Skip to content

sha1-checked: implement hardware acceleration on arm64 and x86_64 #910

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
srijs wants to merge 1 commit into
base: master
Choose a base branch
from
Open

sha1-checked: implement hardware acceleration on arm64 and x86_64 #910

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions Cargo.lock
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

4 changes: 4 additions & 0 deletions sha1-checked/Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,10 +18,14 @@ exclude = [
]

[dependencies]
cfg-if = "1.0"
digest = "0.11"
sha1 = { version = "0.11.0-rc.5", default-features = false }
zeroize = { version = "1.8", default-features = false, optional = true }

[target.'cfg(any(target_arch = "aarch64", target_arch = "x86", target_arch = "x86_64"))'.dependencies]
cpufeatures = "0.3"

[dev-dependencies]
digest = { version = "0.11", features = ["dev"] }
hex-literal = "1"
Expand Down
16 changes: 14 additions & 2 deletions sha1-checked/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,8 +18,20 @@ unsuitable for further use in any security critical capacity, as it is
But, this crate provides the detection [algorithm] pioneered by git, to detect hash collisions when they
occur and prevent them. The [paper] has more details on how this works.

This implementation will be slower to use than the pure SHA-1 implementation, as it has to do more computations and
it can not rely on hardware acceleration available on some CPUs.
## Performance

This implementation is slower than plain SHA-1, since it does extra work per block to detect collisions.
Measured against this crate's own benchmarks, at throughput relative to plain, undetected SHA-1 on the same backend:

| architecture | scalar | hardware-accelerated |
|--------------|--------|----------------------|
| `aarch64` | 64% | 63% |
| `x86_64` | 55% | 38% |

Where the CPU's SHA-1 instructions are available, most blocks run through them, falling back to scalar
compression only when a potential collision is flagged. On `aarch64` this keeps detection roughly the same
fraction of plain SHA-1's speed as without hardware acceleration. On `x86_64`, `sha1`'s own hardware backend
speeds up by more than `sha1-checked`'s fixed per-block bookkeeping does, so the gap widens.

## Examples

Expand Down
Loading
Loading