NEVER ROPEN murdock: nice compile for better test performance #17918
Conversation
github-actions
bot
added
the
Area: CI
kfessel
added
the
CI: ready for build
|
let's see if that has any visible effect on total build times. (the one error so far seems unrelated) |
kfessel
requested review from
smlng,
leandrolanzieri,
aabadie,
MichelRottleuthner and
fjmolinas
as code owners
| @@ -8,6 +8,9 @@ USEMODULE += base64 | |||
| # Use a terminal that does not introduce extra characters into the stream. | |||
| RIOT_TERMINAL ?= socat | |||
|
|
|||
| #ensure the rx buffer has some room even with large test patterns | |||
| CFLAGS += -DSTDIO_UART_RX_BUFSIZE=128 | |||
There was a problem hiding this comment.
(some boards have issues with larger uart transfers, like, the samr21-xpro over it's edbg)
There was a problem hiding this comment.
But that isn't an issue of the stdio-buffer-size being to big --i think
There was a problem hiding this comment.
In the pbkdf2-test are transfers of upto 67 byte (65+2) only limited by the speed of the serial connection.
The default rx buffer is 64 bytes -> if the uart isr keep coming in and hence the bytes not processed, it is getting tight in that buffer.
|
Checking mudock: |
|
@kaspar030 this seem to neither help nor hurt the cpu runtime another PR 17920 https://ci.riot-os.org/RIOT-OS/RIOT/17920/f5e626841800c0af22bf6bbe82d549ff615e8434/output.html total cpu runtime: 7d:3h:51m:13.0s this PR https://ci.riot-os.org/RIOT-OS/RIOT/17918/9bdd7043721e7a0a5004a8fefb2f325e5e6a3937/output.html total cpu runtime: 7d:2h:24m:10.2s they happen at a similar time of day sadly the output shows that this isn't a complete solution as 2 test did fail (maybe increased load due to change of day?) |
|
hm, honestly I would conclude that this doesn't solve the issue of the emulation being inaccurate timing wise. :( |
|
Wo how to still run the test but consider them as warnings? |
|
keep this closed its pointing to the wrong branch |
kfessel
changed the title
NEVER REOPEN - wrong branch
Contribution description
Title
Testing procedure
Issues/PRs references
addresses #17894