Changes between Version 2 and Version 3 of Performance/Sep2011


Ignore:
Timestamp:
2011-09-26T05:01:56Z (9 years ago)
Author:
warner
Comment:

oops, all MDMF test files were k=3. Need proper results.

Legend:

Unmodified
Added
Removed
Modified
  • Performance/Sep2011

    v2 v3  
    2323== MDMF (trunk) ==
    2424
    25 MDMF is fast! Trunk downloads 1MB/10MB/100MB MDMF files at around 4MBps. The
    26 download speed seems unaffected by k (from 1 to 60). Partial reads take the
    27 expected amount of time: O(data_read), slightly quantized near the 128KiB
    28 segment size.
     25MDMF is fast! Trunk downloads 1MB/10MB/100MB MDMF files at around 4MBps.
     26~~The download speed seems unaffected by k (from 1 to 60)~~. Partial
     27reads take the expected amount of time: O(data_read), slightly quantized
     28near the 128KiB segment size.
     29
     30(update 25-Sep: the test harness appears to have been flawed, and all
     31MDMF files were actually k=3. A new test is in the works to get proper
     32data on how MDMF read speed varies with k).
    2933
    3034* MDMF read versus k, 100MB attachment:MDMF-100MB-vs-k.png (timing5.out)
     
    7478* use unencrypted HTTP for share reads
    7579
    76 readv() is the least-work/most-promising, since MDMF has readv() and achieves
    77 high speeds. First step is to do whatever MDMF is doing
     80readv() is the least-work/most-promising, ~~since MDMF has readv() and
     81achieves high speeds. First step is to do whatever MDMF is doing~~
    7882
    7983== Future Tests ==