What needs positioning
Profile-Guided Optimization (PGO) build-level performance improvements in Percona Server for MySQL, shipping in the next 8.4 and 9.7 point releases. PGO is a compiler optimization: the binary is rebuilt using runtime profiling data from representative workloads, so the compiler can optimize hot code paths (inlining, branch layout, code locality). The result is better runtime performance with no application, schema, or query changes. Customers get the improvement by upgrading to the PGO-built binaries.
This builds on the build-level optimization already in our packages (Link-Time Optimization, WITH_LTO=ON, in recent 8.0 builds). PGO is the next step in that line.
What is shipping and where
- Percona Server for MySQL 8.4.10
- Percona Server for MySQL 9.7.1
Why it matters (Performance pillar)
- Performance customers get for free on upgrade: higher throughput and lower latency from the same hardware, with no code, schema, or configuration changes.
- Better price/performance: more work per core supports cost-optimization conversations (do more on existing infrastructure, or right-size down).
- Open source, no feature gating: this is in the standard open source builds, not a paid tier.
Target audience
- Existing Percona Server for MySQL users planning routine upgrades (the benefit lands automatically on upgrade).
- High-throughput and latency-sensitive workloads evaluating price/performance.
- Cost-optimization and consolidation buyers comparing run-rate on existing hardware.
Pillar alignment
Primary: Performance and Reliability at Scale. Supporting: Optimized TCO / cost optimization (price/performance, more work per core) and Future readiness (continued engine performance investment in the open). Intended to slot into the pillar-alignment work rather than stand alone.
Open items before publication
- Exact performance claims (throughput or latency percentages) need MySQL engineering sign-off before publication. Do not publish specific numbers until confirmed.
- Add the 8.4.10 and 9.7.1 release notes and any PS Jira reference once they are public.
- PXC scope is confirmed: PGO build scripts cover Percona Server and Percona XtraDB Cluster (PS-11306), and PS features inherit into PXC at the same version (PXC engineering).
Existing references
What needs positioning
Profile-Guided Optimization (PGO) build-level performance improvements in Percona Server for MySQL, shipping in the next 8.4 and 9.7 point releases. PGO is a compiler optimization: the binary is rebuilt using runtime profiling data from representative workloads, so the compiler can optimize hot code paths (inlining, branch layout, code locality). The result is better runtime performance with no application, schema, or query changes. Customers get the improvement by upgrading to the PGO-built binaries.
This builds on the build-level optimization already in our packages (Link-Time Optimization,
WITH_LTO=ON, in recent 8.0 builds). PGO is the next step in that line.What is shipping and where
Why it matters (Performance pillar)
Target audience
Pillar alignment
Primary: Performance and Reliability at Scale. Supporting: Optimized TCO / cost optimization (price/performance, more work per core) and Future readiness (continued engine performance investment in the open). Intended to slot into the pillar-alignment work rather than stand alone.
Open items before publication
Existing references
use-cases-value-pillars/performance-reliability.mduse-cases-value-pillars/cost-optimization.md