Re: [Nolug] Go

On Thu, Dec 12, 2013 at 4:02 PM, John Souvestre <johns@sstar.com> wrote:

> Ultimately, I believe that as processor power increases horizontally
> (cores) rather than vertically (speed) that the only way to make use of it
> is to run parallel. For any given task, especially those which are
> centered on doing one big thing, you can parallelize by hand. But if you
> simply want your average day-to-day code to run faster without having to
> think about it and do any extra work then I think that the programming
> language you use is key. Otherwise the average programmer won’t want to do
> the extra work it normally takes now.
>

For highly parallel processing; we are probably already there, with GPU
computing and CUDA; that allows you to run C++ code on the video boards,
as long as your computation is suitable. nVidia has consumer level
video cards such as the GTX690 that ship with 4GB of GDDR5 and over 3000
CUDA cores; there are also such GPUs as the Tesla K1 with 16GB of GPU
memory.

With video cards that can do over 1 billion MD5 hashes per second;
all the weak password hashing algorithms will be giving way shortly.

How can you call a NT or SMB password hash secure; if you can enumerate
all 8 character combinations all uppercase lowercase letters, and all
symbols on a keyboard with a mid-end gaming PC's video card in less than 2
hours?

Processor performances that increase sequential processing (speed) are much
more beneficial for ordinary applications than mere multiplication of
cores, but increasing speed beyond a point becomes much harder and
exponentially more expensive.

Most applications will not have great benefits beyond a few cores, even
if programmed concurrently; your operating system can easily sabotage your
efforts to take effective advantage of more than 3 or 4 cores.

You really need to have less than 1% "sequential overhead", for ideal
parallelism; which means no data dependencies.

--
-JH
___________________
Nolug mailing list
nolug@nolug.org