In theory, getting 1 Storage Obtain for 1M should be less overhead than 256 4K versions. The virtual storage algorithms to get the data is supposed to be less (need less levels of hashing). So I would understand the potential. The conversion to 1M memory to 4K memory and vice versa, seems to be a bit painful; so you better size what you nearly actually need. I don't always mind the Pied Piper approach; but I'd always want metrics before and after. Ditto about HiperDispatch. The theory sounds good, but how is it measured? At least to be sure no bad things are being done to you.
-----Original Message----- From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of Gerhard Adam Sent: Sunday, August 15, 2010 Sunday 12:24 PM To: [email protected] Subject: Re: Anyone Using z10 Large Pages in anger > We are still using large pages but have been doing it in steps over > the last 6 months (IPL with larger LFAREA, convert more WAS regions to > 64-bit). OK, I give up. Why? What is the benefit versus the cost? Even the original literature suggested that it could cause performance degradation for some applications (although obviously not because of bugs). So I'm curious. Who is actually measuring this? What is being measured? and how did anyone determine that it would be beneficial? Adam ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
