Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations IamaSherpa on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

The free lunch is over

Status
Not open for further replies.

chiph

Programmer
Jun 9, 1999
9,878
US
The article says that processor power has topped out, and that for any future performance gains, programmers must learn how to write concurrent software to take advantage of multi-core CPUs and hyperthreading.


Chip H.


____________________________________________________________________
Click here to learn Ways to help with Tsunami Relief
If you want to get the best response to a question, please read FAQ222-2244 first
 
jsteph, I agree entirely. But look at it another way: Virtually universally, the more senior a manager is, the larger his chair. I've never seen any evidence that managers develop larger or more delicate bottoms as they get promoted. In the same way, they always have bigger desks, even though it's their secretary who probably needs the bigger desk.

There are not a lot of senior managers who will tolerate having an older, slower cpu than their staff. I'm sure that's one reason why desktops get so vastly powerful.
 
Lets go back a few years one of my last non pc boxes was an Atari Falcon 16mb with 1gb drive (trust me that was huge and the 16mb cost me close on £200) and an accelator (32mhz i think)card.
On this I ran Steinbergs Cubase Audio, a word proccesor and a decent imaging program.
Many a time we put it up against my friends state of the art P90 (clocked to 110mhz) with 128mb and a 2gb drive running 95.

In nearly every case the Atari trounced the pc in rendering, opening, saving, converting various files. As for Audio well the Atari was in a league of it's own.

Now bearing in mind the proccesing and memory differences, why did the Atari win. In my opion it's simple, as the Atari had so little processor power and memory (a standard falcon was 4mb and 16mhz), the programmers had no choice but to write good quality code.
It was the same with games. Poor graphics and poor sound mean't one thing, to survive you needed gameplay. Now if you have a poor game, stick lots of pretty sound and graphics and hope no one notices.
I'm sure the writers of the game MDK stated they developed on low spec, poor quality machines. If the game slowed down, crashed or generally was of poor quality, they went back and rewrote the code until it worked. Now that was quality programming.

Rant over.....
Stu..

Only the truly stupid believe they know everything.
Stu.. 2004
 
Talking about quality programming, anyone ever play Chuck Yeager's Air Combat ?
It was a flight sim from 1991. It played very well at the time, and guess what ? IT STILL DOES !
That's right, a game from 1991 that ran on a 386 at 40Mhz now runs on a machine that is 20000+ times faster, and everything still works just like it should.
The guys that coded that game were so good that they managed to make their game react to totally unforseen hardware modifications, and do so gracefully. In 1991, nobody even dreamed of multi-gigahertz processors, or DDR memory, and yet they managed to make their circa-1991 code operate flawlessly on circa-2005 hardware.
Is that quality programming, or what ?

Pascal.
 
Darn, you had a 40 mhz 386? Mine was only 16 mhz. DX or SX? :)

A programmer named Pascal, how did that happen?



BocaBurger
<===========================||////////////////|0
The pen is mightier than the sword, but the sword hurts more!
 
DX obviously, an AMD version (one that didn't totally break the floating point unit). It was fun.

As for the name, ask my parents ;-)

Pascal.
 
You know, I can almost guess why. If you had EGA or VGA graphics, you drew on the real screen; therefore to avoid drawing catastrophes (noise, messy images) you had to coordinate with the raster beam, which meant the program speed was determined largely by the screen refresh rate - which is governed by what the human eye will put up with - which is still the same!

(Of course they probably still did a bit of speed-checking. If your machine is so slow that it takes more than a screen refresh to do all the calculations, then the program will suddenly jump in speed as processor speed improves. And I'm certainly not denying that it was quality programming. It takes a bit of skill to work round a raster beam and get good performance out of a minimal EGA/VGA)
 
Nano-tech might improve processor power.
And hopefully some genius will invent a new computer model different from the Turing Machine (all computers are turing machines).
 
FYI -
AMD is thinking of selling the dual-core Opterons at a discount, making them comparable in price with the single-core processors. They'll likely go up after the introductory period.

Chip H.


____________________________________________________________________
If you want to get the best response to a question, please read FAQ222-2244 first
 
I've just had this link in a newsletter, it is certainly relevant to this thread:
What this means, of course, is that operating systems that only support single CPU systems (eg XP Home) will have to be rewritten or replaced with an SMP aware OS to take advantage of the dual core operation.

John
 
I didn't see a mention of that in the article but it stands to reason. I would not be surprised to see some XP Home SMP Edition arrive on the scene though, but surely a Longhorn Home would have this ability from the start. Worst comes to worst, you just use XP Pro.

Relatively few people bother with a full motherboard/CPU upgrade anymore, and would buy a new machine. Since the existing machine with XP Home probably has an OEM license (not transferable to a new machine) there is no impact there.

Those who DO major upgrades to use such a dual-core processor chip would probably be faced with an upgrade to Pro. Wouldn't you think most people inclined to do this are already running Pro though?

So I guess I agree with you but I don't see a major impact. White box OEMs like those in Microsoft's System Builder program would just use another OEM license to keep the cost relatively low.
 
In fact, re reading the article above, its not even that simple. From that article:

article said:
In effect you get two Prescott P4 CPUs each with Hyper Threading (HT) in a single package, which can help boost PC performance when running suitable multithreaded applications - to Windows, the 840 appears as four virtual CPUs.

XP Pro and 2K Pro only support 2 CPUs. For more than that you need a server OS.
With current systems, this means Windows 2000 Server or Windows 2003 Server on the desktop PC just to take full advantage of the new CPUs.
This will of course need to change in the future to accomodate the new generation of CPUs.
The alternative is to disable Hyperthreading features to avoid the need for expensive server operating system licenses.

John
 
jrbarnett:

What this means, of course, is that operating systems that only support single CPU systems (eg XP Home) will have to be rewritten or replaced with an SMP aware OS to take advantage of the dual core operation.

Unless some clever programmer somewhere comes up with code that sits on top of the OS and parses instructions to the different CPUs.
 
I think people running XP Home will be out of luck. AFAIK, it can't even take advantage of hyperthreading.

From what I've read, Microsoft is trying to decide what to do with their licensing. They make a lot of money when companies have to step up to Windows Advanced Server because they need to go to 4 CPUs, and I can't see them giving that up. Don't forget that the companies also need to buy SQL Server Advanced edition to run on that 4-cpu box. The regular SQL Server will only recognize 2 cpus.

Chip H.


____________________________________________________________________
If you want to get the best response to a question, please read FAQ222-2244 first
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top