Author: Matt Taylor
Date: 23:01:04 12/02/02
Go up one level in this thread
On December 02, 2002 at 09:20:05, Robert Hyatt wrote: >On December 02, 2002 at 02:59:19, Gian-Carlo Pascutto wrote: > >>On December 01, 2002 at 23:02:44, Robert Hyatt wrote: >> >>>Doesn't make _me_ wonder at all. Perhaps they had a prototype chip that was >>>not stamped with any sort of identifier at the time... I can think of several >>>reasons, none of which would discredit text. >>> >>>Only "children" pay more attention to the pictures than to the words written, >>>so most of us wouldn't consider that a big deal. I personally don't pay much >>>attention to photos of motherboards, chipsets, processors, etc. I consider it >>>a high-bandwidth waste of time... >> >>If it is clear that the pictures were messed with without mention, >>you grow suspect of the rest of the article as well. If we can't trust >>the picture, how are we supposed to trust te rest of the data we can't >>verify? > >For the reasons I gave... > >> >>It's like writing a paper on parallel searching and forgetting to >>mention you calculated instead of measured half of the data. > >First I measured the _important_ data, the speedup numbers. > >Second, I don't see anything related in the two. A photograph has _no_ >technical content. IE what can you tell by looking at the top of one of my >xeon 700's? Absolutely nothing. > >I think this concept of "amateurs" questioning "professionals" is a bit beyond >credibility. The folks at Tom's Hardware could hardly get away with falsifying >results. I'm sure AMD would be ready to launch a legal action immediately >should they feel wronged. They don't. So the conclusion is pretty obvious. > >IMHO of course. > >But you and others that post this stuff about "that can't be trusted" ought to >look at yourselves very carefully. That's _not_ very scientific, IMHO... Neither are 90% of the benchmarks that get quoted. A scientific benchmark would also quantify all variables and make attempts to rationalize the results with existing data. I have seen few articles that do this. In total, I have seen -1- well-written article favoring P4 and -2- well-written articles favoring Athlon, 1 written prior to P4's release. Does THG even attempt explanation of the benchmarks? I think not. Then again, I've never seen anyone explain why Quake is a good CPU benchmark... The most damaging evidence is visible in the following two links: http://www17.tomshardware.com/cpu/01q3/010917/heatvideo-01.html http://www.vanshardware.com/articles/2001/october/011031_AMD_Video_Response/011031_AMD_Video_Response.htm In a nutshell, Tom's Hardware has a video that shows the Athlon and AthlonXP (Palomino) failing when the heatsink is remove. The Pentium 3 shuts itself down, and the Pentium 4 continues operating, albeit rather slowly. The link on Van's Hardware quotes an AMD spokesperson and completely contradicts the THG findings. At worst, THG is guilty of deception. At best, they are guilty of sloppiness and lack of research. Either case does not paint them favorably.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.