Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!samsung!uunet!sjsca4!poffen From: poffen@sj.ate.slb.com (Russ Poffenberger) Newsgroups: comp.sys.ibm.pc Subject: Re: Memory upgrade and refrest rate... Message-ID: <1991May15.182437.7165@sj.ate.slb.com> Date: 15 May 91 18:24:37 GMT References: <1991May3.042704.31762@iitmax.iit.edu> <6710006@pollux.svale.hp.com> Reply-To: poffen@SunOS (Russ Poffenberger) Organization: Schlumberger Technologies, ATE division, San Jose, Ca. Lines: 30 In article <6710006@pollux.svale.hp.com> jgraf@pollux.svale.hp.com (John Graf) writes: >Changing the refresh rate will NOT buy you any performance gain. >The only reason to put faster-access DRAMs in your system is if >you can also decrease the number of wait-states added to memory >cycles by the DRAM controller. Typically the wait-states are >hard-wired into the DRAM controller (they assume a certain speed >DRAM), but some controllers have programmable wait-state registers. >You should try to get a Technical Reference Manual from Tandy (if >one exists) to find out if there's any way to change the wait- >states. If you can't change 'em, don't waste your money on the >faster DRAMs -- your system will still assume you have the slow >DRAMs. > Agreed that putting faster DRAMS than necessary in a machine is a waste of money since the access is fixed byt the number of wait states. However, changing the systems refresh interval CAN cause a small, but perceptible increase in performance. During refresh, the CPU is locked out from ALL memory access, increasing the delay between successive refreshes will allow the CPU to access memory with fewer delays. There are several PD programs available to change the refresh interval. You do run the risk of getting more memory errors if the interval is set too large. Russ Poffenberger DOMAIN: poffen@sj.ate.slb.com Schlumberger Technologies UUCP: {uunet,decwrl,amdahl}!sjsca4!poffen 1601 Technology Drive CIS: 72401,276 San Jose, Ca. 95110 (408)437-5254