Re: Hardware has finally outrun bloatware?



On Sun, 1 Feb 2009 20:40:59 -0800 (PST), Robert Myers <rbmyersusa@xxxxxxxxx> wrote:

On Feb 1, 10:01 pm, Alphonse Q Muthafuyer <muthafu...@xxxxxxxxx>
wrote:
OK, you've just spent $1k on what should be a very, very nice system. I
should probably keep my mouth shut, but ...

Have you ever heard of i/o?<g> Are your disks electro-mechanical? Do you
not still have i/o-bound processes whose lack of speed is correlated
with the volume of (disk) data to be processed?


Well, sure. I could pick a set of disk-intensive tasks that
bottleneck with disk i/o. If that has prospects for making the system
unresponsive, I guess I don't understand how. Those applications will
only be able to proceed as fast as they can get and put information,
but that's life. What won't happen is that the system won't come to a
screeching halt as it mixes up heavy disk i/o from applications with a
need to swap memory to disk because I have too many applications
open. If my view is naive, educate me.

The history of computing is replete with repetitions of sequences like:

a.) Hardware developers invent new viable technology.
b.) Software developers (who have been watching carefully)
write code to take advantage of a.).
c.) Business, research, etc folk find ways to make use of new
computing abilities/resources.

In the case of, say, generations of memory technology (ie DDR, DDR2, etc)
the sequence can proceed relatively quickly. New, faster memory, they've
already been working on 64-bit systems that allow addressing beyond
old restrictive boundaries (ie 4 gb limit), things proceed with no
major obstacles.

In the case of parallelism, things are different. To implement, a
lot of software has to be rewritten, and there are limits on what
processes lend themselves to running in parallel. See:

http://www.informit.com/articles/article.aspx?p=1193856&rll=1

On the other hand, if you are not working for anyone, and are merely
designing experiments for yourself, you can limit yourself to the
context of your available resources and, I suppose, be happy.

So, maybe the step up for this system is solid state disks rather than
more memory.

With the right (I assume server-grade) pipe, SSD can be very much
like adding memory for reads. Writes take longer.

Your 6 gb sounds very adequate for ordinary and perhaps some not-
so-ordinary desktop applications. Outside the domain of desktop
apps, 6-12 gb is not so impressive. It all depends on the apps.

Seriously, you are running hard-core simulation software that is
fully multi-threaded? You've checked and you get a nice, even
distribution of cycles across the 4 cpu's?

Parallelism is well in hand for the kinds of tasks that interest me.

That is a less-than-detailed response. Are you adept with Task
Manager?

Not certain I understand the issue with the display driver. Brought
the whole system down? More detail might be helpful.

Yes indeed. Shut the system down. Literally. There have been two
incidents with the display driver. One made the system burp, with
some kind of announcement from which it recovered. The other produced
a panic shutdown. The diagnostic information from the panic shutdown
said that the problem occurred in the display driver. I'm not a
windows guy, so I don't understand how the OS is unable to protect
itself from a device driver, but apparently that's the case.

Your ATI® Radeon HD 4850 driver is fully supported for Vista Home
Premium 64? Have you made support inquiries? There are "little holes"
in all Windows products: it's possible that a problem with your
driver has found one. With such a system, I'd follow the usual
support channels first. If that yields no relief, you might want
to take a peek at :

technet.microsoft.com/en-us/sysinternals/default.aspx

AQ

"The monkey and the baboon was playing 7-up.
The monkey won the money but he scared to pick it up.
The monkey stumbled, mama.
The baboon fell.
The monkey grab the money and he run like hell!"
- from "Dirty Motherfuyer", Roosevelt Sykes, around 1935
.

Relevant Pages

  • Re: non-paged memory inside a kernel driver
    ... There is just one instance of the driver and it will be ... disk and one flash. ... so I need 100MB of non-paged memory inside my driver. ... allocate memory, the virtual address space has to have a big enough gap ...
    (microsoft.public.development.device.drivers)
  • Re: Help reading minidump
    ... Corruption in the NTFS file system or bad blocks on the hard disk can induce this error. ... Another possible cause is depletion of nonpaged pool memory. ... If the nonpaged pool memory is completely depleted, this error can stop the system. ... However, during the indexing process, if the amount of available nonpaged pool memory is very low, another kernel-mode driver requiring nonpaged pool memory can also trigger this error. ...
    (microsoft.public.windows.vista.hardware_devices)
  • Re: Help - hooking up DSP to XScale running CE6!
    ... There's no reason to try to use something that appears like a "disk" as the ... define what interface you're going to have, then implement the right driver ... the DSP host port can be mapped into memory. ... We have a BSP, but not the source of the BSP. ...
    (microsoft.public.windowsce.platbuilder)
  • Re: Memory Performance Issue with Fedora Core 2 Kernels
    ... Memory Performance Issue with Fedora Core 2 Kernels ... >>reserve the top of physical RAM for my driver. ... >>pointer to save the data to disk. ...
    (Fedora)
  • Re: 2.6.23-rc4-mm1 OOPS in forcedeth?
    ... # Generic Driver Options ... # PCI IDE chipsets support ... # SCSI support type (disk, tape, CD-ROM) ...
    (Linux-Kernel)
Loading