Details
-
Enhancement
-
Status: Resolved
-
P3
-
Resolution: Fixed
-
None
-
b26
Description
Create intrinsics for the ChaCha20 algorithm for x86_64 and aarch64. On x86_64, support for AVX, AVX2 and AVX-512 should be taken into account to support a wide array of x86 chips. For the aarch64 platform, a baseline ASIMD implementation should be created, with possible enhancements from the SVE instruction set for supporting processors.
The current approach will accelerate the core 20-round ChaCha20 block function, as outlined in RFC 7539, section 2.3.
The current approach will accelerate the core 20-round ChaCha20 block function, as outlined in RFC 7539, section 2.3.
Attachments
Issue Links
- is blocked by
-
JDK-8296233 C2 crashes with Unschedulable Graph in ChaCha20Poly1305 microbenchmark with ChaCha20 intrinsics and UseAVX=3
-
- Closed
-
- relates to
-
JDK-8302007 Improve ChaCha20 intrinsics for single-part encryption/decryption on AVX-512
-
- Open
-
-
JDK-8288047 Accelerate Poly1305 on x86_64 using AVX512 instructions
-
- Resolved
-
-
JDK-8298055 AArch64: fastdebug build fails after JDK-8247645
-
- Resolved
-
-
JDK-8298592 Add java man page documentation for ChaCha20 and Poly1305 intrinsics
-
- Resolved
-
(2 links to)