article thumbnail

The Return of the Frame Pointers

Brendan Gregg

I'm sure this delivered large performance improvements and I wouldn't try arguing against it. 2005-2023: The winter of broken profilers However, the change was then applied to x86-64 (64-bit) as well, which had sixteen registers and didn't benefit so much from a seventeenth.

Java 145
article thumbnail

Simple Parameterization and Trivial Plans — Part 2

SQL Performance

Simple parameterization has a number of quirks in this area, which can result in more parameterized plans being cached than expected, or finding different results compared with the unparameterized version. Reputation = 252552 ; These statements result in six cached plans, three Adhoc and three Prepared : Different guessed types.

Cache 90
Insiders

Sign Up for our Newsletter

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

Trending Sources

article thumbnail

Simple Parameterization and Trivial Plans — Part 1

SQL Performance

Both target performance and efficiency for workloads frequently submitting simple statements. In this first part, after a quick introduction, I look at the effects of simple parameterization on the plan cache. The aim is to reduce compilations by increasing cached plan reuse. Simple Parameterization. Shell Plans. sp_configure.

Cache 61
article thumbnail

The Surprising Effectiveness of Non-Overlapping, Sensitivity-Based Performance Models

John McCalpin

This was a keynote presentation at the “2nd International Workshop on Performance Modeling: Methods and Applications” (PMMA16), June 23, 2016, Frankfurt, Germany (in conjunction with ISC16 ). Here I assumed a particular analytical function for the amount of memory traffic as a function of cache size to scale the bandwidth time.

article thumbnail

How Parallel Plans Start Up – Part 1

SQL Performance

It includes initializing each operator, and adding hidden profilers to collect runtime performance data like actual row count and elapsed time. The fundamentals of row mode parallel execution haven’t changed since SQL Server 2005, so the following discussion is broadly applicable. Compatibility level 150 is used exclusively.

Cache 98
article thumbnail

The Amazing Evolution of In-Memory Computing

ScaleOut Software

From Distributed Caches to Real-Time Digital Twins. Emerging in the early 2000s, the first such platforms provided distributed caching on clustered servers with straightforward APIs for storing and retrieving in-memory objects.

article thumbnail

The Amazing Evolution of In-Memory Computing

ScaleOut Software

From Distributed Caches to Real-Time Digital Twins. Emerging in the early 2000s, the first such platforms provided distributed caching on clustered servers with straightforward APIs for storing and retrieving in-memory objects.