JDK-8351488 : Improve performance of floating point reduction kernels
  • Type: Enhancement
  • Component: hotspot
  • Sub-Component: compiler
  • Affected Version: 25
  • Priority: P4
  • Status: Open
  • Resolution: Unresolved
  • CPU: x86_64
  • Submitted: 2025-03-10
  • Updated: 2025-09-26
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
Other
tbdUnresolved
Related Reports
Relates :  
JDK-8368778 - x86 backend support for add/mul reduction operations for Float16
Description
Our recent runs on different leading edge platforms show that Floating point reduction kernels are better off SLP vectorization than with SLP.

While ADD MUL shows degraded performance on Granite Rapids AP.
Min/Max shows performance degradation on Sierra Forest SP (AVX2 only Xeon)

Please find attached the data.