c 
Sutter's Mill
FEBRUARY 12, 2024
I generally give one or two courses a year on C++ and related technologies. Because “high-performance low-latency” is kind of C++’s bailiwick, and because it’s my course, you’ll be unsurprised to learn that the topics and code focus on C++ and include coverage of modern C++17/20/23 features.
HammerDB
OCTOBER 23, 2023
HammerDB has always used stored procedures as a design decision because the original benchmark was implemented as close as possible to the example workload in the TPC-C specification that uses stored procedures. Additionally, reviewing official TPC-C full disclosure reports highlighted that all vendors also use stored procedures.
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
CSS Wizardry
OCTOBER 16, 2023
When serving and storing files on the web, there are a number of different things we need to take into consideration in order to balance ergonomics, performance, and effectiveness. Plotted on the same horizontal axis of 1.6s, the waterfalls speak for themselves: 201ms of cumulative latency; 109ms of cumulative download. It gets worse.
Brendan Gregg
OCTOBER 8, 2018
Screenshot: tracing read latency for PID 181: # bpftrace -e 'kprobe:vfs_read /pid == 30153/ { @start[tid] = nsecs; } kretprobe:vfs_read /@start[tid]/ { @ns = hist(nsecs - @start[tid]); delete(@start[tid]); }'. Here's a comparison for an average-sized DTraceToolkit script (seeksize.d) Attaching 2 probes. ^C variable curthread curtask.
The Morning Paper
AUGUST 29, 2019
We group the DBMS design choices and tradeoffs into three broad categories, which result from the need for dealing with (A) external storage; (B) query executors that are spun on demand; and (C) DBMS-as-a-service offerings. InS does now offer an NVMe variant too, and the authors perform limited testing on that as well. Key findings.
Brendan Gregg
OCTOBER 7, 2018
Screenshot: tracing read latency for PID 181: # bpftrace -e 'kprobe:vfs_read /pid == 30153/ { @start[tid] = nsecs; } kretprobe:vfs_read /@start[tid]/ { @ns = hist(nsecs - @start[tid]); delete(@start[tid]); }'. Here's a comparison for an average-sized DTraceToolkit script (seeksize.d) Attaching 2 probes. ^C variable curthread curtask.
The Netflix TechBlog
JANUARY 24, 2022
To that end, it’s important that we prevent significant performance regressions from reaching the production app. Any performance regression that makes it into a product release will degrade user experience, so the challenge is to detect and fix such regressions before they ship. What do we mean by Performance?
Expert insights. Personalized for you.
40 
Let's personalize your content