优化原生结果编码传输与浏览器缓存,记录八路性能基线

原生结果series通过字节索引直传,C端使用Ryu精确回读编码和64 KiB批量写出;网页采用Float64缓存和CSV工作线程,减少结果处理与保存等待。

补充八路AME曲线核查、全流程分阶段计时、独立编码基准和复现工具,固定后续优化采用修正八路及rtol=1e-8。C写出1.1808→0.1638 s,点击到可查看8.0100→6.9756 s。

验证:最终10项编码专项、29项相关后端回归通过;8份原生结果逐位一致,16次网页结果/CSV/刷新恢复通过。前端构建及缓存/CSV专项在本轮结果处理工作中通过。环境、原始大结果与临时构建不纳入Git。
This commit is contained in:
lujingze committed 2026-09-11 15:09:15 +00:00
1 parent 808c484f5b
commit 3bc4be3c06
55 files changed
+20522 -141

No files matched your search

+1 -1
View File
@@ -52,7 +52,7 @@ def toolchain() -> tuple[str, Path, str]:
def build_native(program: NativeProgram, *, cache_dir: Path | None = None) -> NativeBuild:
start = time.perf_counter()
compiler, sundials, compiler_version = toolchain()
runtime = sorted(NATIVE.rglob("*.c")) + sorted((NATIVE / "include").glob("*.h"))
runtime = sorted(NATIVE.rglob("*.c")) + sorted((NATIVE / "include").rglob("*.h"))
flags = ["-std=c11", "-O3", "-Wall", "-Wextra", "-Werror", "-ffp-contract=off", "-fno-fast-math"]
executable_name = "model.exe" if os.name == "nt" else "model"
if os.name == "nt":