by Grigoriy Gubankov
06/24/2003 | 05:36 PM
Welcome to the new issue of X-bit Watch. Despite Summer time and calmness in the market and particularly in new product announcements, some interesting articles and reviews have appeared over the Internet and we are hurrying to bring you the creams of the Web.
<%BANNER[article]%>Of course, Intel Pentium 4 is the king of the hardware review hill these days. I think, you have already read our review of this great microprocessor, but if you want to learn even more about this CPU, you should look through the following articles:
AnandTech – “Intel Pentium 4 3.20GHz - The Real Slim Shady”;
TweakTown – “Survival of the Fittest – Intel Pentium 4 3.20GHz vs. AMD Athlon XP 3200+”;
HEXUS.net – “Intel Pentium 4 3.20GHz review”;
SharkyExtreme - “Intel Pentium 4-3.20GHz Review”;
Computerbase.de – “The Last Northwood” (in German | translated);
K-hardware.de – “Pentium 4 3.20GHz in Test” (in German | translated).
It is widely known fact that default multiplier and voltage on AMD processors can be easily modified using “golden bridges” on the top of such CPUs. Furthermore, some even know that these bridges are directly connected to CPU pins, but the correspondence between the bridges and the pins remained unknown until now. OCinside.de, a web-site we are familiar with for its interactive CPU multiplier guide, has posted a script called “Pin-Mod guide”. The name speaks for itself - you can pick CPU type, multiplier, Vcore and immediately see, which pins are needed to be shortened in order to get such settings. If you want to shorten pins not on the CPU, but on the mainboard socket or mainboard PCB, you may switch between different viewpoints. Find the script here.
Yesterday Apple released its Macs based on G5 processors and immediately claimed that these chips are the fastest on Earth. As the proof of its claims Steve Jobs’ company showed SPECCPU2000 scores obtained on G5-equipped Mac and on the Dell Precision workstation, based on the Pentium 4 3.0GHz CPU. Even leaving aside the fact that Pentium 4 3.0GHz is not the fastest x86 processor from Intel now and the lack of comparisons to AMD processors, there are some other things in Apple’s presentation, which can be called at least “doubtful”. The most “doubtful” decision that Apple made is that it turned off SSE2 use in compiler during compilation of SPECCPU2000 tests. As Intel Pentium 4 relies largely on this instruction set during floating point calculations, it is not a surprise that performance of this CPU dropped dramatically, compared to official SPEC scores. Read about this and other aspects of Apple’s approach of doing comparisons in this article written, incidentally, by a Mac user. There are also letters from angry Mac users about the blurb and its author.
Real World Technologies has posted an article called “Escape From the Planet of x86”. As you can guess from the title, the article is about the history of x86 instructions set, from its introduction in Intel 8086/88 family of processors by nowadays and its battle with RISC solutions. There are also debates about shortcomings of x86 and huge hardships for transition to another possible (and maybe more practical) instruction set, as there is an immense amount of software written for x86 processors. Read the article here.
SloTech.com web-site interviews Richard Huddy, a former Developer Relations Manager for NVIDIA who has been working for ATI Technologies for more than a year now. Richard himself is a very nice person and an excellent presenter, as some of you certainly know. This time he talks on variety of topics from the current state of graphics market in general and ATI Technologies in particular to the long-term future of 3D graphics and prospects of such exotic things like Virtual Reality headsets and ray-tracing. An extremely exciting and useful read can be found over here.