On Sunday, June 13, 2021 at 12:51:20 PM UTC-6, Quadibloc wrote:
> On Thursday, June 10, 2021 at 10:17:57 PM UTC-6, Quadibloc wrote:

> So an instruction set for those header formats has now been defined, and space
> in it is used to provide load and store instructions for aligned operands using the
> banks of 128 registers. While there are those two restrictions on memory reference
> for those instructions, the other painful restrictions are gone; indexing is available,
> and long displacements may be used.
>
> http://www.quadibloc.com/arch/cp0104.htm

And one point which would not be obvious without referring back to the
list of header formats on the initial page:

http://www.quadibloc.com/arch/ct14int.htm

Since this alternate extended instruction format is based on a 32-bit unit
instead of a 32-bit unit, there are fewer prefix bits in the headers, as noted.

This means that, with headers being limited to a maximum length of
64 bits, in this case, *unlike* the case of Extended Mode instructions,
it's possible to have headers with _both_ explicit indication of parallelism
and instruction predication. So, since full VLIW operation is possible with
this instruction set, it's appropriate that it's the one that has good
memory-reference instructions for the large register banks.

John Savard

On 2021-06-08, MitchAlsup wrote:
> On Tuesday, June 8, 2021 at 8:58:30 AM UTC-5, George Neuner wrote:
>>
>>
>> IME, Chrome by itself will happily use more than 8GB.
>>
> It never ceases to amaze me that I can have four (4) 8GB documents open in WORD
> and only use 750KB of memory, yet if I close Chrome with 8 windows open, I go from
> 70% memory utilization to 30% memory utilization ! on an 8GB machine.
>

One thing about web browsers is that they are effectively acting as an
application execution platform, and the web pages are the applications.
The browser uses a whole bunch of tricks to keep memory usage down, but
you need things like in-memory image decoding caches and multiple copies
of rendering surfaces etc to get acceptable performance /and/ security
(e.g. process isolation), and ultimately the memory consumption of the
browser is entirely dependent on the web pages/applications.

I've worked on several web browser engines, and it's /very/ hard to
optimize performance & memory usage since you have no control over
what web devs do. E.g. consider a web page with 100 streaming video
elements. Or a document tree (DOM) with a depth of 100-1000 layers,
for which you need to build rendering acceleration structures. And of
course a user is free to have hundreds of tabs open at the same time.

That said, Chrome is hardly the most memory efficient engine in history.

BTW, a web page can consume considerable CPU+GPU too. A few years ago I
made a fairly simple experiment:

https://hot.bitsnbites.eu/

Beware: I've seen it crash Android phones (required reboot) and make
regular desktop OS:es unresponsive. If nothing else, it will spin fans
and drain batteries. ;-)

/Marcus

On Monday, June 14, 2021 at 3:57:29 AM UTC-6, Marcus wrote:

> https://hot.bitsnbites.eu/
>
> Beware: I've seen it crash Android phones (required reboot) and make
> regular desktop OS:es unresponsive. If nothing else, it will spin fans
> and drain batteries. ;-)

I dared to click on the link. I saw an image of a hot pepper, and some text,
on an otherwise blank page. I closed the page, rather than waiting for bad
things to happen.

But shortly after I closed it, I got a pop-up screen I had never seen before:
AMD had detected a driver timeout, did I want to report it?

So the page was busy doing some things, even if they were too much
for my computer.

John Savard

On Monday, June 14, 2021 at 9:17:27 AM UTC-5, Quadibloc wrote:
> On Monday, June 14, 2021 at 3:57:29 AM UTC-6, Marcus wrote:
>
> > https://hot.bitsnbites.eu/
> >
> > Beware: I've seen it crash Android phones (required reboot) and make
> > regular desktop OS:es unresponsive. If nothing else, it will spin fans
> > and drain batteries. ;-)
> I dared to click on the link. I saw an image of a hot pepper, and some text,
> on an otherwise blank page. I closed the page, rather than waiting for bad
> things to happen.
>
> But shortly after I closed it, I got a pop-up screen I had never seen before:
> AMD had detected a driver timeout, did I want to report it?
<
Back at CMU circa 1973 I had an account on IBM 360/67 running TSS.
Now TSS kept track of your CPU usage and of your memory usage.
So I got in a habit of //unload xxxxx after running an application. This
in turn, cause the OS to issue of application timeout, and exposed a
fault in the OS code that could cause it to crash. When the finger pointed
around to my direction, I explained why I was doing that and that nothing
in the manual indicated I should not be doing that, and that I was saving
my account money by doing that. They decided to fix it instead of bothering
me...........
<
>
> So the page was busy doing some things, even if they were too much
> for my computer.
<
You are not the first.
>
> John Savard