Re: [PATCH v2 1/4] gdbsupport: use dynamic partitioning in gdb::parallel_for_each

[Tom Tromey <tom@tromey.com>]

>>>>> "Simon" == simon marchi <simon.marchi@polymtl.ca> writes:

Simon> From: Simon Marchi <simon.marchi@polymtl.ca>
Simon> gdb::parallel_for_each uses static partitioning of the workload, meaning
Simon> that each worker thread receives a similar number of work items.  Change
Simon> it to use dynamic partitioning, where worker threads pull work items
Simon> from a shared work queue when they need to.

Finally getting back to this.

Simon> Note that gdb::parallel_for_each is currently only used for processing
Simon> minimal symbols in GDB.  I am looking at improving the startup
Simon> performance of GDB, where the minimal symbol process is one step.

I do wonder if this functionality should just be intrinsic to
thread-pool somehow, like separate the notions of a "task" and a "job
queue".

But anyway I think we should move forward with what exists.

Simon> +  auto task =
Simon> +    [&next, first, last, n_worker_threads, &args_tuple] () {

I think the "{" should go on the next line.

Approved-By: Tom Tromey <tom@tromey.com>

Tom