Mark the ConfigKit performance request complete
This commit is contained in:
@@ -1,6 +1,7 @@
|
|||||||
# Computational Efficiency Improvements
|
# Computational Efficiency Improvements
|
||||||
|
|
||||||
Status: Accepted scope following review; implementation has not started.
|
Status: Accepted scope following review. The ConfigKit request (item 1) is
|
||||||
|
complete; local performance implementation has not started.
|
||||||
|
|
||||||
## Purpose
|
## Purpose
|
||||||
|
|
||||||
@@ -45,25 +46,16 @@ their benefits still need measurement.
|
|||||||
|
|
||||||
### 1. Submit A ConfigKit Performance Feature Request
|
### 1. Submit A ConfigKit Performance Feature Request
|
||||||
|
|
||||||
Submit an appropriate feature request for ConfigKit's consideration, describing
|
Completed on 2026-10-08. The request is recorded in ConfigKit's
|
||||||
Promptkit's downstream configuration-construction workload and the measured
|
`docs/roadmap/promptkit-performance-request.md`, committed as `d53021e`.
|
||||||
allocation cost. Include the fixture shape, selected ConfigKit version,
|
It includes the downstream fixture, dependency and toolchain versions,
|
||||||
benchmark method, and allocation-profile evidence so upstream maintainers can
|
repeated benchmark results, reproduction commands, and allocation-profile
|
||||||
reproduce and assess it. The existing
|
observations, while preserving existing configuration and inspection semantics.
|
||||||
[construction benchmark](../../internal/configuration/benchmark_test.go) and
|
|
||||||
[configuration adapter](../../internal/configuration/load.go) provide the
|
|
||||||
downstream context.
|
|
||||||
|
|
||||||
Request more efficient construction while retaining the source composition,
|
Acceptance, design, and implementation remain with ConfigKit. Upstream
|
||||||
resolution, final-value handoff, and operator-inspection capabilities Promptkit
|
improvements will become available through a later ConfigKit dependency upgrade;
|
||||||
uses. Explain why libraries that build multiple engines or load large catalogs
|
no further Promptkit-specific work remains for this item at present. Local
|
||||||
benefit, and distinguish startup cost from ordinary prompt execution.
|
performance improvements can proceed independently.
|
||||||
|
|
||||||
Leave acceptance, internal design, and implementation to ConfigKit. Submission
|
|
||||||
and upstream consideration are the local outcome for this item; adoption of a
|
|
||||||
future upstream improvement can be assessed when one is available. Promptkit's
|
|
||||||
other improvements can proceed independently. Reimplementing ConfigKit's loader
|
|
||||||
or bypassing its configuration ownership is outside this scope.
|
|
||||||
|
|
||||||
### 2. Avoid Response Copies In Basic Validation
|
### 2. Avoid Response Copies In Basic Validation
|
||||||
|
|
||||||
@@ -162,8 +154,8 @@ allocation costs. Follow with request encoding and identity-only inspection
|
|||||||
|
|
||||||
Keep prepared-state consolidation (item 6) conditional on measured savings and
|
Keep prepared-state consolidation (item 6) conditional on measured savings and
|
||||||
a clear ownership analysis; it has the greatest risk of weakening isolation or
|
a clear ownership analysis; it has the greatest risk of weakening isolation or
|
||||||
concurrency behavior. The ConfigKit request (item 1) can proceed independently
|
concurrency behavior. The ConfigKit request (item 1) is complete and does not
|
||||||
and does not block local improvements. This order is a recommendation, not a
|
block local improvements. This order is a recommendation, not a
|
||||||
dependency between otherwise independent changes.
|
dependency between otherwise independent changes.
|
||||||
|
|
||||||
## Benchmarking And Verification
|
## Benchmarking And Verification
|
||||||
|
|||||||
Reference in New Issue
Block a user