On 03/03/25 4:04 pm, Brendan Jackman wrote:
On Mon, Mar 03, 2025 at 03:48:38PM +0530, Dev Jain wrote:
On 03/03/25 3:17 pm, Brendan Jackman wrote:
On Fri, Feb 28, 2025 at 11:06:35PM +0530, Dev Jain wrote:
Taking a cursory look at the test, it creates three threads for each cpu.
The bounding of the variable is fine but that being the reason to rename the
variable is not making sense to me.
Hmm yeah the name needs to be more abstract. Do you think nr_workers
would be confusing? Or even just "parallelism" or nr_parallel? Or any
other ideas?
FWIW I briefly looked at just cleaning this up to remove the global
variable but that's a bigger time investment than I can afford here I
think. (The local variable in stress() would still need a better name
anyway).
Thanks for the review BTW!
Your welcome.
I personally prefer leaving it as is; unless someone comes up and completely
cleans up the structure, let us save our collective brain cycles for more
meaningful battles than renaming variables :)
Hmm, I think that's a false economy on brain cycles. A variable called
nr_cpus that isn't a number of CPUs is bound to waste a bunch of
mental energy at some point in the future.
Unless you strongly object I'll go for nr_parallel. It's not a great
name but, well... I think that probably just suggests it's not a great
variable, and I don't have time to fix that.
nr_parallel sounds better for sure. In case you send out a new patch:
Reviewed-by: Dev Jain <dev.jain@xxxxxxx>
