143057
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This CL adds more combined load/store test cases for 386/amd64.
Change-Id: I0a483a6ed0212b65c5e84d67ed8c9f50c389ce2d
Reviewed-on: https://go-review.googlesource.com/c/142878
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
ARMv7's MULAF/MULSF/MULAD/MULSD are not fused,
this CL fixes the confusing test cases.
Change-Id: I35022e207e2f0d24a23a7f6f188e41ba8eee9886
Reviewed-on: https://go-review.googlesource.com/c/142439
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Akhil Indurti <aindurti@gmail.com>
Reviewed-by: Giovanni Bajo <rasky@develer.com>
x = map[string(byteslice)] is already optimized by the compiler to avoid a
string allocation. This CL generalizes this optimization to:
x = map[T1{ ... Tn{..., string(byteslice), ...} ... }]
where T1 to Tn is a nesting of struct and array literals.
Found in a hot code path that used a struct of strings made from []byte
slices to make a map lookup.
There are no uses of the more generalized optimization in the standard library.
Passes toolstash -cmp.
MapStringConversion/32/simple 21.9ns ± 2% 21.9ns ± 3% ~ (p=0.995 n=17+20)
MapStringConversion/32/struct 28.8ns ± 3% 22.0ns ± 2% -23.80% (p=0.000 n=20+20)
MapStringConversion/32/array 28.5ns ± 2% 21.9ns ± 2% -23.14% (p=0.000 n=19+16)
MapStringConversion/64/simple 21.0ns ± 2% 21.1ns ± 3% ~ (p=0.072 n=19+18)
MapStringConversion/64/struct 72.4ns ± 3% 21.3ns ± 2% -70.53% (p=0.000 n=20+20)
MapStringConversion/64/array 72.8ns ± 1% 21.0ns ± 2% -71.13% (p=0.000 n=17+19)
name old allocs/op new allocs/op delta
MapStringConversion/32/simple 0.00 0.00 ~ (all equal)
MapStringConversion/32/struct 0.00 0.00 ~ (all equal)
MapStringConversion/32/array 0.00 0.00 ~ (all equal)
MapStringConversion/64/simple 0.00 0.00 ~ (all equal)
MapStringConversion/64/struct 1.00 ± 0% 0.00 -100.00% (p=0.000 n=20+20)
MapStringConversion/64/array 1.00 ± 0% 0.00 -100.00% (p=0.000 n=20+20)
Change-Id: I483b4d84d8d74b1025b62c954da9a365e79b7a3a
Reviewed-on: https://go-review.googlesource.com/c/116275
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This change adds codegen tests for the intrinsification on ppc64 of
the OnesCount{64,32,16,8}, and TrailingZeros{64,32,16,8} math/bits
functions.
Change-Id: Id3364921fbd18316850e15c8c71330c906187fdb
Reviewed-on: https://go-review.googlesource.com/c/141897
Reviewed-by: Lynn Boger <laboger@linux.vnet.ibm.com>
This CL adds tests of fused multiplication-accumulation
on arm/arm64.
Change-Id: Ic85d5277c0d6acb7e1e723653372dfaf96824a39
Reviewed-on: https://go-review.googlesource.com/c/141652
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Instead of
MOVB go.string."foo"(SB), AX
do
MOVB $102, AX
When we know the global we're loading from is readonly, we can
do that read at compile time.
I've made this arch-dependent mostly because the cases where this
happens often are memory->memory moves, and those don't get
decomposed until lowering.
Did amd64/386/arm/arm64. Other architectures could follow.
Update #26498
Change-Id: I41b1dc831b2cd0a52dac9b97f4f4457888a46389
Reviewed-on: https://go-review.googlesource.com/c/141118
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
Do []byte(string) conversions more efficiently when the string
is a constant. Instead of calling stringtobyteslice, allocate
just the space we need and encode the initialization directly.
[]byte("foo") rewrites to the following pseudocode:
var s [3]byte // on heap or stack, depending on whether b escapes
s = *(*[3]byte)(&"foo"[0]) // initialize s from the string
b = s[:]
which generates this assembly:
0x001d 00029 (tmp1.go:9) LEAQ type.[3]uint8(SB), AX
0x0024 00036 (tmp1.go:9) MOVQ AX, (SP)
0x0028 00040 (tmp1.go:9) CALL runtime.newobject(SB)
0x002d 00045 (tmp1.go:9) MOVQ 8(SP), AX
0x0032 00050 (tmp1.go:9) MOVBLZX go.string."foo"+2(SB), CX
0x0039 00057 (tmp1.go:9) MOVWLZX go.string."foo"(SB), DX
0x0040 00064 (tmp1.go:9) MOVW DX, (AX)
0x0043 00067 (tmp1.go:9) MOVB CL, 2(AX)
// Then the slice is b = {AX, 3, 3}
The generated code is still not optimal, as it still does load/store
from read-only memory instead of constant stores. Next CL...
Update #26498Fixes#10170
Change-Id: I4b990b19f9a308f60c8f4f148934acffefe0a5bd
Reviewed-on: https://go-review.googlesource.com/c/140698
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Teach samesafeexpr to handle arithmetic unary and binary ops.
It makes map lookup optimization possible in
m[k+1] = append(m[k+1], ...)
m[-k] = append(m[-k], ...)
... etc
Does not cover "+" for strings (concatenation).
Change-Id: Ibbb16ac3faf176958da344be1471b06d7cf33a6c
Reviewed-on: https://go-review.googlesource.com/135795
Run-TryBot: Iskander Sharipov <iskander.sharipov@intel.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Makes go binary smaller by 0.2%.
I noticed this in autogenerated equal methods, and there are
probably a lot of those.
Change-Id: I4e04eb3653fbceb9dd6a4eee97ceab1fa4d10b72
Reviewed-on: https://go-review.googlesource.com/135379
Reviewed-by: Ilya Tocar <ilya.tocar@intel.com>
This adds some improvements to the rules for PPC64 to eliminate
unnecessary zero or sign extends, and fix some rule for truncates
which were not always using the correct sign instruction.
This reduces of size of many functions by 1 or 2 instructions and
can improve performance in cases where the execution time depends
on small loops where at least 1 instruction was removed and where that
loop contributes a significant amount of the total execution time.
Included is a testcase for codegen to verify the sign/zero extend
instructions are omitted.
An example of the improvement (strings):
IndexAnyASCII/256:1-16 392ns ± 0% 369ns ± 0% -5.79% (p=0.000 n=1+10)
IndexAnyASCII/256:2-16 397ns ± 0% 376ns ± 0% -5.23% (p=0.000 n=1+9)
IndexAnyASCII/256:4-16 405ns ± 0% 384ns ± 0% -5.19% (p=1.714 n=1+6)
IndexAnyASCII/256:8-16 427ns ± 0% 403ns ± 0% -5.57% (p=0.000 n=1+10)
IndexAnyASCII/256:16-16 441ns ± 0% 418ns ± 1% -5.33% (p=0.000 n=1+10)
IndexAnyASCII/4096:1-16 5.62µs ± 0% 5.27µs ± 1% -6.31% (p=0.000 n=1+10)
IndexAnyASCII/4096:2-16 5.67µs ± 0% 5.29µs ± 0% -6.67% (p=0.222 n=1+8)
IndexAnyASCII/4096:4-16 5.66µs ± 0% 5.28µs ± 1% -6.66% (p=0.000 n=1+10)
IndexAnyASCII/4096:8-16 5.66µs ± 0% 5.31µs ± 1% -6.10% (p=0.000 n=1+10)
IndexAnyASCII/4096:16-16 5.70µs ± 0% 5.33µs ± 1% -6.43% (p=0.182 n=1+10)
Change-Id: I739a6132b505936d39001aada5a978ff2a5f0500
Reviewed-on: https://go-review.googlesource.com/129875
Reviewed-by: David Chase <drchase@google.com>
math.RoundToEven can be done by one arm64 instruction FRINTND, intrinsify it to improve performance.
The current pure Go implementation of the function Abs is translated into five instructions on arm64:
str, ldr, and, str, ldr. The intrinsic implementation requires only one instruction, so in terms of
performance, intrinsify it is worthwhile.
Benchmarks:
name old time/op new time/op delta
Abs-8 3.50ns ± 0% 1.50ns ± 0% -57.14% (p=0.000 n=10+10)
RoundToEven-8 9.26ns ± 0% 1.50ns ± 0% -83.80% (p=0.000 n=10+10)
Change-Id: I9456b26ab282b544dfac0154fc86f17aed96ac3d
Reviewed-on: https://go-review.googlesource.com/116535
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The CL 132915 added the wrong codegen test for math.Copysign(c, -1),
it should test that AND is not emitted. This CL fixes this error.
Change-Id: Ida1d3d54ebfc7f238abccbc1f70f914e1b5bfd91
Reviewed-on: https://go-review.googlesource.com/134815
Reviewed-by: Giovanni Bajo <rasky@develer.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Giovanni Bajo <rasky@develer.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Add some rules to match the Go code like:
y &= 63
x << y | x >> (64-y)
or
y &= 63
x >> y | x << (64-y)
as a ROR instruction. Make math/bits.RotateLeft faster on arm64.
Extends CL 132435 to arm64.
Benchmarks of math/bits.RotateLeftxxN:
name old time/op new time/op delta
RotateLeft-8 3.548750ns +- 1% 2.003750ns +- 0% -43.54% (p=0.000 n=8+8)
RotateLeft8-8 3.925000ns +- 0% 3.925000ns +- 0% ~ (p=1.000 n=8+8)
RotateLeft16-8 3.925000ns +- 0% 3.927500ns +- 0% ~ (p=0.608 n=8+8)
RotateLeft32-8 3.925000ns +- 0% 2.002500ns +- 0% -48.98% (p=0.000 n=8+8)
RotateLeft64-8 3.536250ns +- 0% 2.003750ns +- 0% -43.34% (p=0.000 n=8+8)
Change-Id: I77622cd7f39b917427e060647321f5513973232c
Reviewed-on: https://go-review.googlesource.com/122542
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This CL implements the math/bits.OnesCount{8,16,32,64} functions
as intrinsics on s390x using the 'population count' (popcnt)
instruction. This instruction was released as the 'population-count'
facility which uses the same facility bit (45) as the
'distinct-operands' facility which is a pre-requisite for Go on
s390x. We can therefore use it without a feature check.
The s390x popcnt instruction treats a 64 bit register as a vector
of 8 bytes, summing the number of ones in each byte individually.
It then writes the results to the corresponding bytes in the
output register. Therefore to implement OnesCount{16,32,64} we
need to sum the individual byte counts using some extra
instructions. To do this efficiently I've added some additional
pseudo operations to the s390x SSA backend.
Unlike other architectures the new instruction sequence is faster
for OnesCount8, so that is implemented using the intrinsic.
name old time/op new time/op delta
OnesCount 3.21ns ± 1% 1.35ns ± 0% -58.00% (p=0.000 n=20+20)
OnesCount8 0.91ns ± 1% 0.81ns ± 0% -11.43% (p=0.000 n=20+20)
OnesCount16 1.51ns ± 3% 1.21ns ± 0% -19.71% (p=0.000 n=20+17)
OnesCount32 1.91ns ± 0% 1.12ns ± 1% -41.60% (p=0.000 n=19+20)
OnesCount64 3.18ns ± 4% 1.35ns ± 0% -57.52% (p=0.000 n=20+20)
Change-Id: Id54f0bd28b6db9a887ad12c0d72fcc168ef9c4e0
Reviewed-on: https://go-review.googlesource.com/114675
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
The codegen testsuite uses regexp to parse the syntax, but it doesn't
have a way to tell line comments containing checks from line comments
containing English sentences. This means that any syntax error (that
is, non-matching regexp) is currently ignored and not reported.
There were some tests in memcombine.go that had an extraneous space
and were thus effectively disabled. It would be great if we could
report it as a syntax error, but for now we just punt and swallow the
spaces as a workaround, to avoid the same mistake again.
Fixes#25452
Change-Id: Ic7747a2278bc00adffd0c199ce40937acbbc9cf0
Reviewed-on: https://go-review.googlesource.com/113835
Run-TryBot: Giovanni Bajo <rasky@develer.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
We generate MOVBLZX for byte-sized LoadReg, so
(MOVBQZX (LoadReg (Arg))) is the same as
(LoadReg (Arg)). Remove those zero extension where possible.
Triggers several times during all.bash.
Fixes#25378
Updates #15300
Change-Id: If50656e66f217832a13ee8f49c47997f4fcc093a
Reviewed-on: https://go-review.googlesource.com/115617
Run-TryBot: Ilya Tocar <ilya.tocar@intel.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Memmove can use AVX/prefetches/other optional instructions, so
only do it for small sizes, when call overhead dominates.
Change-Id: Ice5e93deb11462217f7fb5fc350b703109bb4090
Reviewed-on: https://go-review.googlesource.com/112517
Run-TryBot: Ilya Tocar <ilya.tocar@intel.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Munday <mike.munday@ibm.com>
