[lkml]   [2006]   [Jan]   [13]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    SubjectRe: [PATCH/RFC] Shared page tables
    On Thursday 12 January 2006 23:15, Brian Twichell wrote:

    > Hi,
    > We evaluated page table sharing on x86_64 and ppc64 setups, using a
    > database OLTP workload. In both cases, 4-way systems with 64 GB of memory
    > were used.
    > On the x86_64 setup, page table sharing provided a 25% increase in
    > performance,
    > when the database buffers were in small (4 KB) pages. In this case,
    > over 14 GB
    > of memory was freed, that had previously been taken up by page tables.
    > In the
    > case that the database buffers were in huge (2 MB) pages, page table
    > sharing provided a 4% increase in performance.


    Is that 25%-50% percent of overall performance (e. g. transaction throughput),
    or is this a measurement of, say, DB process startup times, or what? It
    seems to me that the impact of the shared page table patch would mostly be
    noticed at address space construction/destruction times, and for a big OLTP
    workload, the processes are probably built once and stay around forever, no?

    If the performance improvement is in overall throughput, do you understand why
    the impact would be so large? TLB reloads? I don't understand why one
    would see that kind of overall performance improvement, but I could be
    overlooking something. (Could very likely be overlooking something...:-) )

    Oh, and yeah, was this an AMD x86_64 box or what?
    Ray Bryant
    AMD Performance Labs Austin, Tx
    512-602-0038 (o) 512-507-7807 (c)

    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to
    More majordomo info at
    Please read the FAQ at

     \ /
      Last update: 2006-01-13 23:37    [W:0.022 / U:6.668 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site