RE: [PATCH 0/4] sched: Various reweight_entity() fixes

[Doug Smythies]

On 2026.02.11 23:52 Peter Zijlstra wrote:
> On Wed, Feb 11, 2026 at 09:54:40PM -0800, Doug Smythies wrote:
>
>> Same here, 6 hours and 20 minutes so far, and I'll leave it overnight:
>
> Excellent!
>
>> By the way, thank you for the cool parsing and histogram command line.
>
> No problem. I love awk :-) Its just small enough to re-learn by reading
> the man page.
>
> I always struggle with things like perl and python because I don't use
> them enough to retain stuff, and then I'm left endlessly googling how
> the heck things worked again. But awk fits in a single 'short' manpage.

We already know all is good, but just for completeness after overnight:
14 hours and 46 minutes:

Cool method:

$ gawk '/^usec/ {next} { if (T) { d=$2-T; bucket[int(d*1000)]++; } T=$2 } END { for (i in bucket) { printf "%0.3f: %d\n", i/1000,
bucket[i] }}' < /dev/shm/turbo.log
1.000: 28858
1.001: 24052
1.002: 87
1.003: 56
1.004: 31
1.005: 11
1.006: 8
1.008: 2

My old way. Interesting differences, likely due to subtle differences in subtracting 2 16 digit numbers:

cat v3.his
1.000, 32974
1.001, 20012
1.002, 86
1.003, 57
1.004, 31
1.005, 11
1.006, 8
1.008, 2

Total: 53181 : Total >= 10 mSec: 0 ( 0.00 percent)