bitshifter / glam-rs
99%

DEFAULT BRANCH: main
Repo Added
Files 154
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

LAST BUILD ON BRANCH main
branch: SELECT
CHANGE BRANCH
x

coverage: 99.058% (+0.005%) from 99.053%
35218159043

push

github

web-flow 
perf(simd): broadcast dot product without an extra shuffle (#833)

The SSE2 broadcast dot product can save a shuffle by computing the
result directly in all four lanes.

## Solution

`dot4_into_m128` now sums the products directly in all four lanes,
keeping the `(x + z) + (y + w)` grouping. Tests and benchmarks are
included.

## Benchmarks

On my Xeon E5-2696 v4:

| Measurement | Before | After | Change |
|---|---:|---:|---:|
| 256-dot-product batch | 301.673 ns | 267.888 ns | -11.2% |
| Throughput | 848.60 M/s | 955.62 M/s | +12.6% |
| Quaternion lerp | 6.491 ns | 6.095 ns | -6.1% |
| Cycles/dot | 3.772 | 3.335 | -11.6% |
| Instructions/dot | 12.104 | 11.104 | -8.3% |
| Allocations | 0 | 0 | Unchanged |

Timing: Criterion on Windows. Counters: five alternating, core-pinned
runs under WSL. Normalization timing stayed within the noise threshold
(+0.7%).

## Testing and linting

- Three million comparisons passed: non-NaN outputs matched bit for bit;
NaNs matched by classification.
- Regression tests cover cancellation, signed zero, subnormals,
overflow, infinities, and NaNs.
- All nine feature configurations, MSRV, and portable SIMD checks
passed.
- Glam-specific Clippy and formatting passed.
- The existing Windows CI-tool lint remains.

🤖 AI disclosure: I used Codex with GPT-6 Astra at xhigh reasoning effort
to help identify the optimization and construct the benchmark. I have
reviewed and understand the resulting implementation.

8 of 8 new or added lines in 2 files covered. (100.0%)

93262 of 94149 relevant lines covered (99.06%)

123554.19 hits per line

Relevant lines Covered
94149 RELEVANT LINES 93262 COVERED LINES
123554.19 HITS PER LINE
Source Files on main

Recent builds

Builds Branch Commit Type Ran Committer Via Coverage
35218159043 main perf(simd): broadcast dot product without an extra shuffle (#833) The SSE2 broadcast dot product can save a shuffle by computing the result directly in all four lanes. ## Solution `dot4_into_m128` now sums the products directly in all four lane... push web-flow github
99.06
35217344094 perf/sse2-dot4-broadcast Merge a43a9e36f into 3ccb58c04 Pull #833 web-flow github
99.06
35210906750 main chore(bench): fold throughput into bench macros and document benchmarking (#852) Batch the criterion bench macros by size with `Throughput::Elements` (defaults `[16, 1024]`) and expand the contributing benchmark and optimization guidance. push web-flow github
99.05
35209744136 bench-macro-throughput Merge 054a2fc33 into 296329c1d Pull #852 web-flow github
99.05
35209470581 bench-macro-throughput Merge d2ebc9849 into 296329c1d Pull #852 web-flow github
99.05
35209138232 bench-macro-throughput Merge dda2a85c1 into 296329c1d Pull #852 web-flow github
99.04
35206671206 bench-macro-throughput Merge 5e9c7e469 into 296329c1d Pull #852 web-flow github
99.05
35092415884 main docs: hide deref helper types from rustdoc (#851) Marks the `Deref` target structs in `src/deref.rs` (`Vec3<T>`, `Vec4<T>`, `Cols2/3/4<V>`) as `#[doc(hidden)]`. These types are implementation details: they live in a private module and are not pu... push web-flow github
99.03
35092044938 docs/deref-doc-hidden Merge 83af27f56 into f18a1b780 Pull #851 web-flow github
99.03
35090265560 main docs: document API guideline deviations in ARCHITECTURE.md (#848) Adds a section documenting where glam deliberately deviates from the Rust API guidelines and why: - Public fields (C-STRUCT-PRIVATE): layout is part of the API - Conversion naming... push web-flow github
99.03
See All Builds (1874)