Hi Paul,
Thanks for the testing. How many iterations of the test did you do? I ran mine 100 times with each config to get a reliable set of measurements.
Colin
-----Original Message-----
From: Paul Menzel <pmenzel@xxxxxxxxxxxxx>
Sent: Tuesday, August 13, 2024 11:15 AM
To: King, Colin <colin.king@xxxxxxxxx>
Cc: Rafael J . Wysocki <rafael@xxxxxxxxxx>; Len Brown <lenb@xxxxxxxxxx>; linux-acpi@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; Steven Rostedt <rostedt@xxxxxxxxxxx>
Subject: Re: [PATCH][next] ACPI: align slab for improved memory performance
Dear Colin,
Thank you for your patch and working on decreasing the boot time.
Am 09.08.24 um 00:21 schrieb Colin Ian King:
> Enabling SLAB_HWCACHE_ALIGN for the ACPI object caches improves boot
> speed in the ACPICA core for object allocation and free'ing especially
> in the AML parsing and execution phases in boot. Testing with 100
> boots shows an average boot saving in acpi_init of ~35000 usecs
> compared to the unaligned version. Most of the ACPI objects being
> allocated and free'd are of very short life times in the critical
> paths for parsing and execution, so the extra memory used for
> alignment isn't too onerous.
>
> Signed-off-by: Colin Ian King <colin.king@xxxxxxxxx>
> ---
> drivers/acpi/osl.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/acpi/osl.c b/drivers/acpi/osl.c index
> 70af3fbbebe5..dab3d5089635 100644
> --- a/drivers/acpi/osl.c
> +++ b/drivers/acpi/osl.c
> @@ -1549,7 +1549,7 @@ void acpi_os_release_lock(acpi_spinlock lockp, acpi_cpu_flags not_used)
> acpi_status
> acpi_os_create_cache(char *name, u16 size, u16 depth, acpi_cache_t **cache)
> {
> - *cache = kmem_cache_create(name, size, 0, 0, NULL);
> + *cache = kmem_cache_create(name, size, 0, SLAB_HWCACHE_ALIGN, NULL);
> if (*cache == NULL)
> return AE_ERROR;
> else
Thank you for the patch. I tested it on the Intel Kaby Lake laptop Dell XPS 13 9360, and unfortunately, my results are different from yours, that the time for `acpi_init()` *in*creased by 20 ms.
Without your patch:
[ 0.201471] initcall acpi_init+0x0/0x590 returned 0 after 112000
usecs
With your patch:
[ 0.222321] initcall acpi_init+0x0/0x590 returned 0 after 132000
usecs
Please find the logs attached.
Kind regards,
Paul
