v8/test/mjsunit/compiler/bigint-int64-lowered.js
Nico Hartmann 99df710d4c [turbofan] Push BigInt truncation over addition and heap constants
This change implements lowering of speculative BigInt addition as well as
BigInt heap constants to corresponding int64 versions, if they are used in
a context where the result is truncated to the least significant 64 bits
(e.g. using asUintN). The JSHeapBroker is extended to provide access to the
BigInt's least significant digit during concurrent compilation. The BigInt
context (required to introduce correct conversions) is recognized in the
RepresentationChanger by either the output type propagated downward or the
TypeCheckKind propagated upward. This is necessary, because the TypeCheckKind
may only be set by nodes that may potentially deopt (and sit in the effect
chain). This is the case for SpeculativeBigIntAdd, but not for BigIntAsUintN.

This CL contains a simple fix to prevent int64-lowered BigInts to flow into
state values as the deoptimizer cannot handle them yet. A more sophisticated
solution to allow the deoptimizer to materialize truncated BigInts will be
added in a following CL.

Bug: v8:9407
Change-Id: I96a293e9077962f53e5f199857644f004e3ae56e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1684183
Commit-Queue: Nico Hartmann <nicohartmann@google.com>
Reviewed-by: Georg Neis <neis@chromium.org>
Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
Reviewed-by: Maya Lekova <mslekova@chromium.org>
Cr-Commit-Position: refs/heads/master@{#62665}
2019-07-12 09:05:29 +00:00

83 lines
2.6 KiB
JavaScript

// Copyright 2019 the V8 project authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
// Flags: --allow-natives-syntax --opt
function TestAsUintN() {
assertEquals(0n, BigInt.asUintN(64, 0n));
assertEquals(0n, BigInt.asUintN(8, 0n));
assertEquals(0n, BigInt.asUintN(1, 0n));
assertEquals(0n, BigInt.asUintN(0, 0n));
assertEquals(0n, BigInt.asUintN(100, 0n));
assertEquals(123n, BigInt.asUintN(64, 123n));
assertEquals(123n, BigInt.asUintN(32, 123n));
assertEquals(123n, BigInt.asUintN(8, 123n));
assertEquals(59n, BigInt.asUintN(6, 123n));
assertEquals(27n, BigInt.asUintN(5, 123n));
assertEquals(11n, BigInt.asUintN(4, 123n));
assertEquals(1n, BigInt.asUintN(1, 123n));
assertEquals(0n, BigInt.asUintN(0, 123n));
assertEquals(123n, BigInt.asUintN(72, 123n));
assertEquals(BigInt("0xFFFFFFFFFFFFFF85"), BigInt.asUintN(64, -123n));
assertEquals(BigInt("0xFFFFFF85"), BigInt.asUintN(32, -123n));
assertEquals(BigInt("0x85"), BigInt.asUintN(8, -123n));
assertEquals(5n, BigInt.asUintN(6, -123n));
assertEquals(5n, BigInt.asUintN(5, -123n));
assertEquals(5n, BigInt.asUintN(4, -123n));
assertEquals(1n, BigInt.asUintN(1, -123n));
assertEquals(0n, BigInt.asUintN(0, -123n));
assertEquals(BigInt("0xFFFFFFFFFFFFFFFF85"), BigInt.asUintN(72, -123n));
}
function TestInt64LoweredOperations() {
assertEquals(0n, BigInt.asUintN(64, -0n));
assertEquals(0n, BigInt.asUintN(64, 15n + -15n));
assertEquals(0n, BigInt.asUintN(64, 0n + 0n));
assertEquals(14n, BigInt.asUintN(32, 8n + 6n));
assertEquals(813n, BigInt.asUintN(10, 1013n + -200n));
assertEquals(15n, BigInt.asUintN(4, -319n + 302n));
for (let i = 0; i < 2; ++i) {
let x = 32n; // x = 32n
if (i === 1) {
x = BigInt.asUintN(64, x + 3n); // x = 35n
const y = x + -8n + x; // x = 35n, y = 62n
x = BigInt.asUintN(6, y + x); // x = 33n, y = 62n
x = -9n + y + -x; // x = 20n
x = BigInt.asUintN(10000 * i, x); // x = 20n
} else {
x = x + 400n; // x = 432n
x = -144n + BigInt.asUintN(8, 500n) + x; // x = 532n
}
assertEquals(20n, BigInt.asUintN(8, x));
}
let x = 7n;
for (let i = 0; i < 10; ++i) {
x = x + 5n;
}
assertEquals(57n, BigInt.asUintN(8, x));
let y = 7n;
for(let i = 0; i < 10; ++i) {
y = BigInt.asUintN(4, y + 16n);
}
assertEquals(7n, y);
}
function OptimizeAndTest(fn) {
%PrepareFunctionForOptimization(fn);
fn();
fn();
%OptimizeFunctionOnNextCall(fn);
fn();
assertOptimized(fn);
fn();
}
OptimizeAndTest(TestAsUintN);
OptimizeAndTest(TestInt64LoweredOperations);