From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Mosberger Date: Fri, 16 Jan 2004 19:18:57 +0000 Subject: Re: shared memory segment size alignment Message-Id: <16392.14625.163460.993083@napali.hpl.hp.com> List-Id: References: <20040116183111.96883.qmail@web60002.mail.yahoo.com> In-Reply-To: <20040116183111.96883.qmail@web60002.mail.yahoo.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-ia64@vger.kernel.org >>>>> On Fri, 16 Jan 2004 10:31:11 -0800 (PST), umut aymakoglu said: umut> Hi - In some kernel versions the SHMLBA is defined as umut> 1MB(1024*1024) and in some as 16K. What would be the umut> difference between aligning the shared memory segment size on umut> 1MB versus 16K before calling shmget()? Would there be any umut> performance gain/lose? We are aligning the segment address on umut> 1MB. The ia64 architecture allows CPUs to have a performance penalty when there is virtual aliasing below a 1MB boundary. So if you don't align to 1MB, it _could_ run slower. No existing processor models do that, but future models might. --david