D.C.Halliday@newcastle.ac.uk (D.C. Halliday) (02/25/90)
Can someone out there tell me what memory speed is required for the STe or standard ST. (I suspect 100ns) Also if I use faster memory such as 70ns will asyncronous tasks such as burst filling a 16Mhz Hypercache or DMA access will these be faster? With the new MMU in the STe and spare/reserved bit patterns in the memory configuration can the STE's memory be expanded beyond 4MB. (I guess adding a new bank size of 8MB would be possible thus allowing the full 16MB of address space.) any comments? Many Thanks Dave H. (D.C.Halliday@newcastle.ac.uk)
knop@duteca (P. Knoppers) (02/27/90)
In article <1990Feb24.222700.22004@newcastle.ac.uk> D.C.Halliday@newcastle.ac.uk (D.C. Halliday) writes: >Can someone out there tell me what memory speed is required for the >STe or standard ST. (I suspect 100ns) Also if I use faster memory such >as 70ns will asyncronous tasks such as burst filling a 16Mhz >Hypercache or DMA access will these be faster? I recently upgraded my 1040 STf to 2.5 MB. I have now 16 256kbit chips that are rated 150 ns on my shelve doing nothing (any takers ?). So the answer to your first question is: 150 ns is sufficiently fast for an 8 MHz MC 68000. (Does the STe run on 8 MHz ?) As to the second question, no, I would not expect DMA to be able to exploit faster RAM chips. I don't know what Hupercache is, but unless it is hardware, it can not exploit faster RAM chips. By the way, even if it could, the improvement in performance would be negligible, because access times for a hard disk are in the order of 25 ms. Transfer rate to the computer must be extremely slow to make any difference. >With the new MMU in the STe and spare/reserved bit patterns in the >memory configuration can the STE's memory be expanded beyond 4MB. (I >guess adding a new bank size of 8MB would be possible thus allowing >the full 16MB of address space.) any comments? Well, I would expect the new MMU to be an improvement over the old one. However, the full 16 MB address range can never be dedicated to RAM, because then there would be no space left for ROM and I/O. I have no documentation on the new MMU, so this is only educated guesswork. -- _____ __ __ __ __ ____ _____ _____ ______ _____ _____ | _ \ | |/ || \| | / \ | _ \ | _ \ | ___|| _ \ / ___| | __/ _ | < | || || || __/ | __/ | >__ | < \__ \ |__| |_| |__|\__||__|\__| \____/ |__| |__| |______||__|\__||_____/ P. Knoppers, Delft Univ. of Technology, The Netherlands, knop@duteca.tudelft.nl
bwhite@oucsace.cs.OHIOU.EDU (Bill White) (03/01/90)
In article <729@duteca.UUCP>, knop@duteca (P. Knoppers) writes: > > I recently upgraded my 1040 STf to 2.5 MB. I have now 16 256kbit chips > that are rated 150 ns on my shelve doing nothing (any takers ?). So the > answer to your first question is: 150 ns is sufficiently fast for an > 8 MHz MC 68000. (Does the STe run on 8 MHz ?) > Keep in mind, however, that both the microprocessor AND the video chip have to access that RAM in one cycle. I've heard that 150ns chips can cause problems with Overscan, for instance. I don't offhand know which I have, I think I have 150ns, but I've had erratic memory problems ever so often, and (though I can't prove anything) I suspect it may be because one of these chips is "not quite" 150ns. So get 120ns and be safe, at the very least. Besides, they aren't that much more expensive these days. Bill White bwhite@oucsace.cs.ohiou.edu