Hi,
Comment from Jack R. Ellis:
=
To set the record straight about Eric's comments of 1-Jul-2005 on
FD-Devel:
First, the UltraDMA drivers have RULES for whether or not they LOAD.
They must find a valid controller chip, one or more UltraDMA disk
drives, and if the "read tests" are in-use each
Hi Tom,
I made just some preliminary tests with RAWREAD, *things*are*better*. I
am not sure exactly what/how.
I will read the source and make some por test starting from that, but
only (sunday maybe?) or monday.
Alain
---
SF.Net email is
Eric Auer wrote:
...
In other words: PLEASE compile LOADFIX into FreeCOM or provide a stand-
alone LOADFIX.COM ... For now, you can use LOADHIGH to get similar
effects than with LOADFIX, but LOADHIGH requires DOS=UMB and some UMBs.
...
um, has anyone heard from Steffen this year? is he still ma
...
either it's a bug - or it's not
if it's a bug, I (speaking generally) might even find it worth fixing
...
tom
This whole discussion is both interesting and amusing (as are many
discussions here...:- )
If you guys do find the cause of the slowdown, and it is fixable, and it
is a k
Hello Alain,
>> just a general hint in benchmarking (for beginners):
>>
>> a) try to measure something with something that you understand what it
>> does; else your measurements might not be interpretable
>> scince you don't know what rawspeed does (*exactly*), or what your
>> indexing app does (
Hello Alain,
> That is why I wish that I could move ONLY the buffers to low memory, but
> as was said here today, this seems too complicated as it is part of the
> kernel internals
BUFFERS= are accessed (and used) only for single sector transfers.
and memcpy()'ing them around is MUCH faster then
Hello Eric,
> part of the problem. AND Jack, I think the 47 seconds were FREEDOS HIMEM
> version 3.10, not MS HIMEM.
interesting 'I think'
WTF is measured ?
> Talking about "every trick in the BOOK", the XMS 3.0 specs explicitly
> tell that you do NOT have to LOCK XMS handles to use the XMS copy
tom ehlert escreveu:
Hello Alain,
just a general hint in benchmarking (for beginners):
a) try to measure something with something that you understand what it
does; else your measurements might not be interpretable
scince you don't know what rawspeed does (*exactly*), or what your
indexing app d
Eric Auer escreveu:
All tests got down to ONLY this umb+udma is very slow (almost the same
as no udma)...
This is what made me think that the UDMA driver could have decided that
UDMA is not possible for some reason caused by enabling UMBs...
that is my guess, but I don know how to test the hi
Hello Alain,
just a general hint in benchmarking (for beginners):
a) try to measure something with something that you understand what it
does; else your measurements might not be interpretable
scince you don't know what rawspeed does (*exactly*), or what your
indexing app does (again, *exctly), i
At 10:28 PM 7/1/2005 +0200, Eric Auer wrote:
Which can for example mean
"EMM386 of FreeDOS has slow VDS". But it can also mean something completely
different.
And has to mean something completely different.
EMM386 doesn't support VDS double-buffering, and reports that when
queried. There i
Hi all...
> >>All tests got down to ONLY this umb+udma is very slow (almost the same
> >>as no udma)...
This is what made me think that the UDMA driver could have decided that
UDMA is not possible for some reason caused by enabling UMBs...
> Uhmm..., Does FreeDOS read/write to an extra buffer
Answer to Jack Ellis, please forward it to him:
Hi, first of all thanks for a lot of information contained in your
message. Then let me say that I did not in any instant complain ABOUT
UMDA, but about WHAT is keeping it from working!!!
Lets take it a part at a time:
- I never herd about the
Hi,
>From Jack Ellis, comments about UMB slowdown.
=
In V6.20 MS-DOS and up (I have V6.22), Microsoft reserves exactly
ONE file "BUFFER" no matter WHAT appears in the CONFIG.SYS file as the
"BUFFERS=" value!
Microsoft GAVE UP on file "BUFFERS" (A) after XMS memory began to be
used and (B)
Hi all,
Johnson Lam escreveu:
Where can I find an explanation about overlap. What is it? I could not
understad from the docs what is overlapping with what.
The XDMA document have detail information. In short, overlap means
XDMA will buffer all output and return to the user's program. If it's
d
Hi all,
Eric Auer escreveu:
All tests got down to ONLY this umb+udma is very slow (almost the same
as no udma). Only umb and *everything* loaded low is just the same. The
real prblem is that without DOS=UMB, I cannot access umb at all.
This would be VERY strange. So you use DOS=UMB and expli
Hi all...
> I could run tests with freedos-emm and umbpci. I tried some cross-use of
> drivers but those that I tested did not run at all :(
DR DOS EMM386 once worked for me, but later, when FreeDOS EMM386 had
all needed features, I stopped using it. I do have a config with MS
HIMEM and EMM386,
17 matches
Mail list logo