xpra icon
Bug tracker and wiki

Changes between Initial Version and Version 3 of Ticket #797


Ignore:
Timestamp:
01/28/15 01:52:15 (5 years ago)
Author:
Antoine Martin
Comment:

(just re-formatting with wiki tags)

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #797

    • Property Owner changed from Antoine Martin to Nick Centanni
    • Property Status changed from new to assigned
    • Property Milestone changed from to 0.14
  • Ticket #797 – Description

    initial v3  
    11We can take a look at other things to educate ourselves, any result (positive or negative) should tell us more about the trade-offs we have decided to make and whether we should re-evaluate them (my bet: there is bound to be some that need another look).
     2
    23Here are a few worth checking:
    3 0.14 vs 0.15: the obvious one first, we should be doing better - but are we?
    4 x264 tuning: speed and quality impact on bandwidth, framerate, etc
    5 x264: impact of enabling threading ("XPRA_X264_THREADS=1") - same for vpx and webp - related to the CPU tests below
    6 re-run of the detailed x264 vs vp8 comparison
    7 rgb lz4 vs png: are the bandwidth savings worth the CPU pain of using png?
    8 nvenc: kepler vs maxwell 1 vs maxwell2 cards (the new GTX 960 is out too..)
    9 the new multi-delta feature: what is the optimal number of delta buckets (currently set to 5)
    10 CPU tests: measure the impact of clock speed on performance (by artificially capping it), same for number of cores
    11 various environment variables that control aspects of the network layer:: XPRA_READ_BUFFER_SIZE, XPRA_PACKET_JOIN_SIZE, XPRA_INLINE_SIZE, XPRA_YIELD
    12 proxy tuning: PROXY_QUEUE_SIZE
    13 impact of the merge regions code: "XPRA_MERGE_REGIONS"
    14 impact of downscaling, by forcing it with XPRA_SCALING, maybe also forcing CSC with XPRA_FORCE_CSC_MODE
     4* 0.14 vs 0.15: the obvious one first, we should be doing better - but are we?
     5* x264 tuning: speed and quality impact on bandwidth, framerate, etc
     6* x264: impact of enabling threading ("{{{XPRA_X264_THREADS=1}}}") - same for vpx and webp - related to the CPU tests below
     7* re-run of the detailed x264 vs vp8 comparison
     8* rgb lz4 vs png: are the bandwidth savings worth the CPU pain of using png?
     9* nvenc: kepler vs maxwell 1 vs maxwell2 cards (the new GTX 960 is out too..)
     10* the new multi-delta feature: what is the optimal number of delta buckets (currently set to 5)
     11* CPU tests: measure the impact of clock speed on performance (by artificially capping it), same for number of cores
     12* various environment variables that control aspects of the network layer: {{{XPRA_READ_BUFFER_SIZE}}}, {{{XPRA_PACKET_JOIN_SIZE}}}, {{{XPRA_INLINE_SIZE}}}, {{{XPRA_YIELD}}}
     13* proxy tuning: {{{PROXY_QUEUE_SIZE}}}
     14* impact of the merge regions code: "{{{XPRA_MERGE_REGIONS}}}"
     15* impact of downscaling, by forcing it with {{{XPRA_SCALING}}}, maybe also forcing CSC with {{{XPRA_FORCE_CSC_MODE}}}
    1516etc...
     17
    1618The important thing about running those tests is to run them again and again, after either narrowing down on the outliers or discarding them when we understand why they stick out.