From c825214aba30672e992751d3ed36d79555e85c48 Mon Sep 17 00:00:00 2001 From: FritzFlorian Date: Fri, 3 May 2019 15:50:56 +0200 Subject: [PATCH] Add further notes on performance development. --- PERFORMANCE.md | 10 +++++++++- README.md | 2 ++ media/aa27064_fft_average.png | Bin 0 -> 175646 bytes media/d16ad3e_fft_average.png | Bin 0 -> 190756 bytes 4 files changed, 11 insertions(+), 1 deletion(-) create mode 100644 media/aa27064_fft_average.png create mode 100644 media/d16ad3e_fft_average.png diff --git a/PERFORMANCE.md b/PERFORMANCE.md index a57a66d..58397f1 100644 --- a/PERFORMANCE.md +++ b/PERFORMANCE.md @@ -28,7 +28,15 @@ Big improvements of about 6% in our test. This seems like a little, but 6% from the scheduler is a lot, as the 'main work' is the tasks itself, not the scheduler. -### Commit 18b2d744 - Performance problems with higher thread counts +### Commit aa27064 - Performance with ttsa spinlocks (and 'full blocking' top level) + + + +### Commit d16ad3e - Performance with rw-lock and backoff + + + +### Commit 18b2d744 - Performance with lock-free deque After much tinkering we still have performance problems with higher thread counts in the FFT benchmark. Upward from 4/5 threads the diff --git a/README.md b/README.md index 8ec38b4..c42781f 100644 --- a/README.md +++ b/README.md @@ -7,6 +7,8 @@ This section will give a brief introduction on how to get a minimal project setup that uses the PLS library. +Further [general notes](NOTES.md) and [performance notes](PERFORMANCE.md) can be found in +their respective files. ### Installation diff --git a/media/aa27064_fft_average.png b/media/aa27064_fft_average.png new file mode 100644 index 0000000..3441678 Binary files /dev/null and b/media/aa27064_fft_average.png differ diff --git a/media/d16ad3e_fft_average.png b/media/d16ad3e_fft_average.png new file mode 100644 index 0000000..78730f9 Binary files /dev/null and b/media/d16ad3e_fft_average.png differ -- libgit2 0.26.0