On Sun, 17 Feb 2013, Jim Klimov wrote:
I guess I'm reading it wrong, but doesn't "pi" mean "page-in" for pages fetched from on-disk swap space? With so much available RAM, why would it even be used?
No, I don't think that this is quite what it means. Page-in means that a VM page was added to the process address space. There are many types of page-faults associated with "page-in", and not just those which create a new page based on a memory mapped file, or a dirty page from swap.
Also, 4k*180561 = 722244Kb paged in, and 2048k*1207 = 2471936Kb paged in, so in the second case the system actually does 3x more IO (even if just some RAM ops) to page-in the data. Perhaps some value in-between should give you better results in the particular test run (smaller pages and requests amounting to 800mb-1gb, and fewer requests having less overhead and RAM fragmentation).
Actually, hardly any I/O is being done at all. The larger pages mean that more virtual memory is given to the program per request. A highly compressed file is being "exploded" into quite a lot of RAM.
This system has 128GB of RAM and the program is 64-bit. All of the disks are completely idle (as reported by iostat) while the program is running.
And you do skip the first few lines of vmstat output, right? (that should show lifetime stats average during this OS uptime). Trust it for dynamic data only after the second or third line.
Yes, I do skip those lines. Bob -- Bob Friesenhahn [email protected], http://www.simplesystems.org/users/bfriesen/ GraphicsMagick Maintainer, http://www.GraphicsMagick.org/ ------------------------------------------- illumos-discuss Archives: https://www.listbox.com/member/archive/182180/=now RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175430-2e6923be Modify Your Subscription: https://www.listbox.com/member/?member_id=21175430&id_secret=21175430-6a77cda4 Powered by Listbox: http://www.listbox.com
