Brendan Gregg

MARCH 16, 2024

2015-2020: Overhead As part of production rollout I did many performance overhead tests, which I've described publicly before: The overhead of adding frame pointers to everything (libc and Java) was usually less than 1%, with one exception of 10%. The actual overhead depends on your workload. Others have reported around 1% and around 2%.

Java 145

Java Cache Google Hardware 145

Percona

MARCH 27, 2023

A lot of useful information can be retrieved from this schema, for example, table metadata and foreign key relations, but trying to query I_S can induce performance degradation if your server is under heavy load, as shown in the following example test. The same tests have been executed in Percona Server for MySQL 5.7

Cache 99

Cache Hardware Servers Benchmarking 99

John McCalpin

APRIL 25, 2023

I have a lot of historical data using my ReadOnly benchmark (as described in some of the earliest entries in this blog [link] A read-only access pattern removes the need to understand and explain the many complexities associated with the “streaming stores” typically used in the STREAM benchmark (e.g., Stay tuned!

Benchmarking 97

Benchmarking Cache Latency Tuning 97

The Morning Paper

SEPTEMBER 27, 2020

The wikipedia page on floating point numbers describes a number of related accuracy problems including the difficulty of testing for equality. In day-to-day usage, beyond judicious use of ‘within’ for equality testing, I suspect most of us ignore the potential difficulties of floating point arithmetic even if we shouldn’t.

Benchmarking 101

Benchmarking Hardware Programming Code 101

HammerDB

JUNE 27, 2022

HammerDB is a software application for database benchmarking. It enables the user to measure database performance and make comparative judgements about database hardware and software. Databases are highly sophisticated software, and to design and run a fair benchmark workload is a complex undertaking. The HammerDB name.

Benchmarking 62

Benchmarking Open Source C++ Cache 62

Scalegrid

JANUARY 25, 2024

Key metrics like throughput, request latency, and memory utilization are essential for assessing Redis health, with tools like the MONITOR command and Redis-benchmark for latency and throughput analysis and MEMORY USAGE/STATS commands for evaluating memory. It depends upon your application workload and its business logic.

Metrics 130

Metrics Monitoring Latency Cache 130

The Morning Paper

NOVEMBER 3, 2019

For the rest of us, if you really need that extra performance (maybe what you get out-of-the-box or with minimal tuning is good enough for your use case) then you can upgrade hardware and/or pay for a commercial license of a tuned distributed (RHEL). This causes average slowdowns of 66% across the select , poll , and epoll tests.

Performance 108

Performance Benchmarking Tuning Hardware 108

HammerDB

APRIL 19, 2023

Therefore, before we attempt to measure our database performance, we should know the system or cloud instance to be tested in detail. Benchmarking the target Two of the more popular database benchmarks for MySQL are HammerDB and sysbench. This allows us to know our operating environment and its capability. 4.22 %usr 38.40

Performance 54

Performance Benchmarking Cache Storage 54

Percona

MARCH 23, 2023

Hardware Memory The amount of RAM to be provisioned for database servers can vary greatly depending on the size of the database and the specific requirements of the company. Benchmark before you decide. The optimal value can be decided after testing multiple settings, starting from eight is a good choice. I hope this helps!

Performance 123

Performance Hardware Tuning Storage 123

Percona

JUNE 22, 2023

Defining high availability In general terms, high availability refers to the continuous operation of a system with little to no interruption to end users in the event of hardware or software failures, power outages, or other disruptions. If a primary server fails, a backup server can take over and continue to serve requests.

Availability 105

Availability Database Open Source Hardware 105

HammerDB

FEBRUARY 25, 2021

HammerDB is a load testing and benchmarking application for relational databases. All the databases that HammerDB tests implement a form of MVCC (multi-version concurrency control). This post explains why HammerDB made the language decisions it made to make it the best performing and most usable database benchmarking software.

Programming 67

Programming Benchmarking Virtualization C++ 67

Percona

AUGUST 22, 2023

Some of the most important elements include: No single point of failure (SPOF): You must eliminate any SPOF in the database environment, including any potential for an SPOF in physical or virtual hardware. Redundancy provides backups and safeguards against data loss in case of hardware failures. there cannot be high availability.

Availability 77

Availability Hardware Open Source Database 77

Percona

MAY 1, 2023

Key areas include: Configuration parameter tuning : This tuning involves altering variables such as memory allocation, disk I/O settings, and concurrent connections based on specific hardware and requirements. This not only results in cost savings by minimizing hardware requirements but also has the potential to decrease cloud expenses.

Tuning 52

Tuning Efficiency Database Performance 52

Percona

JUNE 22, 2023

Indexing efficiency Monitoring indexing efficiency in MySQL involves analyzing query performance, using EXPLAIN statements, utilizing performance monitoring tools, reviewing error logs, performing regular index maintenance, and benchmarking/testing. This KPI is also directly related to Query Performance and helps improve it.

Performance 103

Performance Monitoring Traffic Database 103

KeyCDN

AUGUST 15, 2019

Web performance is a broad subject, and you’ll find no shortage of performance testing tips and tutorials all over the web. Before you begin tuning your website or application, you must first figure out which metrics matter most to your users and establish some achievable benchmarks. What is Performance Testing?

Testing Tools 55

Testing Tools Best Practices Performance Testing Testing 55

Speed Curve

MARCH 7, 2022

These numbers should not be taken as a benchmark for your own site. You can see this by looking at the synthetic test result for Sears.com (again, available via our Industry Benchmarks ). Don't assume hardware and networks will mitigate page bloat. Not all pages are getting bigger. The CLS score for this page is 1.0468.

Mobile 145

Mobile Benchmarking Google Metrics 145

The Morning Paper

MAY 27, 2019

Last week we saw the benefits of rethinking memory and pointer models at the hardware level when it came to object storage and compression ( Zippads ). The protections are hardware implemented and cannot be forged in software. At hardware reset the boot code is granted maximally permissive architectural capabilities.

C++ 61

C++ Hardware Virtualization Benchmarking 61

Brendan Gregg

SEPTEMBER 25, 2021

As a Xen guest, this profile was gathered using perf(1) and the kernel's software cpu-clock soft interrupts, not the hardware NMI. There's also a test and println() in the loop to, hopefully, convince the compiler not to optimize-out an otherwise empty loop. This will slow this test a little.) But I'm not completely sure.

Speed 126

Speed Java AWS Virtualization 126

Adrian Cockcroft

JUNE 21, 2022

I became the Sun UK local specialist in performance and hardware, and as Sun transitioned from a desktop workstation company to sell high end multiprocessor servers I was helping customers find and fix scalability problems. We had specializations in hardware, operating systems, databases, graphics, etc.

Tuning 52

Tuning Benchmarking Engineering C++ 52

HammerDB

DECEMBER 5, 2020

A full understanding of why this is important requires some knowledge of the evolution of database hardware and software. The HammerDB TPROC-C workload by design intended as CPU and memory intensive workload derived from TPC-C – so that we get to benchmark at maximum CPU performance at a much smaller database footprint.

C++ 40

C++ Benchmarking Virtualization Open Source 40

Alex Russell

APRIL 29, 2021

As an engineer on a browser team, I'm privy to the blow-by-blow of various performance projects, benchmark fire drills, and the ways performance marketing (deeply) impacts engineering priorities. With each team, benchmarks lost are understood as bugs. is access to hardware devices. This is as it should be. Lower is better.

Media 145

Media Games Education Engineering 145

Percona

JANUARY 19, 2023

Example: Creating four simple tables to store strings but using different data types: db1 test> CREATE TABLE tb1 (id int auto_increment primary key, test_text char(200)); Query OK, 0 rows affected (0.11 sec) db1 test> CREATE TABLE tb2 (id int auto_increment primary key, test_text varchar(200)); Query OK, 0 rows affected (0.05

Open Source 86

Open Source Storage Database Big Data 86

Percona

AUGUST 5, 2023

This means that multiple partitions can be processed simultaneously, making better use of available hardware resources and further enhancing query performance. Test, Test, and Test Again : Before implementing partitioning in production, thoroughly test it in a controlled environment. Periodically review and adjust.

Strategy 40

Strategy Storage Database Best Practices 40

VoltDB

JUNE 1, 2017

Hardware Optimizers” want to get the maximum utilization out of hardware. These systems were designed to have a lifetime of half a decade or more, and rapidly changing hardware meant that the initial deployment had to be sized for 5-7 years out. Attendees could be broken down into several distinct groups. Where VoltDB fits.

Hardware 56

Hardware Latency Virtualization Cloud 56

Alex Russell

MARCH 6, 2021

Hardware Past As Performance Prologue. Using a global ASP as a benchmark can further mislead thanks to the distorting effect of ultra-high-end prices rising while shipment volumes stagnate. But the hardware future is not evenly distributed, and web workloads aren't heavily parallel. Today, either method returns a similar answer.

Performance 124

Performance Network Cache Metrics 124

Brendan Gregg

SEPTEMBER 25, 2021

As a Xen guest, this profile was gathered using perf(1) and the kernel's software cpu-clock soft interrupts, not the hardware NMI. There's also a test and println() in the loop to, hopefully, convince the compiler not to optimize-out an otherwise empty loop. This will slow this test a little.)

Speed 52

Speed Java AWS Virtualization 52

Scalegrid

JULY 13, 2020

These new applications are a great way for enterprise companies to test out PostgreSQL before migrating their entire infrastructure. Oracle support for hardware and software packages is typically available at 22% of their licensing fees. So Which Is Best?

Open Source 244

Open Source Tuning C++ Database 244

SQL Server According to Bob

SEPTEMBER 26, 2016

When we released Always On Availability Groups in SQL Server 2012 as a new and powerful way to achieve high availability, hardware environments included NUMA machines with low-end multi-core processors and SATA and SAN drives for storage (some SSDs). As we moved towards SQL Server 2014, the pace of hardware accelerated.

Availability 40

Availability Servers Hardware Benchmarking 40

Abhishek Tiwari

DECEMBER 27, 2017

A recent performance benchmark completed by Intel and BlueData using the BigBench benchmarking kit has shown that the performance ratios for container-based Hadoop workloads on BlueData EPIC are equal to and in some cases, better than bare-metal Hadoop [7]. Performance.

Big Data 63

Big Data Storage Benchmarking Hardware 63

The Morning Paper

MAY 14, 2019

Last time around we looked at the DeathStarBench suite of microservices-based benchmark applications and learned that microservices systems can be especially latency sensitive, and that hotspots can propagate through a microservices architecture in interesting ways. When available, it can use hardware level performance counters.

Big Data 45

Big Data Cloud Performance Hardware 45

VoltDB

JUNE 1, 2017

Hardware Optimizers” want to get the maximum utilization out of hardware. These systems were designed to have a lifetime of half a decade or more, and rapidly changing hardware meant that the initial deployment had to be sized for 5-7 years out. Attendees could be broken down into several distinct groups. Where VoltDB fits.

Hardware 40

Hardware Latency Virtualization Cloud 40

Alex Russell

JANUARY 30, 2024

HTML, CSS, images, and fonts can all be parsed and run at near wire speeds on low-end hardware, but JavaScript is at least three times more expensive, byte-for-byte. India's speed test medians are moving quickly, but variance is orders-of-magnitude wide, with 5G penetration below 25% in the most populous areas. 30% of worldwide volume."

Performance 90

Performance Network Mobile Speed 90

Speed Curve

JULY 3, 2023

These numbers should not be taken as a benchmark for your own site. Amazon is a good example of a site that serves large, fast pages, as you can see in this recent look at our Retail Page Speed Benchmarks : In the waterfall chart for this test run , you can see why the Amazon home page ranks fastest for Start Render.

Mobile 52

Mobile Google Metrics Retail 52

Brendan Gregg

SEPTEMBER 25, 2021

As a Xen guest, this profile was gathered using perf(1) and the kernel's software cpu-clock soft interrupts, not the hardware NMI. There's also a test and println() in the loop to, hopefully, convince the compiler not to optimize-out an otherwise empty loop. This will slow this test a little.) But I'm not completely sure.

Speed 40

Speed Java AWS Virtualization 40

HammerDB

OCTOBER 12, 2018

This post complements the previous best practice guides this time with the focus on MySQL and MariaDB and achieving top levels of performance with the HammerDB MySQL TPC-C test. As is also the case this limitation is at the database level (especially the storage engine) rather than the hardware level. hardware limits: 1000 MHz - 3.80

Best Practices 40

Best Practices Scalability Performance C++ 40

John McCalpin

JANUARY 22, 2018

After almost four weeks, we have a solid characterization of the behavior — no tested code exceeds an execution rate of 12 vector pipe instructions every 7 cycles (6/7 of the nominal peak) when executed on a single core — but we are unable to propose a testable quantitative model for the source of the throughput limitation.

Latency 40

Latency Hardware Code Testing 40

Alex Russell

OCTOBER 22, 2017

Budgets are scaled to a benchmark network & device. Automating testing against an objective baseline. Deciding what benchmark to use for a performance budget is crucial. Simulated packet loss and variable latency, however, can make benchmarking extremely difficult and slow. What’s going on here?

Performance 68

Performance Benchmarking Network Mobile 68