Re: [RFC PATCH] locking/percpu-rwsem: use this_cpu_{inc|dec}() for read_count

[peterz]

On Wed, Sep 16, 2020 at 08:32:20PM +0800, Hou Tao wrote:

> I have simply test the performance impact on both x86 and aarch64.
> 
> There is no degradation under x86 (2 sockets, 18 core per sockets, 2 threads per core)

Yeah, x86 is magical here, it's the same single instruction for both ;-)
But it is, afaik, unique in this position, no other arch can pull that
off.

> However the performance degradation is huge under aarch64 (4 sockets, 24 core per sockets): nearly 60% lost.
> 
> v4.19.111
> no writer, reader cn                               | 24        | 48        | 72        | 96
> the rate of down_read/up_read per second           | 166129572 | 166064100 | 165963448 | 165203565
> the rate of down_read/up_read per second (patched) |  63863506 |  63842132 |  63757267 |  63514920

Teh hurt :/