-
### Interesting Upcoming Features from Low Latency, Parallelism and Concurrency From Kona 2023, Tokyo 2024, and St. Louis 2024
PAUL E. MCKENNEY,
MAGED MICHAEL
& MICHAEL WONG
20
24
September 15
0 码力 |
56 页 |
514.85 KB
| 1 年前 3
-
## +24
## Work Contracts Rethinking Task Based Concurrency and Parallelism for Low Latency C++
MICHAEL A MANISCALCO
### “We cannot solve our problems with the same thinking we used when we created them Michael A Maniscalco
○ Software Architect and Principal Developer at Lime Trading
○ We develop low latency market data and trading software for use in HFT
## • Personal
github.com/buildingcpp
Work Contracts hash_task<64>; // duration = approx 1.1 $ \mu $ s
using low_content_task = hash_task<256>; // duration = approx 4.2 $ \mu $ s
void bench()
{
bench<low_content_task>();
bench();
0 码力 |
142 页 |
2.80 MB
| 1 年前 3
-
Google™
Go GC:
Latency Problem Solved
Rick Hudson
Google Engineer
## My Codefendants: The Cambridge Runtime Gang


#1 Barrier: GC Latency
## When is the best time to do a GC?
When nobody is looking.
Using camera to track eye movement jpg)
## Pop up a network wait icon
## Waiting
## Or Trade Throughput for Reduced GC Latency
## Google
## Latency
## Nanosecond
1: Grace Hopper Nanosecond 11.8 inches Microsecond
5.4: Time light travels
0 码力 |
20 页 |
897.11 KB
| 2 年前 3
-
## Designing an ultra low-overhead multithreading runtime for Nim
Mamy Ratsimbazafy
mamy@numforge.co
## Hello!
## I am Mamy Ratsimbazafy
During the day blockchain/Ethereum 2 developer (in Nim)
During viable runtime plan in a weekend
## Understanding the design space
Concurrency vs parallelism, latency vs throughput
Cooperative vs preemptive, IO vs CPU
## Parallelism is not concurrence
Concurrent multithreading runtime level.
## The problem
## How to schedule M tasks on N hardware threads?
## Latency vs Throughput
## Do we want to do all the work in a minimal amount of time?
Numerical computing
0 码力 |
37 页 |
556.64 KB
| 1 年前 3
-
Leveraging C++20/23 Features for Low Level Interactions
©
## Overview
In a baremetal environment, we’re going to demonstrate effective use of C++
How did we end up with C language HW interactions?
What cleaner, safer, and more testable?
Why is C so prolific?
It's the kernel, silly!
30 years ago, all low level interaction was done with C
Organizations are very comfortable with C in embedded applications
0 码力 |
56 页 |
5.39 MB
| 1 年前 3
-
## Leveraging Istio for Creating API Tests Low Effort API Testing for Microservices
## mesh dynamics
## Structure
## • What has changed?
– Migration to microservices triggering need for extensive API testing earlier
Create and maintain a balanced test pyramid
Create different types of tests with low effort

Service
0 码力 |
21 页 |
1.09 MB
| 2 年前 3
-
thermostats)
▶ Medical equipment (e.g. pacemakers)
## ▶ Characteristics:
▶ Limited hardware resources
▶ Low power consumption
▶ May have real-time performance constraints
## Machine learning on embedded devices devices
Alternative to cloud-based machine learning
Advantages:
▶ Real-time processing
▶ Low latency
▶ Reduced bandwidth usage
▶ Offline operation
▶ Improved privacy
## ▶ Disadvantages:
▶ Compatibility devices
C++ is widely used in embedded systems:
▶ Was designed with efficiency in mind
▶ Provides low-level access to hardware resources
▶ Provides high-level abstractions
Image classification
## Supervised
0 码力 |
51 页 |
1.78 MB
| 1 年前 3
-
Options Automated Trading Systems team at Optiver, a global market maker company
- Worked on low-latency systems for 15 years in Trading and Defense

Ultra low-latency execution (< 1µs)
## Order book deep dive
“The order book is the heart of any trading system
0 码力 |
123 页 |
5.89 MB
| 1 年前 3
-
dcf/p9_2.jpg)
| TPS | Latency(ms) | CPU |
| Consumer | Producer |
| 优化前 | 81986 | 1 业务线程
业务线程
业务线程
| TPS | Latency(ms) | CPU | | Consumer | Producer | 原始单连接*10 0 码力 |
17 页 |
2.22 MB
| 2 年前 3 -

if there is a low probability
## STABILIZER repeatedly randomizes layout
## this speedup is real

found 8,000,000 similar images
(latency)
(thruput)
How fast do results come back?
## Progress Points

Luke wants to minimize response time.
He adds latency progress points.
## Progress Points
Little's Law: We Lear
latency = transactions / throughput
![Image](/uploads/documents/6/9 0 码力 |
197 页 |
11.90 MB
| 1 年前 3
|
|