Two questions

On 2015-09-19 04:27:59 +0000, David Taylor
said:

On 18/09/2015 23:40, PeterN wrote:
[]
Sorry Alan. On my laptop, the optimum memory for me is 12 GB. Given the
price of memory, I plan to put in 32. Yes there is some loss with more,
as the increase in memory requires more memory to manage it, at the
current price, I can always take some out.

No need to worry about needing "more memory to manage more memory".
It's a /very/ small fractional increase, and more than compensated for
by the increased ability of the OS to cache data in RAM.

If the delta price between 16 GB and 32 GB isn't too great, go for it,
although as Alan say, 16 GB is likely enough. I use MRTG to monitor
memory usage, and here's what my 16 GB system shows:

http://www.satsignal.eu/mrtg/performance_kiruna.php

Even the 8 GB systems here feel comfortable (all running Win-10).

On my Mac I have 16GB, of that I have 9.8GB allocated to PS, I
currently have 4 layers of a 4288 x 2412 @360 tiff open. The memory
usage is 4.53GB. There is still some headroom.
In Lightroom I have left the RAW cache setting at 1GB, and LR is open
and using a massive 450.3MB.
I have also been experimenting with Affinity Photo, and that is open
with a NEF and three layers, and it is using 806.9MB.
The Creative Cloud is using 1.03GB (I don't know why)
Adobe Desktop Services is using 977.9MB.(I don't know why)
My web browser is using up 1.61GB. (I have lots of tabs open)
My Mail app is using 241.2MB.
This Usenet Client is using 167.4MB
My Wacom driver is eating up 24.4MB

The system refuses to give up the 2.24GB it is using.
There is some other stuff running is using 1.8GB.

So of the 16 GB installed, and with heavy multi-task, multi-app usage I
am currently using 13.8GB. If I need to free up some of that RAM there
are a few apps I can quit to give PS that 9.8GB I have assigned it.

I could also go to 24GB or 32GB, but for now I am managing with 16GB.


--
Regards,

Savageduck

On 19/09/2015 06:24, Savageduck wrote:
[]
On my Mac I have 16GB, of that I have 9.8GB allocated to PS, I currently
have 4 layers of a 4288 x 2412 @360 tiff open. The memory usage is
4.53GB. There is still some headroom.
In Lightroom I have left the RAW cache setting at 1GB, and LR is open
and using a massive 450.3MB.
I have also been experimenting with Affinity Photo, and that is open
with a NEF and three layers, and it is using 806.9MB.
The Creative Cloud is using 1.03GB (I don't know why)
Adobe Desktop Services is using 977.9MB.(I don't know why)
My web browser is using up 1.61GB. (I have lots of tabs open)
My Mail app is using 241.2MB.
This Usenet Client is using 167.4MB
My Wacom driver is eating up 24.4MB

The system refuses to give up the 2.24GB it is using.
There is some other stuff running is using 1.8GB.

So of the 16 GB installed, and with heavy multi-task, multi-app usage I
am currently using 13.8GB. If I need to free up some of that RAM there
are a few apps I can quit to give PS that 9.8GB I have assigned it.

I could also go to 24GB or 32GB, but for now I am managing with 16GB.

Wow! I hadn't realised just how profligate with memory (as well as with
money) those Adobe programs are! The Apple software too. Fortunately,
I am spared those problems as much simpler (and lower cost) programs
completely satisfy my photographic organisation and processing needs.
It seems wrong to me that a program needs memory pre-allocated.

On this 8 GB Windows PC, one of my own programs uses 526 MB (it does
have a 124 MPix image and four 30 MPix images - all monochrome
fortunately), Firefox uses 186 MB, Mail 90 MB and Thunderbird (news) 84
MB etc. etc. Oh, and there is a 700 MB (max) RAMdisk for the satellite
reception buffering, but that sizes dynamically and isn't using a lot
right now.

Another difference is that I very rarely use RAW - perhaps reducing the
memory requirements somewhat. I'm less seriously into post-processing
than many in the group, preferring to get things right in the camera,
and I rarely print, the display is my main output medium.

--
Cheers,
David
Web: http://www.satsignal.eu

In article , David Taylor
wrote:

On my Mac I have 16GB, of that I have 9.8GB allocated to PS, I currently
have 4 layers of a 4288 x 2412 @360 tiff open. The memory usage is
4.53GB. There is still some headroom.
In Lightroom I have left the RAW cache setting at 1GB, and LR is open
and using a massive 450.3MB.
I have also been experimenting with Affinity Photo, and that is open
with a NEF and three layers, and it is using 806.9MB.
The Creative Cloud is using 1.03GB (I don't know why)
Adobe Desktop Services is using 977.9MB.(I don't know why)
My web browser is using up 1.61GB. (I have lots of tabs open)
My Mail app is using 241.2MB.
This Usenet Client is using 167.4MB
My Wacom driver is eating up 24.4MB

The system refuses to give up the 2.24GB it is using.
There is some other stuff running is using 1.8GB.

So of the 16 GB installed, and with heavy multi-task, multi-app usage I
am currently using 13.8GB. If I need to free up some of that RAM there
are a few apps I can quit to give PS that 9.8GB I have assigned it.

I could also go to 24GB or 32GB, but for now I am managing with 16GB.

Wow! I hadn't realised just how profligate with memory (as well as with
money) those Adobe programs are! The Apple software too. Fortunately,
I am spared those problems as much simpler (and lower cost) programs
completely satisfy my photographic organisation and processing needs.
It seems wrong to me that a program needs memory pre-allocated.

only because you don't understand how photoshop works, and the memory
is not preallocated anyway.

In article , Eric Stevens
wrote:

I suspect that SCSA (The Secure Content Storage Association), trading
as Vidity will have a major impact on who does what with which in the
near future.

how would that affect apple transitioning to another cpu platform?

I believe it will require a secure chipset.

no

We shall have to wait and
see.

yes

On 19/09/2015 07:28, Tony Cooper wrote:
[]
Just an added comment...people who talk about getting it right in
camera seem to saying that if you do that you don't need to be adept
in post. There are shots you can't get right enough in camera to not
need some post work. Those of us who do work extensively in post are
not failing to get it right in camera, and we only do work extensively
in post when it was not possible to get it entirely right in camera.

I take your point, Tony, but I find that for the great majority of shots
I take, careful setting of the camera exposure produces a quite
acceptable result, with very little need for exposure adjustment
afterwards. Of, course other adjustments may be needed, but they don't
require the greater dynamic range available from RAW.

I'm not making large prints, and others may disagree with "great
majority" and "quite acceptable".

--
Cheers,
David
Web: http://www.satsignal.eu

In article , Pablo wrote:

Sandman:
When you move the cursor from, say, 512,512 to 800,800, no CPU
cycles are used to visually depict that movement.

Unless anything is keeping track of where it is. Which is what
happens in GUI software. All the ****ing time.

Not true, plus - keeping track uses less CPU cycles than *managing* the mouse
cursor, which was the topic. A software cursor means the OS handles mouse input,
mouse position and cursor rendering, and then you have software tracking on *top*
of those CPU cycles.

Bored now. And as I so obviously have already forgotten more than
you and spammy will ever know, I'm off to play another game.

Haha, it's always amazing to see ignorant children find their way onto usenet and
pretend to know things.

--
Sandman

In article , Pablo wrote:

Pablo:
Bored now. And as I so obviously have already forgotten more
than you and spammy will ever know, I'm off to play another
game.

android:
If you're not careful you might become one...

I have to admit, I'm genuinely gobsmacked how stuff has moved on
since my programming days ie; that the firmware on a graphics card
can query a mouse without going through the processor!

No it doesn't. You haven't learned a thing. Did you take English lessons from one
of the resident rpd trolls?

Mouse *movement* is rendered automatically in hardware by the GPU, put directly
into the frame buffer without first being handled and rendered by any software
and thus any CPU cycles.

I've just looked around and I can't find any info on the interface.

There is no "interface", it's automatic. For any software developer, the API to
access cursor information is identical.

I'm guessing there's a way the card queries the PCI bus for USB
activity or whatever. Presumably the card's processor does port IO
the same as an interrupt handler would do it. Sounds interesting.
Does anyone know where I can read about this technology? Purely for
interests sake - I don't intend doing any programming, as I'm
obviously so long out of the loop that it would be impossible.

Well, that's painfully obvious.

--
Sandman

In article , nospam wrote:

nospam:
cursor position information is maintained by the hardware
controlling the cursor, not the cpu.

Pablo:
Now we have hardware controlling stuff. Magic dust anyone? Clue:
hardware doesn't *do* anything.

Incidentally, I've written *loads* of protected mode interrupt
handlers. And even when a processor wants to get a byte from a
port (which involves proc clocks) it does it in an interrupt,
otherwise it would need a dedicated thread - read processor.

You really know *nothing* about this.

Did your trackball controllers run in their own little world
away from the processor?

Sandman:
They do. Most modern graphics cards have support for hardware
cursor, meaning that when you move your input device, the cursor
is drawn on screen by the GPU, directly in the frame buffer, not
something calculated and handled by the CPU. Software can poll and
track the cursor position, but the actual displaying of the cursor
on screen is handled by the GPU in hardware.

the problem is that he's not using modern hardware and doesn't
understand any of the advancements that have happened in the past
couple of decades. all he knows is interrupt handlers.

Yeah, hello 1995

--
Sandman

In article , Pablo
wrote:

the mouse doesn't use cpu.

No, as you say, the graphics card must communicate directly with the mouse
to know what it's up to.

That's very clever. I wonder how it works?

maybe you should stop wondering and try learning something for a
change. a lot has happened since interrupt driven mouses.

On 2015-09-19 00:02, nospam wrote:
In article , Alan Browne
wrote:

I'm betting Apple abandon Intel within the next 5 years (possibly 3) and
that they go all ARM (under the Apple Ax processor line) for OS X. Note
that iOS has high source code commonality with OS X (kernel to just
below the UI)

low end macs will probably move to arm fairly soon to gain significant
battery life improvements as well as reduce components cost.

high end macs will likely remain intel for the foreseeable future.

I'd bet that Apple would convert the entire line right up to the Mac Pro
once they got things going. Maybe not as quick as the Intel transition,
but PDQ.

i doubt it, since intel wins in compute power. where arm shines is
power efficiency, making it suitable for portable products. desktop
products are not on a tight power budget so there's not a strong reason
to make the switch there.

Apple strive to simplify things. More cores is not an issue.