wolfordj@ihlpa.ATT.COM (452is-Wolford) (02/15/89)
QUESTION: Is there a problem with using EXTENDED memory and a communications program or any other program. I read in a recent PC-MAG, FEB 14, 1989, they state that you will loose characters if you use EXTENDED memory for you DISK CACHE while you are using a Communications program. They give some other statements that when you use EXTENDED memory it is slower than EXPANDED because you have to switch in and out of protected mode. DOES this also apply to ram-disks running on extended memory ? Is there a problem and why is it a problem ? Jeff Wolford att!iwsag!jww att!ihlpa!wolfordj harvard-+ | ucbvax--+---- att --+ iwsag!jww | decvax--+
cgs@umd5.umd.edu (Chris Sylvain) (02/15/89)
In article <11702@ihlpa.ATT.COM> wolfordj@ihlpa.ATT.COM (452is-Wolford) writes:
< QUESTION: Is there a problem with using EXTENDED memory and
< a communications program or any other program.
<
< [...]
< Is there a problem and why is it a problem ?
It is a potential problem caused by the fact that using the BIOS service 15h
to move a 64kbyte block (maximum) into extended memory will disable
interrupts for 1 msec, measured on a 8 MHz IBM-PC/AT. The return to real
mode itself takes 400 microsec.
The BIOS service itself disables all interrupts, rather than deciding for
the user which interrupts are OK in protected mode, and which are not.
It is said that to use the BIOS 15h service, one has already done nearly
all the work to skip using the BIOS altogether and enable those interrupts
which are desired to be serviced in protected mode.
My source is _The IBM-PC from the Inside Out/Includes the IBM-PC/AT_ by
Sargeant and Shoemaker III, pub. by Addison and Wesley.
---> To further examine the problem, I would like to know what the expected
---> frequency of interrupts is for the AT I/O chip (the 16450?) and the
---> XT I/O chip (8450B?) when operated at 9600 baud.
--
--==---==---==--
.. He chortled in his joy. ..
ARPA: cgs@umd5.UMD.EDU BITNET: cgs%umd5@umd2
UUCP: ..!uunet!umd5.umd.edu!cgsswh@hpsmtc1.HP.COM (Steve Harrold) (02/15/89)
Re: Extended memory and disk caching I have not read the referenced article, but my own experience is quite the contrary. For the last 2+ years, I have been using extended memory for both RAM disk and disk caching on my AT clone, using VDISK and Super PC Kwik. At the same time, I spend a lot of the day connected to other minis via TCP/IP LAN card and by RS232 at 9600 baud, doing terminal I/O and file transfers. I have never lost a byte.
Sean.Conner@f7.n369.z1.FIDONET.ORG (Sean Conner) (02/26/89)
Chris Sylvain writes: >---> To further examine the problem, I would like to know what the expected >---> frequency of interrupts is for the AT I/O chip (the 16450?) and the >---> XT I/O chip (8540?) when operated at 9600 baud. Well, at 9600, you have 9,600 bits of information per second, or one bit per 104 uSeconds (micro seconds). Since there's 10 bits sent per byte (Start, 8 data (7 data, 1 parity) and Stop), you have (assuming that data is comming is ASAP) a Received-Data interrupt every 1.04 mSeconds (milliseconds). The frequency if interrupts is independant of the I/O chip (as it seems to me). Hope that helps somewhat. -Sean