加入 Gitee
与超过 1200万 开发者一起发现、参与优秀开源项目,私有仓库也完全免费 :)
免费加入
文件
克隆/下载
fptaylor.mli 1.56 KB
一键复制 编辑 原始数据 按行查看 历史
Alexey Solovyev 提交于 2020-12-07 20:17 . Updated js
(* ========================================================================== *)
(* FPTaylor: A Tool for Rigorous Estimation of Round-off Errors *)
(* *)
(* Author: Alexey Solovyev, University of Utah *)
(* *)
(* This file is distributed under the terms of the MIT license *)
(* ========================================================================== *)
(* -------------------------------------------------------------------------- *)
(* Main FPTaylor functions *)
(* -------------------------------------------------------------------------- *)
type error_type =
Err_abs_approx | Err_abs_exact |
Err_rel_approx | Err_rel_exact |
Err_ulp_approx | Err_ulp_exact
type error_result = {
error_type : error_type;
(* Total error *)
(* Lower bounds of error intervals represent lower bounds
returned by a global optimization procedure.
low = neg_infinity if a lower bound is not returned. *)
error : Interval.interval option;
(* Second order error *)
total2 : Interval.interval option;
(* Error model *)
error_model : Expr.expr option;
}
type result = {
task : Task.task;
real_bounds : Interval.interval;
errors : error_result list;
elapsed_time : float;
}
val error_type_name : error_type -> string
val validate_options : unit -> unit
val fptaylor : input_files:string list -> result list
Loading...
马建仓 AI 助手
尝试更多
代码解读
代码找茬
代码优化