Brendan Gregg

APRIL 14, 2022

offer letter logo (2014) flame graphs (2014) eBPF tools (2014-2019) PMC analysis (2017) my pandemic-abandoned desk (2020); office wall I joined Netflix in 2014, a company at the forefront of cloud computing with an attractive work culture. It was the most challenging job among those I interviewed for.

Java 52

Java Cache Cloud AWS 52

Alex Russell

APRIL 29, 2021

Delayed three years ( Chrome 40, November 2014 vs. Safari 11.1, A subset ( element.animate() ) has enabled developers to more easily create high-performance visual effects with lower risk of visual stuttering in Chrome and Firefox since 2014. Critical in adapting web content to mobile, particularly regarding multi-touch gestures.

Media 145

Media Games Education Engineering 145

Abhishek Tiwari

NOVEMBER 3, 2018

They often get blindsided by vendor’s pitch and end-up making decision based on some fancy demos (see my post from 2014 on Adobe AEM ). Circa 2014, I was working with a big Japanese automotive brand in Australia. However, it is important to note that JAMstack can serve both static as well as dynamic content.

Systems 63

Systems Cache Website Network 63

Nick Desaulniers

JUNE 2, 2018

ccache , the compiler cache, is a fantastic way to speed up build times for C and C++ code that I previously recommended. Usually when this happens with ccache, there’s something non-deterministic about the builds that prevents cache hits. Cold Cache. kB max cache size 5.0 kB max cache size 5.0 Ubuntu SMP.

Speed 46

Speed Cache C++ Systems 46

O'Reilly Software

JUNE 1, 2018

ccache, the compiler cache, is a fantastic way to speed up build times for C and C++ code that I previously recommended. Usually when this happens with ccache, there’s something non-deterministic about the builds that prevents cache hits.

Speed 40

Speed Cache C++ Code 40

Speed Curve

JUNE 14, 2021

To overcome some of these challenges, we built our own Java-based player, complete with caching, content compression, and even bandwidth detection so it could switch between video, audio, and text versions of a course depending on network speed. Ultimately the business didn’t survive the dotcom bust, but it lit a spark.

Retail 52

Retail Internet Internet Education 52

Brendan Gregg

FEBRUARY 28, 2023

Titus, the Netflix container management platform, is now open source,” [link] , Apr 2018 [Cutress 19] Dr. DDR6: Here's What to Expect in RAM Modules,” [link] , Nov 2020 [Salter 20] Jim Salter, “Western Digital releases new 18TB, 20TB EAMR drives,” [link] , Jul 2020 [Spier 20] Martin Spier, Brendan Gregg, et al.,

Performance 52

Performance Latency Cache Virtualization 52

The Morning Paper

JANUARY 23, 2020

Today we’ll be digging into the analysis of an incident that took place at Etsy on December 4th, 2014. This is part 2 of our look at Allspaw’s 2015 master thesis (here’s part 1 ). 1:06pm reports of the personalised homepage having issues start appearing from multiple sources.

Internet 53

Internet Internet Cache Engineering 53

SQL Server According to Bob

JUNE 3, 2016

In the Sep 2014 the SQL Server CSS and Development teams performed a deep dive focused on scalability and performance when running on current and new hardware configurations. “SQL Server 2016 running on the same hardware as SQL Server 2014, 2012, 2008, 2008 R2 or 2005 uses fewer resources and executes a wide range of workloads faster.

Hardware 40

Hardware Software Engineering Transportation Scalability 40

SQL Server According to Bob

JUNE 3, 2016

Use SQL Server Management Studio (SSMS) or your favorite query editor to connect to a SQL Server 2012 or 2014 instance. Note: You may need to execute the dbcc a second time so buffer cache is hot, eliminating I/O sub-system variants. SQL Server 2014. Paste the script below in a new query window. or newer release. SQL Server.

Hardware 40

Hardware Servers Cache Software Engineering 40

ACM Sigarch

MARCH 18, 2019

For example, how many buffers must a cache have to record outstanding misses if it receives 2 memory references per cycle at 2.5 He is a fellow of IEEE and the ACM, serves as Chair of the Computer Community Consortium (2018-20) and served as Wisconsin Computer Sciences Department Chair 2014-17. Figure 2: (Sub-)System for Little’s Law.

Systems 60

Systems Latency Performance Analytics 60

SQL Performance

APRIL 8, 2019

The L3 cache size is 64MB. The older Intel Xeon E5-26xx v3 (Haswell) series which was introduced in Q3 of 2014, had a maximum memory bandwidth of 2133MHz. The L3 cache size is 64MB. On the other hand, Azure VM CPU and storage performance has come a long way since I wrote about it back in 2014 ! lanes for I/O connectivity.

Azure 42

Azure Virtualization Storage Benchmarking 42

Brendan Gregg

AUGUST 29, 2021

Reads usually have apps waiting on them; writes may not (write-back caching). This is a 64-Gbyte memory system, and 48 Gbytes is in the page cache (the file system cache). Along with the numbers from the problem statement, this gives me a theory: Do the 100-Gbyte files bust the page cache, whereas 40-Gbyte files fit?

Cache 40

Cache C++ AWS Systems 40

SQL Performance

MAY 29, 2019

It can be activated from SQL Server 2008 to 2014 inclusive using documented trace flag 610. GO. -- Clear the plan cache. GO. -- Clear the plan cache. GO. -- Clear the plan cache. Remember that DMLRequestSort is cached with the plan — it is not evaluated on each execution of a reused plan. NULL , -- Partition ID.

Cache 64

Cache Testing Servers Code 64

ACM Sigarch

MARCH 25, 2019

For the previous cache miss buffer example, the 32-buffer answer is minimal for 100-ns average miss latency. He is a fellow of IEEE and the ACM, serves as Chair of the Computer Community Consortium (2018-20) and served as Wisconsin Computer Sciences Department Chair 2014-17. The M/M/1 design seems to over-designed by 2x (100ms ?

Systems 53

Systems Latency Performance C++ 53

SQL Performance

JUNE 21, 2019

For SQL Server 2008R – 2014, you could lower this threshold using trace flag 2371. If you are on SQL Server 2008 – SQL Server 2014, for a single NUMA node with less than 8 logical processors, keep MAXDOP at or below the number of logical processors. That can take a long time on larger tables.

Servers 49

Servers Azure Hardware Database 49

CSS Wizardry

MARCH 3, 2019

To this end, having a solid caching strategy can make all the difference for your visitors. ?? How is your knowledge of caching and Cache-Control headers? That being said, more and more often in my work I see lots of opportunities being left on the table through unconsidered or even completely overlooked caching practices.

Cache 266

Cache Latency Strategy Servers 266

Brendan Gregg

JULY 2, 2021

cachestat File system cache statistics line charts 7. As an extreme example, I wrote cachestat(8) while on vacation in 2014 for use on the Netflix cloud, which was a mix of Linux 3.2 execsnoop New processes (via exec(2)) table 2. opensnoop Files opened table 3. ext4slower Slow filesystem I/O table 4. It depends on the tool.

Open Source 40

Open Source Latency Cache Energy 40

SQL Performance

JULY 16, 2019

This situation was so bad that Microsoft offered a 25% discount on the cost of SQL Server processor core licenses for SQL Server 2012 and SQL Server 2014, if you ran on qualifying AMD Opteron processors with six or more cores. They will also have up to 256MB of L3 cache per processor.

Servers 46

Servers Hardware Architecture Retail 46

SQL Performance

AUGUST 23, 2019

The SQL Server batch mode implementation of a Bloom filter is optimized for modern CPU cache architectures and is known internally as a complex bitmap. Complex bitmaps have supported multiple join columns and all batch mode data types since SQL Server 2014. Bitmap Choice.

Servers 75

Servers Engineering Efficiency Retail 75

SQL Performance

MARCH 8, 2021

SQL Server 2014 SP2 and later produce runtime (“actual”) execution plans that can include elapsed time and CPU usage for each execution plan operator (see KB3170113 and this blog post by Pedro Lopes). Total elapsed time is 933ms with 6,673ms of CPU time with a warm cache.

Servers 95

Servers Tuning Architecture Performance 95

SQL Server According to Bob

DECEMBER 31, 2018

iBytesRead = s.Read(bData, 0, c_iMaxReadSize); } }. . // Do we need to start the read-ahead to suck data into file system cache if (iCookie > -1) { tReadAhead.Start(iCookie); } }. MergeXEvents consumes XEvent files, using optimized parallel sort algorithms and outputting events into the physical file in event_sequence order.

Efficiency 40

Efficiency Code Servers Cache 40

Percona

JUNE 4, 2023

GA 31 March 2014 10.0.10 MyISAM is modified with segmented key caches, the default temporary table storage engine is Aria (which is a crash-safe MyISAM), the FederatedX engine is a modified FEDERATED engine, and there are more: CONNECT, Mroonga, OQGRAPH, Sequence, SphinxSE, SPIDER, TokuDB and of course MyRocks. GA 28 April 2011 5.5.11-20.2

Open Source 52

Open Source Servers Storage Scalability 52