Assist JIT in eliminating bounds checks

[xtqqczze]

Follow-up: #81001.

Change condition in loops from index != T.Length to index < T.Length where T:

• U[]
• string
• Span<U>
• ReadOnlySpan<U>

cc: @stephentoub

[ghost]

I couldn't figure out the best area label to add to this PR. If you have write-permissions please help me learn by adding exactly one area label.

[gfoidl]

Do these need uint-casts to eliminate the bound checks?

[xtqqczze]

In this case casts appear to regress bounds check elimination.

Perhaps this is unnecessary because srcIndex is a local variable that is zero-initialized?

runtime/src/libraries/System.Private.CoreLib/src/System/Buffers/Text/Utf8Parser/Utf8Parser.Number.cs

Line 34 in ce18ccf

int srcIndex = 0;

[stephentoub]

In this case casts appear to regress bounds check elimination.

Can you elaborate? I wouldn't expect this to remove the bounds check on source[srcIndex] without casting srcIndex to uint here.

[xtqqczze]

We already do checks on source.Length before the loop. Adding uint casts seems to cause JIT to "forget" about these.

[stephentoub]

I still don't understand what you mean. Can you share a sharplab.io example?

[xtqqczze]

It's worth noting that SharpLab is .NET 7

@tannergooding For future reference, does the codegen from Compiler Explorer have the same optimizations as a local disassembly using .NET 8 nightly build.

Currently, .e.g. https://csharp.godbolt.org/z/cKhTfjvqd, seems to be using:
// crossgen2 8.0.0-preview.3.23152.99+8d2ddc45e8f6f2a726c2b91c1ba2b9f7208b1f3b

[EgorBo]

@xtqqczze for godbolt you need to pass -O or better --Ot argument to compiler to enable optimizations. godbolt uses crossgen (R2R) which is, on average, worse than what JIT emits (sharplab). Although, even sharplab's codegen is a bit synthetic since in the real world our JIT is tiered and Tier1 codegen might be different from what you see in sharplab.

[xtqqczze]

for godbolt you need to pass -O or better --Ot

@EgorBo Not seeing any differeences in codegen for this example: https://csharp.godbolt.org/z/cKhTfjvqd, is this expected?

[tannergooding]

Yes. However, its showing the crossgen code and so isn't strictly the same as what sharplab shows or what a user will see for Tier 1 codegen.

You can customize the instruction set targeted and I'd recommend using --instruction-set x86-x64-v3 as that targets AVX2, BMI1, BMI2, LZCNT, MOVBE, and FMA

You can customize the OS: --targetos windows (defaults to linux)
and the architecture: --targetarch arm64 (defaults to x64)

and in general pass in other options supported by crossgen (such as --Ot to optimize for speed)

[xtqqczze]

@tannergooding Great help with your tips, I knew about the parameters but couldn't find a useful reference with a search engine.

[ghost]

Tagging subscribers to this area: @dotnet/area-meta
See info in area-owners.md if you want to be subscribed.

Issue Details

---

Follow-up: #81001.

Change condition in loops from index != T.Length to index < T.Length where T:

• U[]
• string
• Span<U>
• ReadOnlySpan<U>

cc: @stephentoub

Author:	xtqqczze
Assignees:	-
Labels:	area-Meta, community-contribution
Milestone:	-

[hughbe]

Should we create an issue to add support to JIT to eliminate range checks for these?

[xtqqczze]

Should we create an issue to add support to JIT to eliminate range checks for these?

Issue opened: #84697.

[stephentoub]

LGTM. Most of these won't actually help eliminate bounds checks like the title suggests, but they're fine for clarity.

[stephentoub]

LGTM. Most of these won't actually help eliminate bounds checks like the title suggests, but they're fine for clarity.