The previous change sped up the pure computation form of LeadingZeros8.
This places it somewhat close to the table lookup form.
Depending on something that varies from toolchain to toolchain
(alignment, perhaps?), the slowdown from ditching the table lookup
is either 20% or 5%.
This benchmark is the best case scenario for the table lookup:
It is in the L1 cache already.
I think we're close enough that we can switch to the computational version,
and trust that the memory effects and binary size savings will be worth it.
Code:
func f8(x uint8) { z = bits.LeadingZeros8(x) }
Before:
"".f8 STEXT nosplit size=34 args=0x8 locals=0x0
0x0000 00000 (x.go:7) TEXT "".f8(SB), NOSPLIT, $0-8
0x0000 00000 (x.go:7) FUNCDATA $0, gclocals·2a5305abe05176240e61b8620e19a815(SB)
0x0000 00000 (x.go:7) FUNCDATA $1, gclocals·33cdeccccebe80329f1fdbee7f5874cb(SB)
0x0000 00000 (x.go:7) MOVBLZX "".x+8(SP), AX
0x0005 00005 (x.go:7) MOVBLZX AL, AX
0x0008 00008 (x.go:7) LEAQ math/bits.len8tab(SB), CX
0x000f 00015 (x.go:7) MOVBLZX (CX)(AX*1), AX
0x0013 00019 (x.go:7) ADDQ $-8, AX
0x0017 00023 (x.go:7) NEGQ AX
0x001a 00026 (x.go:7) MOVQ AX, "".z(SB)
0x0021 00033 (x.go:7) RET
After:
"".f8 STEXT nosplit size=30 args=0x8 locals=0x0
0x0000 00000 (x.go:7) TEXT "".f8(SB), NOSPLIT, $0-8
0x0000 00000 (x.go:7) FUNCDATA $0, gclocals·2a5305abe05176240e61b8620e19a815(SB)
0x0000 00000 (x.go:7) FUNCDATA $1, gclocals·33cdeccccebe80329f1fdbee7f5874cb(SB)
0x0000 00000 (x.go:7) MOVBLZX "".x+8(SP), AX
0x0005 00005 (x.go:7) MOVBLZX AL, AX
0x0008 00008 (x.go:7) LEAL 1(AX)(AX*1), AX
0x000c 00012 (x.go:7) BSRL AX, AX
0x000f 00015 (x.go:7) ADDQ $-8, AX
0x0013 00019 (x.go:7) NEGQ AX
0x0016 00022 (x.go:7) MOVQ AX, "".z(SB)
0x001d 00029 (x.go:7) RET
Change-Id: Icc7db50a7820fb9a3da8a816d6b6940d7f8e193e
Reviewed-on: https://go-review.googlesource.com/108942
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This was an artifact from when we had a separate ssa.Type interface to
break circular dependency between packages ssa and gc. It's no longer
needed now that package ssa directly uses package types.
Change-Id: I6a93e5d79082815f7f0eb89507381969cc6cb403
Reviewed-on: https://go-review.googlesource.com/109137
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
Currently, each architecture lowers OpConvert to an arch-specific
OpXXXconvert. This is silly because OpConvert means the same thing on
all architectures and is logically a no-op that exists only to keep
track of conversions to and from unsafe.Pointer. Furthermore, lowering
it makes it harder to recognize in other analyses, particularly
liveness analysis.
This CL eliminates the lowering of OpConvert, leaving it as the
generic op until code generation time.
The main complexity here is that we still need to register-allocate
OpConvert operations. Currently, each arch's lowered OpConvert
specifies all GP registers in its register mask. Ideally, OpConvert
wouldn't affect value homing at all, and we could just copy the home
of OpConvert's source, but this can potentially home an OpConvert in a
LocalSlot, which neither regalloc nor stackalloc expect. Rather than
try to disentangle this assumption from regalloc and stackalloc, we
continue to register-allocate OpConvert, but teach regalloc that
OpConvert can be allocated to any allocatable GP register.
For #24543.
Change-Id: I795a6aee5fd94d4444a7bafac3838a400c9f7bb6
Reviewed-on: https://go-review.googlesource.com/108496
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
Focus on "isfoo" funcs that take a *Node, and conver them to isFoo
methods instead. This makes for more idiomatic Go code, and also more
readable func names.
Found candidates with grep, and applied most changes with sed. The funcs
chosen were isgoconst, isnil, and isblank. All had the same signature,
func(*Node) bool.
While at it, camelCase the isliteral and iszero function names. Don't
move these to methods, as they are only used in the backend part of gc,
which might one day be split into a separate package.
Passes toolstash -cmp on std cmd.
Change-Id: I4df081b12d36c46c253167c8841c5a841f1c5a16
Reviewed-on: https://go-review.googlesource.com/105555
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
When a neither of a conditional block's successors follows,
the block must end with a conditional branch followed by a
an unconditional branch. If the (conditional) branch is
"unlikely", invert it and swap successors to make it
likely instead.
This doesn't matter to most benchmarks on amd64, but in one
instance on amd64 it caused a 30% improvement, and it is
otherwise harmless. The problematic loop is
for i := 0; i < w; i++ {
if pw[i] != 0 {
return true
}
}
compiled under GOEXPERIMENT=preemptibleloops
This the very worst-case benchmark for that experiment.
Also in this CL is a commoning up of heavily-repeated
boilerplate, which made it much easier to see that the
changes were applied correctly. In the future this should
allow un-exporting of SSAGenState.Branches once the
boilerplate-replacement is done everywhere.
Change-Id: I0e5ded6eeb3ab1e3e0138e12d54c7e056bd99335
Reviewed-on: https://go-review.googlesource.com/104977
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Insert appropriate race/msan calls before each memory operation during
SSA construction.
This is conceptually simple, but subtle because we need to be careful
that inserted instrumentation calls don't clobber arguments that are
currently being prepared for a user function call.
reorder1 already handles introducing temporary variables for arguments
in some cases. This CL changes it to use them for all arguments when
instrumenting.
Also, we can't SSA struct types with more than one field while
instrumenting. Otherwise, concurrent uses of disjoint fields within an
SSA-able struct can introduce false races.
This is both somewhat better and somewhat worse than the old racewalk
instrumentation pass. We're now able to easily recognize cases like
constructing non-escaping closures on the stack or accessing closure
variables don't need instrumentation calls. On the other hand,
spilling escaping parameters to the heap now results in an
instrumentation call.
Overall, this CL results in a small net reduction in the number of
instrumentation calls, but a small net increase in binary size for
instrumented executables. cmd/go ends up with 5.6% fewer calls, but a
2.4% larger binary.
Fixes#19054.
Change-Id: I70d1dd32ad6340e6fdb691e6d5a01452f58e97f3
Reviewed-on: https://go-review.googlesource.com/102817
Reviewed-by: Cherry Zhang <cherryyz@google.com>
ssa's pos parameter on the Const* funcs is unused, so remove it.
ld's alloc parameter on elfnote is always true, so remove the arguments
and simplify the code.
Finally, arm's addpltreloc never has its return parameter used, so
remove it.
Change-Id: I63387ecf6ab7b5f7c20df36be823322bb98427b8
Reviewed-on: https://go-review.googlesource.com/104456
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
In gdb, "b f" gets confused if the first instruction of "f"
is not marked as a statement in the DWARF line table.
To ensure gdb is not confused, move the first statement
marker in "f" to its first instruction.
The screwy-looking conditional for "what's the first
instruction with a statement marker" will become simpler in
the future.
Fixes#24695.
Change-Id: I2eef81676b64d1bd9bff5da03b89b9dc0c18f44f
Reviewed-on: https://go-review.googlesource.com/104955
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This commit does not change the semantics of the Call method. Its
purpose is to avoid duplication of code by making PrepareCall available
for separate use by the wasm backend.
Updates #18892
Change-Id: I04a3098f56ebf0d995791c5375dd4c03b6a202a3
Reviewed-on: https://go-review.googlesource.com/103275
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
Stores to auto tmp variables can be hoisted to places
where the line numbers make debugging look "jumpy".
Turning those instructions into ones with is_stmt = 0 in
the DWARF (accomplished by marking ssa nodes with NotStmt)
makes debugging look better while still attributing the
instructions with the correct line number.
The same is true for certain register allocator spills and
reloads.
Change-Id: I97a394eb522d4911cc40b4bf5bf76d3d7221f6c0
Reviewed-on: https://go-review.googlesource.com/98415
Run-TryBot: David Chase <drchase@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
Add IsStmt information to src.lico so that suitable lines
for breakpoints (or not) can be noted, eventually for
communication to the debugger via the linker and DWARF.
The expectation is that the front end will apply statement
boundary marks because it has best information about the
input, and the optimizer will attempt to preserve these.
The exact method for placing these marks is still TBD;
ideally stopping "at" line N in unoptimized code will occur
at a point where none of the side effects of N have occurred
and all of the inputs for line N can still be observed.
The optimizer will work with the same markings supplied
for unoptimized code.
It is a goal that non-optimizing compilation should conserve
statement marks.
The optimizer will also use the not-a-statement annotation
to indicate instructions that have a line number (for
profiling purposes) but should not be the target of
debugger step, next, or breakpoints. Because instructions
marked as statements are sometimes removed, a third value
indicating that a position (instruction) can serve as a
statement if the optimizer removes the current instruction
marked as a statement for the same line. The optimizer
should attempt to conserve statement marks, but it is not
a bug if some are lost.
Includes changes to html output for GOSSAFUNC to indicate
not-default is-a-statement with bold and not-a-statement
with strikethrough.
Change-Id: Ia22c9a682f276e2ca2a4ef7a85d4b6ebf9c62b7f
Reviewed-on: https://go-review.googlesource.com/93663
Run-TryBot: David Chase <drchase@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
The ssa.Func has Type field that is described as
function signature type.
It never gets any value and remains nil.
This leads to "<T>" signature printed representation.
Given this function declaration:
func foo(x int, f func() string) (int, error)
GOSSAFUNC printed it as below:
compiling foo
foo <T>
After this change:
compiling foo
foo func(int, func() string) (int, error)
Change-Id: Iec5eec8aac5c76ff184659e30f41b2f5fe86d329
Reviewed-on: https://go-review.googlesource.com/102375
Run-TryBot: Iskander Sharipov <iskander.sharipov@intel.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
Instead of creating a new &nodfp expression for every recover() call,
or a new nodpc variable for every function instrumented by the race
detector, this CL introduces two new uintptr-typed pseudo-variables
callerSP and callerPC. These pseudo-variables act just like calls to
the runtime's getcallersp() and getcallerpc() functions.
For consistency, change runtime.gorecover's builtin stub's parameter
type from "*int32" to "uintptr".
Passes toolstash-check, but toolstash-check -race fails because of
register allocator changes.
Change-Id: I985d644653de2dac8b7b03a28829ad04dfd4f358
Reviewed-on: https://go-review.googlesource.com/99416
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
The first word of an interface is a pointer, but for the purposes
of GC we don't need to treat it as such.
1. If it is a non-empty interface, the pointer points to an itab
which is always in persistentalloc space.
2. If it is an empty interface, the pointer points to a _type.
a. If it is a compile-time-allocated type, it points into
the read-only data section.
b. If it is a reflect-allocated type, it points into the Go heap.
Reflect is responsible for keeping a reference to
the underlying type so it won't be GCd.
If we ever have a moving GC, we need to change this for 2b (as
well as scan itabs to update their itab._type fields).
Write barriers on the first word of interfaces have already been removed.
Change-Id: I643e91d7ac4de980ac2717436eff94097c65d959
Reviewed-on: https://go-review.googlesource.com/97518
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
Not a big improvement, but does help edge cases like the SSA package.
Change-Id: I40e531110b97efd5f45955be477fd0f4faa8d545
Reviewed-on: https://go-review.googlesource.com/92396
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
Completely redesign and reimplement location list generation to be more
efficient, and hopefully not too hard to understand.
RegKills are gone. Instead of using the regalloc's liveness
calculations, redo them using the Ops' clobber information. Besides
saving a lot of Values, this avoids adding RegKills to blocks that would
be empty otherwise, which was messing up optimizations. This does mean
that it's much harder to tell whether the generation process is buggy
(there's nothing to cross-check it with), and there may be disagreements
with GC liveness. But the performance gain is significant, and it's nice
not to be messing with earlier compiler phases.
The intermediate representations are gone. Instead of producing
ssa.BlockDebugs, then dwarf.LocationLists, and then finally real
location lists, go directly from the SSA to a (mostly) real location
list. Because the SSA analysis happens before assembly, it stores
encoded block/value IDs where PCs would normally go. It would be easier
to do the SSA analysis after assembly, but I didn't want to retain the
SSA just for that.
Generation proceeds in two phases: first, it traverses the function in
CFG order, storing the state of the block at the beginning and end. End
states are used to produce the start states of the successor blocks. In
the second phase, it traverses in program text order and produces the
location lists. The processing in the second phase is redundant, but
much cheaper than storing the intermediate representation. It might be
possible to combine the two phases somewhat to take advantage of cases
where the CFG matches the block layout, but I haven't tried.
Location lists are finalized by adding a base address selection entry,
translating each encoded block/value ID to a real PC, and adding the
terminating zero entry. This probably won't work on OSX, where dsymutil
will choke on the base address selection. I tried emitting CU-relative
relocations for each address, and it was *very* bad for performance --
it uses more memory storing all the relocations than it does for the
actual location list bytes. I think I'm going to end up synthesizing the
relocations in the linker only on OSX, but TBD.
TestNexting needs updating: with more optimizations working, the
debugger doesn't stop on the continue (line 88) any more, and the test's
duplicate suppression kicks in. Also, dx and dy live a little longer
now, but they have the correct values.
Change-Id: Ie772dfe23a4e389ca573624fac4d05401ae32307
Reviewed-on: https://go-review.googlesource.com/89356
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
Now that the buffered write barrier is implemented for all
architectures, we can remove the old eager write barrier
implementation. This CL removes the implementation from the runtime,
support in the compiler for calling it, and updates some compiler
tests that relied on the old eager barrier support. It also makes sure
that all of the useful comments from the old write barrier
implementation still have a place to live.
Fixes#22460.
Updates #21640 since this fixes the layering concerns of the write
barrier (but not the other things in that issue).
Change-Id: I580f93c152e89607e0a72fe43370237ba97bae74
Reviewed-on: https://go-review.googlesource.com/92705
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Rick Hudson <rlh@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Turn off append-to-itself optimization if optimizations are turned off.
This optimization triggered a bug when doing
s = append(s, s)
where we write to the leftmost s before reading the rightmost s.
Update #17039
Change-Id: I21996532d20a75db6ec8d49db50cb157a1360b80
Reviewed-on: https://go-review.googlesource.com/81816
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The ssa backend is aggressive about placing constants and
certain other values in the Entry block. It's implausible
that the original line numbers for these constants makes
any sort of sense when it appears to a user stepping in a
debugger, and they're also not that useful in dumps since
entry-block instructions tend to be constants (i.e.,
unlikely to be the cause of a crash).
Therefore, use src.NoXPos for any values that are explicitly
inserted into a function's entry block.
Passes all tests, including ssa/debug_test.go with both
gdb and a fairly recent dlv. Hand-verified that it solves
the reported problem; constructed a test that reproduced
a problem, and fixed it.
Modified test harness to allow injection of slightly more
interesting inputs.
Fixes#22558.
Change-Id: I4476927067846bc4366da7793d2375c111694c55
Reviewed-on: https://go-review.googlesource.com/81215
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This changes the assembly language output to use the
innermost (instead of outermost) position for line
number and file.
The file is printed separately, only when it changes,
to remove redundant and space-consuming noise from the
output.
Unknown positions have line number "?"
The output format was changed slightly to make it
easier to read.
One source of gratuitous variation in debugging output was
removed.
Change-Id: I1fd9c8b0ddd82766488582fb684dce4b04f35723
Reviewed-on: https://go-review.googlesource.com/78895
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Fixes#22703
The fix was already done by Cherry for defer/go of an interface call (CL 23820).
We just need to do it everywhere.
Change-Id: I0115d22e443931fe1bcce44c93c4d0770b5fd268
Reviewed-on: https://go-review.googlesource.com/77450
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Since the slice of names is almost exactly the same as what stringer is
already generating for us.
Passes toolstash -cmp on std cmd.
Change-Id: I3f1e95efc690c0108236689e721627f00f79a461
Reviewed-on: https://go-review.googlesource.com/77190
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Dave Cheney <dave@cheney.net>
Now possible, since stringer just got the -trimprefix flag added.
While at it, simplify a few Op stringifications since we can now use %v,
and no longer have to worry about o<len(opnames).
Passes toolstash -cmp on std cmd.
Fixes#15462.
Change-Id: Icdcde0b0a5eb165d18488918175024da274f782b
Reviewed-on: https://go-review.googlesource.com/76790
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Dave Cheney <dave@cheney.net>
The linker will refuse to work on objects larger than
2e9 bytes (see issue #9862 for why).
With this change, the compiler gives a useful error
message explaining this, instead of leaving it to the
linker to give a cryptic message later.
Fixes#1700.
Change-Id: I3933ce08ef846721ece7405bdba81dff644cb004
Reviewed-on: https://go-review.googlesource.com/74330
Reviewed-by: Robert Griesemer <gri@golang.org>
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Plain blocks that contain only uninteresting instructions
(that do not have reliable Pos information themselves)
need to have their Pos left unset so that they can
inherit it from their successors. The "uninteresting"
test was not properly applied and not properly defined.
OpFwdRef does not appear in the ssa.html debugging output,
but at the time of the test these instructions did appear,
and it needs to be part of the test.
Fixes#22365.
Change-Id: I99e5b271acd8f6bcfe0f72395f905c7744ea9a02
Reviewed-on: https://go-review.googlesource.com/74252
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Currently, assigning a []T where T is a go:notinheap type generates an
unnecessary write barrier for storing the slice pointer.
This fixes this by teaching HasHeapPointer that this type does not
have a heap pointer, and tweaking the lowering of slice assignments so
the pointer store retains the correct type rather than simply lowering
it to a *uint8 store.
Change-Id: I8bf7c66e64a7fefdd14f2bd0de8a5a3596340bab
Reviewed-on: https://go-review.googlesource.com/76027
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
A statement like
foo = bar + qux
might compile to
AX := AX + BX
resulting in a regkill for AX before this instruction.
The buggy behavior is to kill AX "at" this instruction,
before it has executed. (Code generation of no-instruction
values like RegKills applies their effects at the
next actual instruction emitted).
However, bar is still associated with AX until after the
instruction executes, so the effect of the regkill must
occur at the boundary between this instruction and the
next. Similarly, the new value bound to AX is not visible
until this instruction executes (and in the case of values
that require multiple instructions in code generation, until
all of them have executed).
The ranges are adjusted so that a value's start occurs
at the next following instruction after its evaluation,
and the end occurs after (execution of) the first
instruction following the end of the lifetime as a value.
(Notice the asymmetry; the entire value must be finished
before it is visible, but execution of a single instruction
invalidates. However, the value *is* visible before that
next instruction executes).
The test was adjusted to make it insensitive to the result
numbering for variables printed by gdb, since that is not
relevant to the test and makes the differences introduced
by small changes larger than necessary/useful.
The test was also improved to present variable probes
more intuitively, and also to allow explicit indication
of "this variable was optimized out"
Change-Id: I39453eead8399e6bb05ebd957289b112d1100c0e
Reviewed-on: https://go-review.googlesource.com/74090
Run-TryBot: David Chase <drchase@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
We already do this for floor/ceil, but RoundToEven was added later.
Intrinsify it also.
name old time/op new time/op delta
RoundToEven-8 3.00ns ± 1% 0.68ns ± 2% -77.34% (p=0.000 n=10+10)
Change-Id: Ib158cbceb436c6725b2d9353a526c5c4be19bcad
Reviewed-on: https://go-review.googlesource.com/74852
Run-TryBot: Ilya Tocar <ilya.tocar@intel.com>
Reviewed-by: Keith Randall <khr@golang.org>
Handle make(map[any]any) and make(map[any]any, hint) where
hint <= BUCKETSIZE special to allow for faster map initialization
and to improve binary size by using runtime calls with fewer arguments.
Given hint is smaller or equal to BUCKETSIZE in which case
overLoadFactor(hint, 0) is false and no buckets would be allocated by makemap:
* If hmap needs to be allocated on the stack then only hmap's hash0
field needs to be initialized and no call to makemap is needed.
* If hmap needs to be allocated on the heap then a new special
makehmap function will allocate hmap and intialize hmap's
hash0 field.
Reduces size of the godoc by ~36kb.
AMD64
name old time/op new time/op delta
NewEmptyMap 16.6ns ± 2% 5.5ns ± 2% -66.72% (p=0.000 n=10+10)
NewSmallMap 64.8ns ± 1% 56.5ns ± 1% -12.75% (p=0.000 n=9+10)
Updates #6853
Change-Id: I624e90da6775afaa061178e95db8aca674f44e9b
Reviewed-on: https://go-review.googlesource.com/61190
Run-TryBot: Martin Möhrmann <moehrmann@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
The new RoundToEven function can be implemented as a single FIDBR
instruction on s390x.
name old time/op new time/op delta
RoundToEven 5.32ns ± 1% 0.86ns ± 1% -83.86% (p=0.000 n=10+10)
Change-Id: Iaf597e57a0d1085961701e3c75ff4f6f6dcebb5f
Reviewed-on: https://go-review.googlesource.com/74350
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This adds support for math Abs, Copysign to be instrinsics on ppc64x.
New instruction FCPSGN is added to generate fcpsgn. Some new
rules are added to improve the int<->float conversions that are
generated mainly due to the Float64bits and Float64frombits in
the math package. PPC64.rules is also modified as suggested
in the review for CL 63290.
Improvements:
benchmark old ns/op new ns/op delta
BenchmarkAbs-16 1.12 0.69 -38.39%
BenchmarkCopysign-16 1.30 0.93 -28.46%
BenchmarkNextafter32-16 9.34 8.05 -13.81%
BenchmarkFrexp-16 8.81 7.60 -13.73%
Others that used Copysign also saw smaller improvements.
I attempted to make this work using rules since that
seems to be preferred, but due to the use of Float64bits and
Float64frombits in these functions, several rules had to be added and
even then not all cases were matched. Using rules became too
complicated and seemed too fragile for these.
Updates #21390
Change-Id: Ia265da9a18355e08000818a4fba1a40e9e031995
Reviewed-on: https://go-review.googlesource.com/67130
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
Reviewed-by: Keith Randall <khr@golang.org>
Most write barrier calls are inserted by SSA, but copy and append are
lowered to runtime.typedslicecopy during walk. Fix these to set
Func.WBPos and emit the "write barrier" warning, as done for the write
barriers inserted by SSA. As part of this, we refactor setting WBPos
and emitting this warning into the frontend so it can be shared by
both walk and SSA.
Change-Id: I5fe9997d9bdb55e03e01dd58aee28908c35f606b
Reviewed-on: https://go-review.googlesource.com/73411
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
