At least on x86, there's also a benefit in having smaller page tables. In particular, it makes the job of the TLB a lot easier for large memory systems.
On 03/12/2015 03:59 AM, Steven Harris wrote: > Hi Andrew > > The original reasoning on AIX was that as the data is transferred to > tape AIX gets a spin lock on each 4k page and then has to release it, > so that a system using tape heavily spends quite a lot of CPU > acquiring and releasing locks. > > Later versions of AIX have two larger page sizes available and so the > number of locks to be taken and released drops dramatically for the > typical 256KB tape transfer. > > I'm not sure whether this reasoning applies to Linux. If it does then > a larger page size might be beneficial there. > > Regards > > Steve > > Steven Harris > TSM Admin > Canberra Australia. > > On 12/03/2015 3:20 AM, Andrew Galloway wrote: >> We have 2 TSM 6.3.5 servers running on zLinux. Is this something >> we could leverage? >> >> >> >> >> >> __________________________ Andrew Galloway Operational Systems >> Continuity Analyst SSC RCMP Technology Directorate CIO Building >> C1035 R.C.M.P. Headquarters 1200 Vanier Parkway Ottawa ON K1A 0R2 >> 613.998.7707 >> andrew.gallo...@ssc-spc.gc.ca__________________________ >>>>> "Rhodes, Richard L." <rrho...@firstenergycorp.com> 2015-03-11 >>>>> 7:15 AM >>> >> I've been in meetings/training this week about AIX LPM and other >> stuff related with LPARS. We got talking about our TSM servers, >> and the IBM leader mentioned a presentation given a while ago at a >> IBM conference about using LARGE PAGES (16MB) with TSM to prevent >> lock contention. This is supposed to help if you drive some large >> number of tape drives concurrently. He mentioned one company that >> implemented this and saw a 20% increase in tape throughput. >> >> http://www-01.ibm.com/support/docview.wss?uid=swg21649823&aid=1 >> >> I'm curious if anyone out there has heard of this and possibly >> implemented large pages for TSM. >> >> Thanks! >> >> Rick >> >> >> >> >> >> >> ----------------------------------------- The information contained >> in this message is intended only for the personal and confidential >> use of the recipient(s) named above. If the reader of this message >> is not the intended recipient or an agent responsible for >> delivering it to the intended recipient, you are hereby notified >> that you have received this document in error and that any review, >> dissemination, distribution, or copying of this message is strictly >> prohibited. If you have received this communication in error, >> please notify us immediately, and delete the original message. >> -- -- Skylar Thompson (skyl...@u.washington.edu) -- Genome Sciences Department, System Administrator -- Foege Building S046, (206)-685-7354