Brendan Gregg

FEBRUARY 28, 2023

My personal opinion is that I don't see a widespread need for more capacity given horizontal scaling and servers that can already exceed 1 Tbyte of DRAM; bandwidth is also helpful, but I'd be concerned about the increased latency for adding a hop to more memory.

Performance 110

Performance Latency Cache Virtualization 110

SQL Performance

JULY 16, 2019

A close monitoring of the hardware enthusiast community, including many of the most respected hardware analysts and reviewers paints an even more dire picture about Intel in the server processor space. Despite all of this, Intel is not going to lose their entire server processor business any time soon. So, what has changed my mind?

Servers 46

Servers Hardware Architecture Retail 46

Brendan Gregg

JULY 2, 2021

cachestat File system cache statistics line charts. The architecture is: While the bpftrace binary is installed on all the target systems, the bpftrace tools (text files) live on a web server and are pushed out when needed. execsnoop New processes (via exec(2)) table. opensnoop Files opened table. at the time.

Latency 145

Latency Cache Energy Systems 145

Brendan Gregg

FEBRUARY 28, 2023

My personal opinion is that I don't see a widespread need for more capacity given horizontal scaling and servers that can already exceed 1 Tbyte of DRAM; bandwidth is also helpful, but I'd be concerned about the increased latency for adding a hop to more memory.

Performance 52

Performance Latency Cache Virtualization 52

SQL Performance

AUGUST 23, 2019

In traditional row-mode execution plans, SQL Server may introduce a Bitmap operator as part of performing early semi join reduction before a parallel hash or merge join. There have been major improvements since the first appearance of the batch mode execution engine in SQL Server 2012. Background. The Query Optimizer. Bitmap Choice.

Servers 75

Servers Engineering Efficiency Retail 75

Alex Russell

MARCH 6, 2021

A then-representative $200USD device had 4-8 slow (in-order, low-cache) cores, ~2GiB of RAM, and relatively slow MLC NAND flash storage. mid-priced Androids were slightly faster than 2014's iPhone 6. The Moto G4 , for example. The cheapest (high volume) Androids perform like 2012/2013 iPhones, respectively.

Performance 124

Performance Network Cache Metrics 124

ACM Sigarch

MARCH 25, 2019

Figure 1 depicts a simple queue where tasks arrive, are optionally queued in the buffers to the left if the server is already busy, are then serviced one at a time in the circular server, and depart to the right. For the previous cache miss buffer example, the 32-buffer answer is minimal for 100-ns average miss latency.

Systems 53

Systems Latency Performance C++ 53