Benchmarks

[STAstro]

Hi All

I wondered if everyone runs the benchmark at all?  Maybe we should all post our benchmark results and share any best practices?

*******************************************************************************
The Official PixInsight Benchmark version 1.0
Copyright (C) 2014-2015 Pleiades Astrophoto. All Rights Reserved.
*******************************************************************************

Benchmark version ...... 1.00.08
Input checksum ......... 2cd72b67e12fff2812ef5b5da054ab2a70a25e23
Serial number .......... M70B7J5Z775H5ACCBCOMZRGZTL1IWF2T

CPU Identification
CPU vendor ............. GenuineIntel
CPU model .............. Intel(R) Core(TM) i9-7900X CPU @ 3.30GHz

System Information
Platform ............... Windows
Operating system ....... Microsoft Windows 10 Pro
Core version ........... PixInsight Core 01.08.06.1448 (x64)
Logical processors ..... 20
Total memory size ...... 31.706 GiB

Execution Times
Total time ............. 00:55.99
CPU time ............... 00:38.20
Swap time .............. 00:17.75
Swap transfer rate ..... 933.753 MiB/s

Performance Indices
Total performance ...... 8401
CPU performance ........ 9909
Swap performance ....... 5172

*******************************************************************************

[msmythers]

Benchmarks are posted when you hit the submit button and reported here.

http://www.pixinsight.com/benchmark/



Mike

[STAstro]

I didn't know that

Anyway, one thing I have discovered is that PixInsight doesn't seem to use MultiThreads to disks for SSD Optimization, now I understand why...because not everyone has SSDs right!

I have done some bench marking and that's pretty clear, SSDs work optimally when you saturate the queues on the SSD, for example all SSD Vendors publish their performance results based on an optimal workload, usually 4K and an outstanding I/O of 16, I am using a FusionIO IO3 3.2TB PCIe adapter which are pretty much up there with the fastest Flash drives around, however the swap performance was only around 933MiBs/sec, so I have created 4 swap folders on the FusionIO Card and specified them in the Directories and Network settings in PixInsight and I am now close to 2000MiB/sec, I think I can push this futher as I am not even touching the sides on the FusionIO Card yet.

Would be interesting to know exactly how PI works at a workload level such as Block Size, Outstanding IO etc

[msmythers]

My main computer uses an old 6 core AMD processor that only uses one thread per core. I have 2 SSD's in the system. This is the swap configuration that works the best for me. You can hover the mouse over the directory configuration section for more information. There are a lot of older postings about swap configurations and the benchmark script also.


Mike