[Bug 259249] vtnet: Checksum offloading enabled reduces pf NAT performance (~100mbps -> ~200kbps)

[email protected]
Newsgroups gmane.os.freebsd.devel.net
Message-ID <[email protected]/bugzilla/>
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=259249

[email protected] changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[email protected]

--- Comment #4 from [email protected] ---
As fast as as I have testing stand (see #276760), I've re-measured in NATed
jail 

and here clear degradation of download for 14.3p4 with offloading if compare
with 
14-stable or no offloading 

~26 Mbps vs ~800 Mbps normally


FreeBSD Offloading      Download (CV%)  delta %         Upload (CV%)    delta %
14.3p4          yes     26      19.96%                   445    19.96%  
14.3p4          no      813     23.93%  3079%            528    23.93%  19%
14-087fc6ae0a8c yes     489     58.95%  -40%             431    58.95%  -18%
14-087fc6ae0a8c no      762     33.78%  56%              570    33.78%  32%


raw data:
FreeBSD Offloading      Try     Download        Upload

14.3p4          yes     1        21      54 
14.3p4          yes     2        18      769 
14.3p4          yes     3        20      315 
14.3p4          yes     4        28      540 
14.3p4          yes     5        33      495 
14.3p4          yes     6        27      356 
14.3p4          yes     7        32      421 
14.3p4          yes     8        22      325 
14.3p4          yes     9        28      451 
14.3p4          yes     10       27      728 
14.3p4          yes     avg      26      445 
                CV%     19.96%  46.73%

14.3p4          no      1        932     686 
14.3p4          no      2        749     433 
14.3p4          no      3        854     441 
14.3p4          no      4        885     624 
14.3p4          no      5        946     760 
14.3p4          no      6        284     211 
14.3p4          no      7        916     754 
14.3p4          no      8        863     462 
14.3p4          no      9        883     416 
14.3p4          no      10       816     496 
14.3p4          no      avg      813     528 
                CV%     23.93%  33.02%

14-087fc6ae0a8c yes     1        87      510 
14-087fc6ae0a8c yes     2        537     763 
14-087fc6ae0a8c yes     3        884     576 
14-087fc6ae0a8c yes     4        714     414 
14-087fc6ae0a8c yes     5        379     392 
14-087fc6ae0a8c yes     6        724     441 
14-087fc6ae0a8c yes     7        259     563 
14-087fc6ae0a8c yes     8        758     474 
14-087fc6ae0a8c yes     9        60      48 
14-087fc6ae0a8c yes     10       484     128 
14-087fc6ae0a8c yes     avg      489     431 
                CV%     58.95%  48.74%

14-087fc6ae0a8c no      1        964     541 
14-087fc6ae0a8c no      2        943     410 
14-087fc6ae0a8c no      3        980     478 
14-087fc6ae0a8c no      4        981     761 
14-087fc6ae0a8c no      5        861     757 
14-087fc6ae0a8c no      6        874     537 
14-087fc6ae0a8c no      7        765     471 
14-087fc6ae0a8c no      8        596     510 
14-087fc6ae0a8c no      9        327     511 
14-087fc6ae0a8c no      10       328     728 
14-087fc6ae0a8c no      avg      762     570 
                CV%     33.78%  22.59%

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are on the CC list for the bug.
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.