3 ms·

Here's a pretty interesting note on .NET Core 1.1's performance improvements from round 13 of Techempower's framework benchmarks [1] (linked from the announceme

by brandur 10y ago

Here's a pretty interesting note on .NET Core 1.1's performance improvements from round 13 of Techempower's framework benchmarks [1] (linked from the announcement above): > Thanks to Microsoft’s herculean performance tuning effort, ASP.NET—in the new cross-platform friendly form of ASP.NET Core—is now a top performer in our Plaintext test, making it among the fastest platforms at the fundamentals of web request routing. The degree of improvement is absolutely astonishing, going from 2,120 requests per second on Mono in Round 11 to 1,822,366 requests per second on ASP.NET Core in Round 13. That’s an approximately 85,900% improvement, and that doesn’t even account for Round 11’s hardware being faster than our new hardware. Here are the benchmarks in question [2]. There are a few caveats attached, like that it's still not quite at the top and the best improvement is under the plaintext benchmark in particular, but nonetheless a very impressive effort on the part of the .NET team. [1] https://www.techempower.com/blog/2016/11/16/framework-benchmarks-round-13/ https://www.techempower.com/blog/2016/11/16/framework-benchm... [2] https://www.techempower.com/benchmarks/#section=data-r13&hw=ph&test=plaintext https://www.techempower.com/benchmarks/#section=data-r13&hw=...

0xmohit 10y ago
TechEmpower also reports that the performance of ASP.NET Core running on Linux is approximately 760 times faster than it was one year ago. Since TechEmpower started measuring benchmarks in March 2013, they have never seen such a performance improvement as they have observed in ASP.NET Core over the last year.
- arenaninja 10y ago
  The pessimist's take: there was a lot to improve. My exposure to C# and .NET is small enough that I'm not sure what this means in terms of PHP. Is this like a PHP4 to PHP5.3 or more like PHP5 to PHP7 change?
  - phaed 10y ago
    PHP to Hack
  - program_monkey 10y ago
    >The pessimist's take: there was a lot to improve. This is the rational take. The prior results were completely out of scale with competing solutions, running two magnitudes slower. Someone finally fired up a profiler and probably found some dramatic issues, fixed them, and boom people who relied upon the project before, it turns out, were having a grossly under-delivering result. And at the new 859x faster speed, it's still 1/2 the speed of a competitor. So I'm not sure what the take from it is. It is held as a positive, but to me it sounds more like it's a token project.
radicalbyte 10y ago
The most interesting thing to me is the difference between the results from a year ago. The top scores are 1/3rd of what they were a year ago. They changed the test rig from an i7-2600K to a Dual Xeon ES-2660v2. Those 4 cores at 3.4ghz destroy the 20 cores at 2.2ghz. Cloud performance in comparison is terrible. That i7 hits 290k fortunes compared to the cloud VM hitting 30k. You need ~$1500 of VM a month ($0.23 / hour / 30 days / 9) to match a desktop computer from 2011.
- bhauer 10y ago
  After Round 12, we had to find a new physical hardware environment. ServerCentral [1] stepped up and provided some (older) physical hardware for the project. However, the new hardware doesn't match the performance of previous rounds. The physical hardware for each round has been [2]: * Rounds 1 through 8: i7-2600K workstations * Round 9 through 12: 2x Xeon E5-2660v2 * Round 13: 4x Xeon E7-4850 [1] https://www.servercentral.com/ https://www.servercentral.com/ [2] https://www.techempower.com/benchmarks/#section=environment https://www.techempower.com/benchmarks/#section=environment
mythz 10y ago
They're not exactly performance improvements as .NET Core doesn't share any implementation ancestry with Mono (the runtime that was previously benchmarked). .NET Core's just a better engineered and optimized runtime for running Server Apps on Windows, Linux and OSX. Basically .NET Core is now the future for running .NET Server Apps on Linux.

by brandur 10y ago

0xmohit 10y ago

TechEmpower also reports that the performance of ASP.NET Core running on Linux is approximately 760 times faster than it was one year ago. Since TechEmpower started measuring benchmarks in March 2013, they have never seen such a performance improvement as they have observed in ASP.NET Core over the last year.

arenaninja 10y ago
The pessimist's take: there was a lot to improve. My exposure to C# and .NET is small enough that I'm not sure what this means in terms of PHP. Is this like a PHP4 to PHP5.3 or more like PHP5 to PHP7 change?
- phaed 10y ago
  PHP to Hack
- program_monkey 10y ago
  >The pessimist's take: there was a lot to improve. This is the rational take. The prior results were completely out of scale with competing solutions, running two magnitudes slower. Someone finally fired up a profiler and probably found some dramatic issues, fixed them, and boom people who relied upon the project before, it turns out, were having a grossly under-delivering result. And at the new 859x faster speed, it's still 1/2 the speed of a competitor. So I'm not sure what the take from it is. It is held as a positive, but to me it sounds more like it's a token project.

radicalbyte 10y ago

The most interesting thing to me is the difference between the results from a year ago. The top scores are 1/3rd of what they were a year ago. They changed the test rig from an i7-2600K to a Dual Xeon ES-2660v2. Those 4 cores at 3.4ghz destroy the 20 cores at 2.2ghz. Cloud performance in comparison is terrible. That i7 hits 290k fortunes compared to the cloud VM hitting 30k. You need ~$1500 of VM a month ($0.23 / hour / 30 days / 9) to match a desktop computer from 2011.

bhauer 10y ago
After Round 12, we had to find a new physical hardware environment. ServerCentral [1] stepped up and provided some (older) physical hardware for the project. However, the new hardware doesn't match the performance of previous rounds. The physical hardware for each round has been [2]: * Rounds 1 through 8: i7-2600K workstations * Round 9 through 12: 2x Xeon E5-2660v2 * Round 13: 4x Xeon E7-4850 [1] https://www.servercentral.com/ https://www.servercentral.com/ [2] https://www.techempower.com/benchmarks/#section=environment https://www.techempower.com/benchmarks/#section=environment

mythz 10y ago

They're not exactly performance improvements as .NET Core doesn't share any implementation ancestry with Mono (the runtime that was previously benchmarked). .NET Core's just a better engineered and optimized runtime for running Server Apps on Windows, Linux and OSX. Basically .NET Core is now the future for running .NET Server Apps on Linux.