On Fri, Apr 24, 2015 at 10:35:49AM -0400, Waiman Long wrote:
On 04/23/2015 05:23 AM, Mel Gorman wrote:I do not believe that is what he was asking. He was asking if we could
On Wed, Apr 22, 2015 at 04:45:00PM -0700, Andrew Morton wrote:Memory access latency can be more than double for local vs. remote
On Wed, 22 Apr 2015 18:07:50 +0100 Mel Gorman<mgorman@xxxxxxx> wrote:Done.
--- a/arch/x86/KconfigPut this in the "config X86_64" section and skip the "X86_64&&"?
@@ -32,6 +32,7 @@ config X86
select ARCH_SUPPORTS_NUMA_BALANCING if X86_64
select ARCH_SUPPORTS_INT128 if X86_64
+ select ARCH_SUPPORTS_DEFERRED_STRUCT_PAGE_INIT if X86_64&& NUMA
Can we omit the whole defer_meminit= thing and permanently enable theYes. The intent was to have a workaround if there were any failures like
feature? That's simpler, provides better test coverage and is, we
Waiman's vmalloc failures in an earlier version but they are bugs that
should be fixed.
And can this be used on non-NUMA? Presumably that won't speed thingsNothing prevents it. There is less opportunity for parallelism but
up any if we're bandwidth limited but again it's simpler and provides
improving coverage is desirable.
node memory. Bandwidth can also be much lower depending on what kind
of interconnect is between the 2 nodes. So it is better to do it in
a NUMA-aware way.
defer memory initialisation even when there is only one node. It does not
gain much in terms of boot times but it improves testing coverage.
Within a NUMA node, however, we can split theI considered it but discarded the idea. It'd be more complex to setup and
memory initialization to 2 or more local CPUs if the memory size is
the two CPUs could simply end up contending on the same memory bus as
well as contending on zone->lock.