25 Feb
2010
25 Feb
'10
6:05 a.m.
Richard <legalize at xmission.com> wrote:
In article <4B7AF1F7.8030008 at softjar.se>,
Johnny Billquist <bqt at softjar.se> writes:
That was a bad argument against it... :-)
> After searching the net for a while, I think
it is an E&S graphic
> system. ES340 seems most likely, but ES390 might also be it.
I doubt its an E&S system because its raster based. AFAIK, all the E&S terminals and DEC peripherals
were vector based
until the PS/390. If this were a PS/390, then it would have the
characteristic black enclosure for the monitor, keyboard and tablet.
Also, E&S products were high-end and had hardware line drawing at a
minimum, whereas this system appears to be a simple memory-mapped
frame buffer with no additional hardware acceleration.
I think we should keep E&S systems apart from DEC systems. They are not
the same, and mixing them up just confuses the issue.
The system in the video was not something by DEC. However, DEC had
bitmapped graphic systems long before this, but not with the resolution
and depths shown in the video.
The PS/390 was the first "Picture System" to
support raster
operations. The primary market for these peripherals was the
molecular design/molecular modeling community and they needed very
high quality line renderings of their models. Until the antialiasing
technology for raster graphics introduced on the PS/390 the only way
to achieve that quality was to use a vector based display. The raster
display decoupled the display refresh rate from the frame rendering
rate, allowing very high quality renderings of complex models at the
cost of interactivity.
Well, the E&S PS/340 also was a raster display system, as an add-on to
the PS/330.
See http://www.bmsc.washington.edu/people/merritt/graphics/ps330/ps330.html
I think it might very well be that which is shown on the video.
Raster operations are another issue. The video is showing a bitmapped
high resolution picture, and it obviously isn't too fast at loading
another high res picture. And that is all we see them doing. No raster
operations of any kind (unless you count the basic setting and clearing
of pixels).
Similar technology to that used in the PS/390 was used
in E&S's first
workstation the ESV. I worked on the ASIC that performed this process
in the ESV in the summer of 1989. The workstation was released in Q4
1989. All the E&S products were raster from then on, with the
exception of calligraphic light overlays used in the simulation
products.
I have a PS/390 base and several ESV workstations in my collection.
Nice.
Johnny
25 Feb
25 Feb
3:56 p.m.
In article <4B86596F.5000603 at softjar.se>,
Johnny Billquist <bqt at softjar.se> writes:
Well, the E&S PS/340 also was a raster display
system, as an add-on to
the PS/330.
See http://www.bmsc.washington.edu/people/merritt/graphics/ps330/ps330.html
That's the first I've heard of that; it must have been an interim
product before or contemporary with the 390.
Very few early E&S systems survive and since they never sold in high
numbers, its also hard to find people who used them.
I think it might very well be that which is shown on
the video.
I still doubt that. As I mentioned earlier, the PS series was higher
performance than what we see in the video. Also, being a high-end
system the monitors and tablets would have been labelled with the E&S
brand and I don't see that in the video as I have seen on monitors,
keyboards and tablets used for those systems. Now that I think about
it, the lack of a keyboard is another giveaway that this is not an E&S
system. A keyboard for the serial terminal is shown, but not one for
the graphics system directly. The E&S PS series had integral
keyboards and were full-blown terminals and not just a memory mapped
frame buffer.
For the black enclosure look, see these items at the CHM:
<http://www.computerhistory.org/collections/accession/102633847>
<http://www.computerhistory.org/collections/accession/102633848>
<http://www.computerhistory.org/collections/accession/102633849>
PS 330:
<http://www.computerhistory.org/collections/accession/102691276>
For a Picture System brochure, look here:
<http://www.computerhistory.org/collections/accession/102646288>
CHM has a brochure for the PS 340 but its not online:
<http://www.computerhistory.org/collections/accession/102646825>
<http://www.computerhistory.org/collections/accession/102646821>
Ditto PS 320, PS 330:
<http://www.computerhistory.org/collections/accession/102646824>
<http://www.computerhistory.org/collections/accession/102646823>
<http://www.computerhistory.org/collections/accession/102646822>
Raster operations are another issue. The video is
showing a bitmapped
high resolution picture, and it obviously isn't too fast at loading
another high res picture.
The video also shows some painting operations, particularly what would
be called clone brush by most programs today. This is a raster op,
albeit the simplest of raster ops -- copy.
However, with a memory-mapped framebuffer, copy is easily implemented
on the CPU and can be reasonable fast, certainly fast enough to keep
up with the size of brush they were using in the video. Hardware
rasterops become more important when you start running a windowing
system and need to move around big chunks of screen as a common case
(moving windows in response to the mouse).
Bottom line for me is that this looks like a 3rd party memory-mapped
framebuffer and not much more. That makes sense, because it would be
the cheapest way to make graphics available at the time and if you
aren't doing molecular modeling or manipulating complex CAD models, it
isn't worth spending hundreds of thousands of dollars so you can
doodle.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download
<http://legalizeadulthood.wordpress.com/the-direct3d-graphics-pipeline/>
Legalize Adulthood! <http://legalizeadulthood.wordpress.com>
4:01 p.m.
At 02:56 PM 2/25/2010, Richard wrote:
That's the first I've heard of that; it must
have been an interim
product before or contemporary with the 390.
Very few early E&S systems survive and since they never sold in high
numbers, its also hard to find people who used them.
Did I miss the part of the discussion where anyone went to the
YouTube uploader of this clip, and asked where it came from?
- John
5:34 p.m.
In article <4B86E75E.2070403 at bitsavers.org>,
Al Kossow <aek at bitsavers.org> writes:
On 2/25/10 12:56 PM, Richard wrote:
I don't know what machine room presence it would have other than some
boards inserted into the backplane. E&S equipment had its own cabinet
because it had a "GPU". Memory mapped frame buffers aren't much more
than array of memory chips and a simple video scanout circuit. You
can easily fit all of that onto one or two PDP-11 or VAX boards
without requiring an additional cabinet, even with SSI/MSI parts.
In 1984 I was an undergrad helping a grad student with his hardware
project for his PhD. I wirewrapped what was then considered a high
density layout for a graphics system with lots of smarts. A memory
array, a bus interface, and a video scanout circuit are all fairly
straightforward to implement with period parts in a board or two.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download
<http://legalizeadulthood.wordpress.com/the-direct3d-graphics-pipeline/>
Legalize Adulthood! <http://legalizeadulthood.wordpress.com>
Bottom line for me is that this looks like a 3rd
party memory-mapped
framebuffer
I'm not convinced. There was no sign of it in the machine room.
5222
days inactive
5222
days old
4 comments
4 participants
participants (4)
-
aek＠bitsavers.org -
bqt＠softjar.se -
jfoust＠threedee.com -
legalize＠xmission.com