Add range attribute to scalar function results and arguments #128371

andjo403 · 2024-07-29T21:49:22Z

as LLVM 19 adds the range attribute this starts to use it for better optimization.
hade been interesting to see a perf run with the #127513

closes #50156
cc #49572 shall be fixed but not possible to see as there is asserts that already trigger the optimization.

rustbot · 2024-07-29T21:49:30Z

r? @compiler-errors

rustbot has assigned @compiler-errors.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

tgross35 · 2024-07-29T22:14:37Z

tests/codegen/range-attribute.rs

+
+use std::num::NonZero;
+
+// CHECK: noundef range(i64 1, 0) i64 @nonzero_int(i64 noundef range(i64 1, 0) %x)


I think you can do

Suggested change

// CHECK: noundef range(i64 1, 0) i64 @nonzero_int(i64 noundef range(i64 1, 0) %x)

// CHECK-LABEL: @nonzero_int

// CHECK-SAME: noundef range(i64 1, 0) i64 @nonzero_int(i64 noundef range(i64 1, 0) %x)

so you get the better section splitting from CHECK-LABEL if the line doesn't match for some reason.

CHECK-SAME will start matching after the CHECK-LABEL so it is not possible to verify that the result get a range applied in that case unless it is moved to the CHECK-LABEL like
CHECK-LABEL: noundef range(i64 1, 0) i64 @nonzero_int
and then it feels like a CHECK is better

tgross35 · 2024-07-29T22:17:39Z

compiler/rustc_codegen_llvm/src/abi.rs

+        let apply_range_attr = |idx: AttributePlace, abi: rustc_target::abi::Abi| {
+            if cx.sess().opts.optimize != config::OptLevel::No
+                && llvm_util::get_version() >= (19, 0, 0)
+            {
+                if let abi::Abi::Scalar(scalar) = abi {
+                    // If the value is a boolean, the range is 0..2 and that ultimately
+                    // become 0..0 when the type becomes i1, which would be rejected
+                    // by the LLVM verifier.
+                    if let Int(..) = scalar.primitive() {
+                        if !scalar.is_bool() && !scalar.is_always_valid(cx) {


Don't quote me on this but I think it is okay to use let chains in the compiler to avoid all this nesting

Suggested change

let apply_range_attr = |idx: AttributePlace, abi: rustc_target::abi::Abi| {

if cx.sess().opts.optimize != config::OptLevel::No

&& llvm_util::get_version() >= (19, 0, 0)

{

if let abi::Abi::Scalar(scalar) = abi {

// If the value is a boolean, the range is 0..2 and that ultimately

// become 0..0 when the type becomes i1, which would be rejected

// by the LLVM verifier.

if let Int(..) = scalar.primitive() {

if !scalar.is_bool() && !scalar.is_always_valid(cx) {

let apply_range_attr = |idx: AttributePlace, abi: rustc_target::abi::Abi| {

if cx.sess().opts.optimize != config::OptLevel::No

&& llvm_util::get_version() >= (19, 0, 0)

&& let abi::Abi::Scalar(scalar) = abi

// If the value is a boolean, the range is 0..2 and that ultimately

// become 0..0 when the type becomes i1, which would be rejected

// by the LLVM verifier.

&& matches!(scalar.primitive(), Int(..))

&& !scalar.is_bool() && !scalar.is_always_valid(cx) {

One day that will stable :(

nice did not think it was possible to do that.

tgross35 · 2024-07-29T22:24:53Z

Marking as blocked on #127513, which hopefully won't be too far away.

Also, thanks for pushing this forward on both the Rust and LLVM side. The results will be awesome :)

erikdesjardins

It's great to see this!

Ideally you'll be able to revert the assumes added in #103584, now that we can represent this range info properly. (in another PR, perhaps)

erikdesjardins · 2024-07-30T22:35:15Z

tests/codegen/range-attribute.rs

+// Checks that range metadata gets emitted on functions result and arguments
+// with scalar value.


Some more tests that would be nice:

slice length (I don't think that will get range attributes yet, but when it does, it'll resolve Add unreachable hints for length getters? #106737. Having a test is still useful whether it gets the attribute or not, just to show the current state.)

discriminant of an enum with a payload (one that gets scalarpair layout) e.g. enum Foo { A(u64), B(u64), C(u64) }.

Thanks to rust-lang/reference#1499 we're allowed to do this for slices, and per #122926 (comment) it has a decent chance to save a couple megs off the compiler size. So excited to get to do it with range attributes, since doing it with assumes was perf-horrible.

think that I have added the test that you wanted. I'm letting someone more familiar create a PR for adding the range on slices.

nikic · 2024-07-31T15:48:24Z

@bors try @rust-timer queue

bors · 2024-07-31T15:49:35Z

⌛ Trying commit 5ff174c with merge b8ab1d7...

Add range attribute to scalar function results and arguments as LLVM 19 adds the range attribute this starts to use it for better optimization. hade been interesting to see a perf run with the rust-lang#127513 closes rust-lang#50156 cc rust-lang#49572 shall be fixed but not possible to see as there is asserts that already trigger the optimization. r? `@nikic`

bors · 2024-07-31T17:44:49Z

☀️ Try build successful - checks-actions
Build commit: b8ab1d7 (b8ab1d77f3f3c9e0b913842d10cc6b21af0c746a)

rust-timer · 2024-08-01T17:53:57Z

Finished benchmarking commit (b8ab1d7): comparison URL.

Overall result: ✅ improvements - no action needed

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR may lead to changes in compiler perf.

@bors rollup=never
@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.1%	[-1.1%, -1.1%]	1
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

This benchmark run did not return any relevant results for this metric.

Cycles

Results (primary 3.4%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	3.4%	[3.4%, 3.4%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	3.4%	[3.4%, 3.4%]	1

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 755.634s -> 754.893s (-0.10%)
Artifact size: 336.68 MiB -> 336.66 MiB (-0.01%)

andjo403 · 2024-08-01T18:27:48Z

This is not blocked anymore now after the rebase and the perf was ok.

andjo403 · 2024-08-01T18:37:50Z

tests/codegen/cast-optimized.rs

@@ -20,8 +20,6 @@ pub fn u32_index(c: u32) -> [bool; 22] {
 // CHECK-LABEL: @char_as_u32_index
 #[no_mangle]
 pub fn char_as_u32_index(c: char) -> [bool; 22] {
-    // CHECK: %[[B:.+]] = icmp ult i32 %c, 1114112


this is optimized away when the range is added to the input. I only removed it as it do not seem important for the test and more a nice to have check. otherwise I do not know how to update this test.

I believe the CHECK-NOT: call core::panicking::panic below, ensuring that the bounds check gets optimized out, is the important part of this test. Since that still works, I think the spirit of this test is upheld.

You can/should remove the assume (added in https://github.com/rust-lang/rust/pull/124905/files#diff-db08b7c8b00530c7183cf2e42f25dc93b02da93fb40edadbd009eee6ad60e3a1R298-R302) now that it gets optimized out anyways. (Probably need to add min-llvm-version to the test in that case, which is fine.)

do that assume not add value if the cast is not on an input/result of a function call ?

You can leave it there for now, I'll take a look at it in a separate PR.

Technically yes, it could be beneficial. Although, after this PR, we'll have range metadata on loads, arguments, and returns, which covers the vast majority of cases where new values are "introduced" into a function. I think it's unlikely that handing the remaining rare cases would be worthwhile, given that adding assumes costs a bit of compile time and can block other optimizations.

andjo403 · 2024-08-01T19:39:19Z

@rustbot ready

erikdesjardins · 2024-08-01T22:27:49Z

tests/codegen/range-attribute.rs

+
+// CHECK: noundef range(i8 0, 4) i8 @enum1_value(i8 noundef range(i8 0, 4) %x)
+#[no_mangle]
+pub fn enum1_value(x: Enum0) -> Enum0 {


Suggested change

pub fn enum1_value(x: Enum0) -> Enum0 {

pub fn enum1_value(x: Enum1) -> Enum1 {

erikdesjardins · 2024-08-02T03:12:07Z

tests/codegen/range-attribute.rs

+// CHECK: define { i64, i64 } @enum1_value(i64 noundef %x.0, i64 noundef %x.1)
+#[no_mangle]
+pub fn enum1_value(x: Enum1) -> Enum1 {
+    x
+}


Hmm, this should be getting a range attribute, we definitely set the valid range for discriminants. I suspect...

erikdesjardins · 2024-08-02T03:17:06Z

compiler/rustc_codegen_llvm/src/abi.rs

+        let apply_range_attr = |idx: AttributePlace, abi: rustc_target::abi::Abi| {
+            if cx.sess().opts.optimize != config::OptLevel::No
+                && llvm_util::get_version() >= (19, 0, 0)
+                && let abi::Abi::Scalar(scalar) = abi
+                && matches!(scalar.primitive(), Int(..))
+                // If the value is a boolean, the range is 0..2 and that ultimately
+                // become 0..0 when the type becomes i1, which would be rejected
+                // by the LLVM verifier.
+                && !scalar.is_bool()
+                // LLVM also rejects full range.
+                && !scalar.is_always_valid(cx)
+            {
+                attributes::apply_to_llfn(
+                    llfn,
+                    idx,
+                    &[llvm::CreateRangeAttr(cx.llcx, scalar.size(cx), scalar.valid_range(cx))],
+                );
+            }
+        };


...that this isn't called in enough places. Instead of building the attribute here, I would suggest:

add a valid_range field to struct ArgAttributes

set valid_range in adjust_for_rust_scalar when applicable

call llvm::CreateRangeAttr in rustc_codegen_llvm::abi::get_attrs, based on the value of valid_range

Then the existing code should handle the rest.

This solution seems to use a lot more memory if I add a valid_range_and_num_bits: Option<(WrappingRange, u32)> to ArgAttributes the ArgAbi size goes from 56 to 192 and FnAbi from 80 to 224 and the WrappingRange will not be interned but the layout is so even the same types will result in same ranges manytimes.

updated scalarPair also to avoid undoing the work from #100999 but there maybe is some good reason to do like you suggested like other backends want to use the data also.

Is the size of ArgAttributes / FnAbi sensitive? I'd have hoped they only exist temporarily while lowering a call/declaration, but maybe that's not the case.

ended up at this assert and did not know of a good reason to increase it.
https://github.com/andjo403/rust/blob/3cbfa8212a48f8eecce9c5da0615da6c4c883ce9/compiler/rustc_target/src/abi/call/mod.rs#L970-L980

nikic · 2024-08-07T13:20:21Z

@bors r+

andjo403 · 2024-08-11T21:32:59Z

managed to run the DEPLOY=1 ./src/ci/docker/run.sh test-various with the latest commit but
DEPLOY=1 ./src/ci/docker/run.sh i686-gnu fails both on master and this commit for me so seems unrelated.

nikic · 2024-08-12T07:23:44Z

@bors r+

bors · 2024-08-12T07:23:47Z

📌 Commit cfadfab has been approved by nikic

It is now in the queue for this repository.

bors · 2024-08-12T08:11:11Z

⌛ Testing commit cfadfab with merge 9392a60...

Add range attribute to scalar function results and arguments as LLVM 19 adds the range attribute this starts to use it for better optimization. hade been interesting to see a perf run with the rust-lang#127513 closes rust-lang#50156 cc rust-lang#49572 shall be fixed but not possible to see as there is asserts that already trigger the optimization.

rust-log-analyzer · 2024-08-12T09:08:44Z

The job dist-x86_64-msvc failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

[2024-08-12T08:57:09Z DEBUG collector::compile::benchmark] Benchmark iteration 1/1
[2024-08-12T08:57:09Z INFO  collector::compile::execute] run_rustc with incremental=false, profile=Check, scenario=Some(Full), patch=None, backend=Llvm, phase=benchmark
[2024-08-12T08:57:09Z DEBUG collector::compile::execute] "\\\\?\\C:\\a\\rust\\rust\\build\\x86_64-pc-windows-msvc\\stage0\\bin\\cargo.exe" "rustc" "--manifest-path" "Cargo.toml" "-p" "path+file:///C:/a/_temp/msys64/tmp/.tmpXRHLmm#ctfe-stress-5@0.1.0" "--profile" "check" "--" "--wrap-rustc-with" "Eprintln"
[2024-08-12T08:57:14Z INFO  collector::compile::execute] run_rustc with incremental=true, profile=Check, scenario=Some(IncrFull), patch=None, backend=Llvm, phase=benchmark
[2024-08-12T08:57:14Z DEBUG collector::compile::execute] "\\\\?\\C:\\a\\rust\\rust\\build\\x86_64-pc-windows-msvc\\stage0\\bin\\cargo.exe" "rustc" "--manifest-path" "Cargo.toml" "-p" "path+file:///C:/a/_temp/msys64/tmp/.tmpXRHLmm#ctfe-stress-5@0.1.0" "--profile" "check" "--" "--wrap-rustc-with" "Eprintln" "-C" "incremental=C:\\a\\_temp\\msys64\\tmp\\.tmpXRHLmm\\incremental-state"
[2024-08-12T08:57:20Z INFO  collector::compile::execute] run_rustc with incremental=true, profile=Check, scenario=Some(IncrUnchanged), patch=None, backend=Llvm, phase=benchmark
[2024-08-12T08:57:20Z DEBUG collector::compile::execute] "\\\\?\\C:\\a\\rust\\rust\\build\\x86_64-pc-windows-msvc\\stage0\\bin\\cargo.exe" "rustc" "--manifest-path" "Cargo.toml" "-p" "path+file:///C:/a/_temp/msys64/tmp/.tmpXRHLmm#ctfe-stress-5@0.1.0" "--profile" "check" "--" "--wrap-rustc-with" "Eprintln" "-C" "incremental=C:\\a\\_temp\\msys64\\tmp\\.tmpXRHLmm\\incremental-state"
[2024-08-12T08:57:20Z DEBUG collector::compile::benchmark] Benchmark iteration 1/1
[2024-08-12T08:57:20Z INFO  collector::compile::execute] run_rustc with incremental=false, profile=Debug, scenario=Some(Full), patch=None, backend=Llvm, phase=benchmark
[2024-08-12T08:57:20Z DEBUG collector::compile::execute] "\\\\?\\C:\\a\\rust\\rust\\build\\x86_64-pc-windows-msvc\\stage0\\bin\\cargo.exe" "rustc" "--manifest-path" "Cargo.toml" "-p" "path+file:///C:/a/_temp/msys64/tmp/.tmpweH6um#ctfe-stress-5@0.1.0" "--" "--wrap-rustc-with" "Eprintln"
[2024-08-12T08:57:26Z INFO  collector::compile::execute] run_rustc with incremental=true, profile=Debug, scenario=Some(IncrFull), patch=None, backend=Llvm, phase=benchmark
---
   Compiling rustc_driver v0.0.0 (C:\a\rust\rust\compiler\rustc_driver)
[RUSTC-TIMING] rustc_driver test:false 5.120
error: linking with `link.exe` failed: exit code: 1104
  |
  = note: "C:\\Program Files\\Microsoft Visual Studio\\2022\\Enterprise\\VC\\Tools\\MSVC\\14.40.33807\\bin\\HostX64\\x64\\link.exe" "/NOLOGO" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\symbols.o" "C:\\a\\rust\\rust\\build\\x86_64-pc-windows-msvc\\stage1-rustc\\x86_64-pc-windows-msvc\\release\\deps\\rustc_main-abe621a284360afc.rustc_main.3d10d463641474e1-cgu.0.rcgu.o" "C:\\a\\rust\\rust\\build\\x86_64-pc-windows-msvc\\stage1-rustc\\x86_64-pc-windows-msvc\\release\\deps\\rustc_driver-99a7fdabd5344319.dll.lib" "C:\\a\\rust\\rust\\build\\x86_64-pc-windows-msvc\\stage1\\lib\\rustlib\\x86_64-pc-windows-msvc\\lib\\libcompiler_builtins-0a4b04fca3ab5f90.rlib" "psapi.lib" "shell32.lib" "ole32.lib" "uuid.lib" "advapi32.lib" "ws2_32.lib" "ntdll.lib" "advapi32.lib" "kernel32.lib" "ole32.lib" "oleaut32.lib" "bcrypt.lib" "advapi32.lib" "advapi32.lib" "cfgmgr32.lib" "gdi32.lib" "kernel32.lib" "msimg32.lib" "opengl32.lib" "user32.lib" "winspool.lib" "legacy_stdio_definitions.lib" "kernel32.lib" "advapi32.lib" "ntdll.lib" "userenv.lib" "ws2_32.lib" "kernel32.lib" "kernel32.lib" "/defaultlib:libcmt" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\advapi32.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\api-ms-win-core-errorhandling-l1-1-3.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\api-ms-win-core-file-fromapp-l1-1-0.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\api-ms-win-core-handle-l1-1-0.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\api-ms-win-core-ioring-l1-1-0.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\api-ms-win-core-synch-l1-2-0.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\api-ms-win-core-sysinfo-l1-2-0.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\api-ms-win-core-sysinfo-l1-2-3.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\api-ms-win-core-sysinfo-l1-2-4.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\api-ms-win-core-util-l1-1-1.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\api-ms-win-core-wow64-l1-1-1.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\api-ms-win-security-base-l1-2-2.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\avrt.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\bcp47mrm.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\bcryptprimitives.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\clfsw32.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\dbghelp.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\elscore.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\fwpuclnt.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\gdi32.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\icu.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\imagehlp.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\iphlpapi.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\kernel32.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\ktmw32.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\mswsock.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\netapi32.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\normaliz.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\ntdll.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\ntdllk.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\ole32.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\oleaut32.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\psapi.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\rtworkq.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\txfw32.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\user32.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\usp10.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\version.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\windows.networking.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\wofutil.dll_imports_indirect.lib" "C:\\a\\_temp\\msys64\\tmp\\rustcBM1zZI\\ws2_32.dll_imports_indirect.lib" "/NXCOMPAT" "/LIBPATH:C:\\Program Files\\Microsoft Visual Studio\\2022\\Enterprise\\VC\\Tools\\MSVC\\14.40.33807\\atlmfc\\lib\\x64" "/LIBPATH:C:\\a\\rust\\rust\\build\\x86_64-pc-windows-msvc\\stage1-rustc\\x86_64-pc-windows-msvc\\release\\build\\stacker-7601138db8f1aeb3\\out" "/LIBPATH:C:\\Program Files\\Microsoft Visual Studio\\2022\\Enterprise\\VC\\Tools\\MSVC\\14.40.33807\\atlmfc\\lib\\x64" "/LIBPATH:C:\\a\\rust\\rust\\build\\x86_64-pc-windows-msvc\\stage1-rustc\\x86_64-pc-windows-msvc\\release\\build\\psm-fbf8fbd57a82bcdd\\out" "/LIBPATH:C:\\Program Files\\Microsoft Visual Studio\\2022\\Enterprise\\VC\\Tools\\MSVC\\14.40.33807\\atlmfc\\lib\\x64" "/LIBPATH:C:\\a\\rust\\rust\\build\\x86_64-pc-windows-msvc\\stage1-rustc\\x86_64-pc-windows-msvc\\release\\build\\rustc_llvm-196fa4f2fa841609\\out" "/LIBPATH:C:\\a\\rust\\rust\\build\\x86_64-pc-windows-msvc\\llvm\\lib" "/OUT:C:\\a\\rust\\rust\\build\\x86_64-pc-windows-msvc\\stage1-rustc\\x86_64-pc-windows-msvc\\release\\deps\\rustc_main-abe621a284360afc.exe" "/OPT:REF,ICF" "/DEBUG" "/PDBALTPATH:%_PDB%" "/MANIFEST:EMBED" "/MANIFESTINPUT:C:\\a\\rust\\rust\\compiler\\rustc\\Windows Manifest.xml" "/WX"
  = note: LINK : fatal error LNK1104: cannot open file 'C:\a\rust\rust\build\x86_64-pc-windows-msvc\stage1-rustc\x86_64-pc-windows-msvc\release\deps\rustc_main-abe621a284360afc.exe'␍

[RUSTC-TIMING] rustc_main test:false 0.626
error: could not compile `rustc-main` (bin "rustc-main") due to 1 previous error
Build completed unsuccessfully in 0:07:49

bors · 2024-08-12T09:09:24Z

💔 Test failed - checks-actions

nikic · 2024-08-12T09:10:23Z

@bors retry

bors · 2024-08-12T10:20:04Z

⌛ Testing commit cfadfab with merge e08b80c...

bors · 2024-08-12T12:46:42Z

☀️ Test successful - checks-actions
Approved by: nikic
Pushing e08b80c to master...

rust-timer · 2024-08-12T14:03:00Z

Finished benchmarking commit (e08b80c): comparison URL.

Overall result: ✅ improvements - no action needed

@rustbot label: -perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.3%	[-0.3%, -0.3%]	2
Improvements ✅ (secondary)	-1.7%	[-1.7%, -1.7%]	1
All ❌✅ (primary)	-0.3%	[-0.3%, -0.3%]	2

Max RSS (memory usage)

Results (primary -2.5%, secondary 2.1%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	2.1%	[1.2%, 2.8%]	3
Improvements ✅ (primary)	-2.5%	[-2.5%, -2.5%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.5%	[-2.5%, -2.5%]	1

Cycles

This benchmark run did not return any relevant results for this metric.

Binary size

Results (secondary 0.0%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.0%	[0.0%, 0.0%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.0%	[-0.0%, -0.0%]	1
All ❌✅ (primary)	-	-	0

Bootstrap: 753.683s -> 753.452s (-0.03%)
Artifact size: 339.37 MiB -> 341.39 MiB (0.59%)

Stop generating assumes for validity ranges After rust-lang#128371, we represent validity ranges as parameter / return value attributes, so we no longer need to use assumes. r? `@ghost`

jieyouxu · 2024-08-15T04:21:44Z

The job dist-x86_64-msvc failed! Check out the build log: (web) (plain)
Click to see the possible cause of the failure (guessed by this bot)
error: linking with `link.exe` failed: exit code: 1104
[...]
 = note: LINK : fatal error LNK1104: cannot open file 'C:\a\rust\rust\build\x86_64-pc-windows-msvc\stage1-rustc\x86_64-pc-windows-msvc\release\deps\rustc_main-abe621a284360afc.exe'␍

Pinging back #127883.

rustbot assigned compiler-errors Jul 29, 2024

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Jul 29, 2024

tgross35 reviewed Jul 29, 2024

View reviewed changes

tgross35 added the S-blocked Status: Marked as blocked ❌ on something else such as an RFC or other implementation work. label Jul 29, 2024

andjo403 force-pushed the rangeAttribute branch from ee0e5a4 to 5ff174c Compare July 30, 2024 09:07

erikdesjardins reviewed Jul 30, 2024

View reviewed changes

This comment has been minimized.

Sign in to view

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jul 31, 2024

This comment has been minimized.

Sign in to view

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Aug 1, 2024

andjo403 force-pushed the rangeAttribute branch from 5ff174c to 009b4e2 Compare August 1, 2024 18:25

andjo403 commented Aug 1, 2024

View reviewed changes

erikdesjardins reviewed Aug 1, 2024

View reviewed changes

andjo403 force-pushed the rangeAttribute branch from 009b4e2 to 12f8b2d Compare August 1, 2024 23:07

erikdesjardins reviewed Aug 2, 2024

View reviewed changes

andjo403 force-pushed the rangeAttribute branch from 12f8b2d to 5cc0300 Compare August 3, 2024 10:39

This comment has been minimized.

Sign in to view

andjo403 force-pushed the rangeAttribute branch from 5cc0300 to 3cbfa82 Compare August 3, 2024 11:12

erikdesjardins mentioned this pull request Aug 11, 2024

Try to shrink Alignment-related MIR in Layout #128918

Open

Add range attribute to scalar function results and arguments

cfadfab

andjo403 force-pushed the rangeAttribute branch from 01a3604 to cfadfab Compare August 11, 2024 21:19

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Aug 12, 2024

bors added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Aug 12, 2024

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Aug 12, 2024

bors added the merged-by-bors This PR was explicitly merged by bors. label Aug 12, 2024

bors merged commit e08b80c into rust-lang:master Aug 12, 2024
7 checks passed

rustbot added this to the 1.82.0 milestone Aug 12, 2024

andjo403 deleted the rangeAttribute branch August 12, 2024 13:13

erikdesjardins mentioned this pull request Aug 12, 2024

Stop generating assumes for validity ranges #129027

Draft

slanterns mentioned this pull request Aug 14, 2024

Cleanup NonZero<T>::get #129081

Closed

jieyouxu mentioned this pull request Aug 15, 2024

High failure rate on Windows MSVC CI with filesystem errors #127883

Open

lqd mentioned this pull request Sep 5, 2024

SIGTRAP during compilation on MacOS #129887

Closed


		use std::num::NonZero;

		// CHECK: noundef range(i64 1, 0) i64 @nonzero_int(i64 noundef range(i64 1, 0) %x)

	// CHECK: noundef range(i64 1, 0) i64 @nonzero_int(i64 noundef range(i64 1, 0) %x)
	// CHECK-LABEL: @nonzero_int
	// CHECK-SAME: noundef range(i64 1, 0) i64 @nonzero_int(i64 noundef range(i64 1, 0) %x)

		// Checks that range metadata gets emitted on functions result and arguments
		// with scalar value.

	pub fn enum1_value(x: Enum0) -> Enum0 {
	pub fn enum1_value(x: Enum1) -> Enum1 {

Add range attribute to scalar function results and arguments #128371

Add range attribute to scalar function results and arguments #128371

Conversation

andjo403 commented Jul 29, 2024 • edited by nikic Loading

rustbot commented Jul 29, 2024

tgross35 Jul 29, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tgross35 Jul 29, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tgross35 commented Jul 29, 2024 • edited Loading

erikdesjardins left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nikic commented Jul 31, 2024

This comment has been minimized.

bors commented Jul 31, 2024

bors commented Jul 31, 2024

This comment has been minimized.

rust-timer commented Aug 1, 2024

Overall result: ✅ improvements - no action needed

andjo403 commented Aug 1, 2024

Choose a reason for hiding this comment

erikdesjardins Aug 1, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

andjo403 commented Aug 1, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

erikdesjardins Aug 2, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

This comment has been minimized.

nikic commented Aug 7, 2024

andjo403 commented Aug 11, 2024

nikic commented Aug 12, 2024

bors commented Aug 12, 2024

bors commented Aug 12, 2024

rust-log-analyzer commented Aug 12, 2024

bors commented Aug 12, 2024

nikic commented Aug 12, 2024

bors commented Aug 12, 2024

bors commented Aug 12, 2024

rust-timer commented Aug 12, 2024

Overall result: ✅ improvements - no action needed

jieyouxu commented Aug 15, 2024

andjo403 commented Jul 29, 2024 •

edited by nikic

Loading

tgross35 Jul 29, 2024 •

edited

Loading

tgross35 Jul 29, 2024 •

edited

Loading

tgross35 commented Jul 29, 2024 •

edited

Loading

erikdesjardins Aug 1, 2024 •

edited

Loading

erikdesjardins Aug 2, 2024 •

edited

Loading