#Mitsumi Benchmark
16 messages in this thread
I have spent the last half hour reading through the Mitsumi problems (DMA
versus software) and found some good information about the problems I have been
having. I just can't get the Mitsumi that I bought this week and the Colorado
tape drive to be happy with each other. I gave up and dropped DMA.
I just wonder if anyone knows of a benchmark program for testing CD access and
transfer rates. As far as I know, you just have to trust what the manufacturer
claims. Anyway, it would be nice to know what I'm giving up by not trying
harder to get DMA access to work.
Steve — If you have a copy of "The Animals" there is a test program on there
somewhere (but,er,uh I ferget how to use it) — also Profit Press was giving
away a demo cd-rom (MegaDemo) that has several cd-rom tests and other utilities
— well, they want $6 for S&H — check for one of their ads in PCMAG. If you
have the 005, you will find that the figures are close to those advertised, tho
there are a few things not advertised that are a bit slow. Still, a fine drive
for the money.
Craig
Thanks for the info.
It does seem to work well if you don't try to juggle tape drive,
ethernet, Windows 32 bit drive access, etc. with DMA addresses, IRQ and base
addresses. I have the older (apparently) drivers and need to download the
drivers that are dated in '93. The Mitsumi drive is stamped January 1993, but
the drivers are 1992. Maybe the newer drivers will make things easier.
Steve
Steve:
Profitt press reports a *true* speed of 370ms, close enough to the advertised
350 to be true…
Bruce
Steve,
There's a program you may want to check out over on the Zenith Data Systems
Forum (GO ZENITH) Section 5 – DOS Utilities. Here's the abstract:
.. [76703,662] Lib: 5
.. CDTA.ZIP/Bin Bytes: 14661, Count: 206, 06-Nov-92
..
.. Title : CDTA diagnostics and tests for CD-ROM drives
.. Keywords: UTILITY CD-ROM CDROM TIMING ANALYSIS DIAGNOSTICS SPEED TESTS
..
.. CDTA is a "CD-ROM Timing Analyzer." It measures data throughput of a
.. CD-ROM drive/controller and estimates access time. Those numbers should
.. prove useful in comparing CD-ROM drives for speed. CDTA also tests CD-ROM
.. drives, controllers, and disks for diagnosis in the event of problems.
.. Free.
I've used it and it's a good clean program that seems to work as advertised.
>>Anyway, it would be nice to know what I'm giving up by not trying
harder to get DMA access to work.<<
I still maintain that for this type of testing – all the CPU is doing is
processing CD-ROM activity – interrupt driven software reads are just as
fast as DMA. In fact, Richard L Shorter 71740,156 makes a good argument for
why they are FASTER than DMA.
But that is for a situation where the processor doesn't have much else to
do. In theory, when you get into a Multimedia type situation where all
sorts of things are happening as close to "at once" as possible (video
updates, CD-ROM reads, sound, etc.) and the system is pushed near the limit,
letting the DMA handle the CD-ROM transfer should increase system
throughput. I'm still looking for a definitive test to prove or disprove
this idea.
It's late. I gotta put away my soapbox and go to bed.
Good Luck.
Ken
Thanks–
I'll give it a try. The Mitsumi docs say that using the DMA driver is
"preferable", I guess they mean faster. But still, I don't like it when I can't
get something to work like it supposed to.
Steve
Hi,
the answer to the question "What's faster ? DMA or IRQ ?" is in my opinion DMA
because the CPU gives only the command to get something from CD and the rest
runs automatically through the DMA-Controller.
A good test for this is "The 7th Guest" ! I let it run with the IRQ and when
the mentioned "Multimedia-Situation" occurs (means reading Data from CD,
playing FM-Music and playing digitized speech) then the music slows down and
the speech is sometimes "hacked" (as I call it).
I tried the DMA-option (yes, sometimes it works great !) then everything works
fine – no slow down in music and the speech is superb.
By the way i am using a 486DX/50 with 8 megs and a SB PRO
Erik
Erik,
Thanks for your input. I haven't had time to set up a similar test on my
system (hope to get to it soon). The kind of test you ran seems like a valid
way to test the system – real life operation.
I'm also running a 486DX-50 w/16meg and a SB PRO. Not wanting to get too far
off the subject, but do you have any problems initializing the SB PRO in
Windows? I just can't get mine to come up right (the master volume control on
one channel is set to zero). I've been on the Soundblaster forum and they say
"there is a problem with 50MHz processors and the Opti chip set". And "we're
working on a cure".
Ken
Hi Ken,
sorry i've got no probs with my Soundblaster Pro 2 ! This one works fine – but
thanks for mentioning that the SB may have probs with the OPTI-chipset because
I'm thinking of buying a new board (VESA or EISA or both).
Erik
Erik — My original soundblaster works just fine with my OPTI chipset 486-33.
Craig
Ken,
another "real life operation – and failure": I bougt MS's Cinemania Mulitmedia
CD. Because my mainboard (486/33 – Symphony) doesn't like to do a DMA transfer
with the Mitsumi CD ROM, Windows does not play the animated film sequences _at
all_! Even the button to do it is not displayed. Obviously without DMA you
cannot achieve the "max. 40% CPU usage", that is allowed for MPC compliancy;
polling takes probably full CPU power.
Michael
Sorry for butting in, but I too have a Mitsumi and Cinemania. I was just
curious to know which films had animated sequences. I thought that all the
video were just stills. If they exist and I can't play them I'll be very
disappointed.
Richard — I don't have Cinemania, but seem to recall that one of the things
that reviewers didn't like about it was the lack of action videos — and that
was the reason I didn't buy it.
Craig
Craig,
Thanks for your reply. I too wished it had some action videos, but even
without that it is still an entertaining piece of software. Regards, Richard
Michael,
Thanks for your input. It supports my contention that DMA is required for
multimedia. I'm trying to resolve the DMA software problems with Mitsumi. I'll
let you know if there is any success.
Ken
Ken,
that'll be fine. I hope it is only a driver problem.
Thanks Erik