Intel finally to guzzle less!?

Intel is finally thinking about CPU performance/watt and fooling some analysts into thinking they invented the concept. If you read this "Motley Fool" article and you know anything about the history of Sun Sparc, AMD, Transmeta, VIA even PowerPC chips you'll wonder how the author of this article was able to get it so backwards! I first heard "performance per watt" at a Sun customer engineering conference several years ago. I wish someone had published the graph I saw there. Intel was winning the CPU clock frequency race, but how many of us know the SPECs/Watt figures for the P4 and Itantium? Here are a few observations: I didn't know there was a fan on my G3 (PowerPC) powerbook until I set a hot-burning Wintel laptop on top of it. The powerbook fan kicked into high gear and the wintel 'dry heat' seems have partially fixed a powerbook display problem caused by the high humidity here. The Wintel laptop occasionally overheats and shuts down, especially when running a CPU intensive program in summertime (outside of Ireland.) My Powerbook grinds for hours on a video compression, but it has never overheated. When new, the powerbook boasted 10 hours on a set of batteries. Today it is nearly impossible to find a laptop with this much battery life. While I'm frustrated with the inaccurate implications in the "Fool" article (not to mention the fact that Sun stock price and my livelihood are influenced by such nonsense), I don't hold anything against Intel. They must know that consumers aren't going to buy baskets of MHz anymore, they're going to buy solutions. And while Intel promised an efficient multicore CPU sometime in the future, Sun employees stood outside the forum with servers based on more efficient AMD chips which are ready to use today. Welcome to the efficiency club Intel.

Ermmm, More power to you!

Comments:

I think that Intel is very hesitant to push the XSCALE arch as a Desktop/PC architecture on the same footing as x86, this would have solved a lot of their problems (imagine 80200/80321 with a co-processor or in dual processor motherboards). This is one more entry in their long list of errors. I just cannot understand why they do not use XSCALEs as desktop alternatives , Linux/SUN/BSD/RISCOS could have saved their business. We do not like all these useless gigahertzs and wasted watts. An IYONIX (really expensive) type PC from intel could have served well 50% of their market and would have lowered the bills. Waste less , use XSCALE. Sun could push a solution like ODW in this area also. Help us SUN (one more time !!!)

Posted by Vasileios Anagnostopoulos on September 11, 2005 at 07:01 AM GMT+00:00 #

I can understand Intel's reluctance to kill the X86 "golden goose." When the dominant X86 O.S. demands brute force, chipmakers are forced in that direction. I'm glad sysadmins are starting to notice that though X86 PCs are 'cheap', "throw more PCs at it" isn't the most efficient solution to every problem. And now that oil is above $65/barrel, inefficient solutions are seldom the cheapest solution. It reminds me of a bit from the old Bob Newhart show: Bob forgot to thaw the thanksgiving turkey. Howard: "Why don't you turn the oven up to 2000 degrees?" Gerry:"It only goes up to 500." Bob:"Then we'll use four ovens."

Posted by bnitz on September 11, 2005 at 02:54 PM GMT+00:00 #

Post a Comment:
  • HTML Syntax: NOT allowed
About

bnitz

Search

Archives
« April 2014
MonTueWedThuFriSatSun
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
    
       
Today