Get rid of `std.math.F80Repr`. Instead of trying to match the memory
layout of f80, we treat it as a value, same as the other floating point
types. The functions `make_f80` and `break_f80` are introduced to
compose an f80 value out of its parts, and the inverse operation.
stage2 LLVM backend: fix pointer to zero length array tripping LLVM
assertion. It now checks for when the element type is a zero-bit type
and lowers such thing the same way that pointers to other zero-bit types
are lowered.
Both stage1 and stage2 LLVM backends are adjusted so that f80 is lowered
as x86_fp80 on x86_64 and i386 architectures, and identical to a u80 on
others. LLVM constants are lowered in a less hacky way now that #10860
is fixed, by using the expression `(exp << 64) | fraction` using llvm
constants.
Sema is improved to handle c_longdouble by recursively handling it
correctly for whatever the float bit width is. In both stage1 and
stage2.
* F80Repr extern struct needs no explicit padding; let's match the
target padding.
* stage2: fix lowering of f80 constants.
* stage1: decide ABI size and alignment of f80 based on alignment of
u64. x86 has alignof u64 equal to 4 but arm has it as 8.
* stage2: fix Value.floatReadFromMemory to use F80Repr
In the previous commit I got mixed up and cut-pasted instead of
copy-pasting. In this commit I made c_stage1.zig additionally included
for stage1 and everything else included for both. So moving forward we
move stuff over from c_stage1.zig to c.zig instead of copying.
This updates the test runner for stage2 to emit to stdout with the passed, skipped and failed tests
similar to the LLVM backend.
Another change to this is the start function, as it's now more in line with stage1's.
The stage2 test infrastructure for wasm/wasi has been updated to reflect this as well.
- approach by Hacker's Delight with wrapping subtraction
- performance expected to be similar to addo
- tests with all relevant combinations of min,max with -1,0,+1 and all
combinations of sequences +-1,2,4..,max
For large ranges, this is faster than having the caller call setValue() for
each index in the range. Masks wholly covered by the range can be set to
the new mask value in one go, and the two masks at either end that are
partially covered can each set the covered range of bits in one go.
- approach by Hacker's Delight with wrapping addition
- ca. 1.10x perf over the standard approach on my laptop
- tests with all combinations of min,max with -1,0,+1 and combinations of
sequences +-1,2,4..,max
In accordance with the requesting issue (#10750):
- `zig test` skips any tests that it cannot spawn, returning success
- `zig run` and `zig build` exit with failure, reporting the command the cannot be run
- `zig clang`, `zig ar`, etc. already punt directly to the appropriate clang/lld main(), even before this change
- Native `libc` Detection is not supported
Additionally, `exec()` and related Builder functions error at run-time, reporting the command that cannot be run
We're going to remove the first parameter from this function in the
future. Stage2 already ignores the first parameter. So we put an `@as`
in here to make it work for both.
This adds a new path which avoids using compiler_rt generated div
udivmod instructions in the case that a divisor is less than half the
max usize value. Two half-limb divisions are performed instead which
ensures that non-emulated division instructions are actually used. This
does not improve the udivmod code which should still be reviewed
independently of this issue.
Notably this improves the performance of the toString implementation of
non-power-of-two bases considerably.
Division performance is improved ~1000% based on some coarse testing.
The following test code is used to provide a rough comparison between
the old vs. new method.
```
const std = @import("std");
const Managed = std.math.big.int.Managed;
const allocator = std.heap.c_allocator;
fn fib(a: *Managed, n: usize) !void {
var b = try Managed.initSet(allocator, 1);
defer b.deinit();
var c = try Managed.init(allocator);
defer c.deinit();
var i: usize = 0;
while (i < n) : (i += 1) {
try c.add(a.toConst(), b.toConst());
a.swap(&b);
b.swap(&c);
}
}
pub fn main() !void {
var a = try Managed.initSet(allocator, 0);
defer a.deinit();
try fib(&a, 1_000_000);
// Note: Next two lines (and printed digit count) omitted on no-print version.
const as = try a.toString(allocator, 10, .lower);
defer allocator.free(as);
std.debug.print("fib: digit count: {}, limb count: {}\n", .{ as.len, a.limbs.len });
}
```
```
==> time.no-print <==
limb count: 10849
________________________________________________________
Executed in 10.60 secs fish external
usr time 10.44 secs 0.00 millis 10.44 secs
sys time 0.02 secs 1.12 millis 0.02 secs
==> time.old <==
fib: digit count: 208988, limb count: 10849
________________________________________________________
Executed in 22.78 secs fish external
usr time 22.43 secs 1.01 millis 22.43 secs
sys time 0.03 secs 0.13 millis 0.03 secs
==> time.optimized <==
fib: digit count: 208988, limb count: 10849
________________________________________________________
Executed in 11.59 secs fish external
usr time 11.56 secs 1.03 millis 11.56 secs
sys time 0.03 secs 0.12 millis 0.03 secs
```
Perf data for non-optimized and optimized, verifying no udivmod is
generated by the new code.
```
$ perf report -i perf.data.old --stdio
- Total Lost Samples: 0
-
- Samples: 90K of event 'cycles:u'
- Event count (approx.): 71603695208
-
- Overhead Command Shared Object Symbol
- ........ ....... ................ ...........................................
-
52.97% t t [.] compiler_rt.udivmod.udivmod
45.97% t t [.] std.math.big.int.Mutable.addCarry
0.83% t t [.] main
0.08% t libc-2.33.so [.] __memmove_avx_unaligned_erms
0.08% t t [.] __udivti3
0.03% t [unknown] [k] 0xffffffff9a0010a7
0.02% t t [.] std.math.big.int.Managed.ensureCapacity
0.01% t libc-2.33.so [.] _int_malloc
0.00% t libc-2.33.so [.] __malloc_usable_size
0.00% t libc-2.33.so [.] _int_free
0.00% t t [.] 0x0000000000004a80
0.00% t t [.] std.heap.CAllocator.resize
0.00% t libc-2.33.so [.] _mid_memalign
0.00% t libc-2.33.so [.] sysmalloc
0.00% t libc-2.33.so [.] __posix_memalign
0.00% t t [.] std.heap.CAllocator.alloc
0.00% t ld-2.33.so [.] do_lookup_x
$ perf report -i perf.data.optimized --stdio
- Total Lost Samples: 0
-
- Samples: 46K of event 'cycles:u'
- Event count (approx.): 36790112336
-
- Overhead Command Shared Object Symbol
- ........ ....... ................ ...........................................
-
79.98% t t [.] std.math.big.int.Mutable.addCarry
15.14% t t [.] main
4.58% t t [.] std.math.big.int.Managed.ensureCapacity
0.21% t libc-2.33.so [.] __memmove_avx_unaligned_erms
0.05% t [unknown] [k] 0xffffffff9a0010a7
0.02% t libc-2.33.so [.] _int_malloc
0.01% t t [.] std.heap.CAllocator.alloc
0.01% t libc-2.33.so [.] __malloc_usable_size
0.00% t libc-2.33.so [.] systrim.constprop.0
0.00% t libc-2.33.so [.] _mid_memalign
0.00% t t [.] 0x0000000000000c7d
0.00% t libc-2.33.so [.] malloc
0.00% t ld-2.33.so [.] check_match
```
Closes#10630.
This was found on a user's machine when calling "git" as a child process from msys. Instead of getting BROKEN_PIPE on GetOverlappedREsult, it would occur on ReadFile which would then cause the function to hang because the async operation was never started.
These options were removed in 5e63baae8 (CLI: remove --verbose-ast and
--verbose-tokenize, 2021-06-09) but some remainders were left in.
Signed-off-by: Johannes Löthberg <johannes@kyriasis.com>
- use usize to decide if register size is big enough to store
multiplication result or if division is necessary
- multiplication routine with check of integer bounds
- wrapping multipliation and division routine from Hacker's Delight
This commit fixes two related things:
1. If the loop goes all the way through the slice without a match, on
the last iteration `mid == symbols.len - 1` which causes
`&symbols[mid + 1]` to be out of bounds. End one step before that
instead.
2. If the address we're looking for is greater than the address of the
last symbol in the slice, we now match it to that symbol. Previously,
we would miss this case since we only matched if the address was _in
between_ the address of two symbols.
which is the index of the key that already exists in the hash map.
This enables the use case of using `AutoArrayHashMap(void, void)` which
may seem surprising at first, but is actually pretty handy!
This commit includes a proof-of-concept of how I want to use it, with a
new InternArena abstraction for stage2 that provides a compact way to
store values (and types) in an "internment arena", thus making types
stored exactly once (per arena), representable with a single u32 as a
reference to a type within an InternArena, and comparable with a
simple u32 integer comparison. If both types are in the same
InternArena, you can check if they are equal by seeing if their index is
the same.
What's neat about `AutoArrayHashMap(void, void)` is that it allows us to
look up the indexes by key, *without actually storing the keys*.
Instead, keys are treated as ephemeral values that are constructed as
needed.
As a result, we have an extremely efficient encoding of types and
values, represented only by three arrays, which has no pointers, and can
therefore be serialized and deserialized by a single writev/readv call.
The `map` field is denormalized data and can be computed from the other
two fields.
This is in contrast to our current Type/Value system which makes
extensive use of pointers.
The test at the bottom of InternArena.zig passes in this commit.
I hit the "quotes in an RSP file" issue when trying to compile gRPC using
"zig cc". As a fun exercise, I decided to see if I could fix it myself.
I'm fully open to this code being flat-out rejected. Or I can take feedback
to fix it up.
This modifies (and renames) _ArgIteratorWindows_ in process.zig such that
it works with arbitrary strings (or the contents of an RSP file).
In main.zig, this new _ArgIteratorGeneral_ is used to address the "TODO"
listed in _ClangArgIterator_.
This change closes#4833.
**Pros:**
- It has the nice attribute of handling "RSP file" arguments in the same way it
handles "cmd_line" arguments.
- High Performance, minimal allocations
- Fixed bug in previous _ArgIteratorWindows_, where final trailing backslashes
in a command line were entirely dropped
- Added a test case for the above bug
- Harmonized the _ArgIteratorXxxx._initWithAllocator()_ and _next()_ interface
across Windows/Posix/Wasi (Moved Windows errors to _initWithAllocator()_
rather than _next()_)
- Likely perf benefit on Windows by doing _utf16leToUtf8AllocZ()_ only once
for the entire cmd_line
**Cons:**
- Breaking Change in std library on Windows: Call
_ArgIterator.initWithAllocator()_ instead of _ArgIterator.init()_
- PhaseMage is new with contributions to Zig, might need a lot of hand-holding
- PhaseMage is a Windows person, non-Windows stuff will need to be double-checked
**Testing Done:**
- Wrote a few new test cases in process.zig
- zig.exe build test -Dskip-release (no new failures seen)
- zig cc now builds gRPC without error
