杨杰, yangjie

wrong pool unmarshal slize

3

wrong pool unmarshal size. my proto file: ``` // protoc --go_out=. --plugin protoc-gen-go="/Users/jie.yang05/go/bin/protoc-gen-go" --go-vtproto_out=. --plugin protoc-gen-go-vtproto="/Users/jie.yang05/go/bin/protoc-gen-go-vtproto" --go-vtproto_opt=features=marshal+unmarshal+size+pool ./lineentry.proto syntax = "proto3"; package index; option go_package="./proto"; import "github.com/planetscale/vtprotobuf/vtproto/ext.proto"; message lineEntries {...

use table to accelerate bsr calculate

12

在计算bsr的时候，某些操作系统上使用一种for循环的方式计算，时间复杂度是O(n)，n是这个数第从小到大最大的为1的位。 经过golang sdk库math/bits/bits.go/Len函数的启发，我想出了一种算法，具有O（1）的复杂度, 并且计算时间稳定。 并且针对64位操作系统和32位操作系统有不同的优化。 几乎减少了百分之九十的计算时间。

i want to know a process's network file cache. the mincore is not useful.

is sonic support jit perf?

1

in linux specification: jit program could generate a file in /proc/pid/root/tmp/perf-pid.map, which contain jit symbol name and it's address. so perf can use it to tran address to function name....

zstd performance compare

5

i notice you replace zstd.Write to zstd.EncodeAll but without performance compare in https://github.com/mostynb/go-grpc-compression/commit/5f2489304fdefc331e2803e6574af3b8c8f4f00a. i implement different type of zstd encoding and benchmark it. 1, zstd-stream. ``` package main import (...

overhead negligible goroutine sampling

9

The main cost of goroutine performance analysis comes from stop the world. And each goroutine needs to traverse the stack through an expensive unwinder function. We can optimize it by...

**Is your feature request related to a problem? Please describe.** no **Describe the solution you'd like** Now we can get the traceid of goroutine through ebpf. If we can uprobe...

目前我在尝试通过 fp backtrace 的方式回溯每个 goroutine 的调用栈，遇到一些问题。 在 time.sleep，g.sched.pc 可以表示 goroutine 当前的执行函数。 但是如果 goroutine 执行的是一个 for 循环的函数，g.sched.pc 就是 runtime.systemstack_switch。 如何才能获取到该 goroutine 正在执行的函数呢？