Sign in with
Sign up | Sign in
Your question

Intels Pentium Performance Hangs on a Hyper-Thread

Tags:
  • CPUs
  • Performance
  • Font
  • Intel
Last response: in CPUs
Share
December 2, 2002 4:19:00 PM

Okay, now at times THG has been a little biased towards AMD, which is fine. But what the heck is up with this new 'article' "Intel's Pentium Performance Hangs on a Hyper-Thread"? Is Bruce Gain a new employee? Is THG <i>that</i> hard up for articles that they print this crap?

I mean hell, let's start with the very first words:
<font color=red>
Quote:
Intel has quietly begun to deviate from its self-perpetuated myth that processor performance is based on clock-speed alone.

</font color=red>First of all, that myth was <i>not</i> perpetuated solely by Intel. Second of all, only <i>now</i> is Intel 'quietly' deviating? Um, excuse me, but what exactly do you think the Celeron is, if not a testament to the fact that clock speed <i>isn't</i> everything when it comes to performance? Intel has been labelling CPUs with codes to denote FSB and cache differences for how many <b>years</b> now? Intel's HT talk is about as 'quietly' as a foghorn. Only someone monumentally biased could even hope write garbage like this.

Then let's get into the truely biased and unprofessional comments like this one:
<font color=red>
Quote:
Intel will likely continue to throw a substantial portion of its billion-dollar marketing budget to perpetuate the clock-speed myth in the average consumer's mind that processor performance is measured by clock speed alone.</font color=red>

I don't recall <b>ever</b> seeing Intel release a million dollar comercial on TV stating something like "It's all about clockspeed, baby." Have you? The last Intel add that <b>I</b> saw was a bunch of stupid aliens using a Pentium 4. So unless just advertising a P4 without even <b>mentioning it's clockspeed</b> is "<font color=red>to throw a substantial portion of its billion-dollar marketing budget to perpetuate the clock-speed myth in the average consumer's mind</font color=red>", Bruce Gain's comments are completely unwarranted.

Here are some others:
<font color=red>
Quote:
But after the debut of the HT technology in its desktop CPU line, incremental performance gains in new Pentium processors will no longer be measured by pure Megahertz levels, Intel said.

</font color=red>Oh really? Intel said <i>that</i>? First off, why no quotation marks then? Second off, what do you call a P3 1000<b><font color=blue>EB</font color=blue></b> or a P4 2.8<b><font color=blue>B</font color=blue></b>? "<font color=red>no longer be measured by pure Megahertz levels</font color=red>" ... When <i>was</i> it measured by pure Megahertz levels?

<font color=red>
Quote:
Simply put, HT allows one processor to serve as two physical processors while the OS and other programs are tricked into thinking that there are actually two working physical processors.

</font color=red>Um, sorry, but <b>WRONG!</b> That was <i>really</i> badly worded. Simply put, HT allows one processor to <i>pretend to be</i> two processors. It's still only one physical processor, but it administers resources to multiple threads better.

Now lets get into the truely ignorant statements that prove Bruce Gain doesn't even have a clue what he's talking about:
<font color=red>
Quote:
some multithreads were added by developers as a convenience for the software vendors when they were debugging their code. Because the debuggers had dual processor workstations and were writing with multiple threads, the debugging process worked faster.

</font color=red>I'm sorry, but this sounds like complete and total bull-drek to me. I've been a programmer since '95 and I've <b>never</b> heard of a single person ever writing multi-threaded code for the sole purpose of making <i>debugging</i> <b>easier!</b> It's ludicrous. Only someone who has never run a debugger in their life would even suggest such a notion! Multi-threaded code can be a massive pain in the arse to debug. It'd add countless time to product development. And why? Just so that the code runs better on the coder's platform instead of the real-world-user's platform? Oh please! Any <i>good</i> software house would do their best to test and debug the code on a system as close to their client's system as possible, not on some fanciful multi-processor system just to see it run faster.

<font color=red>
Quote:
But for some observers, HT will help the Pentium 4 compensate for several of the less-than-stellar performance benchmarks associated with the Pentium 4 vs. the Athlon and the Pentium 3.

</font color=red>That's nonsense. Anyone who has even half of a clue on the subject knows that the P4 suffers from weaker FPU and bitshifting performance compared to a P3 or Athlon. HT isn't going to do jack to make those run faster. And, of course, since Bruce Gain is clearly biased, absolutely no place is "<font color=red>less-than-stellar performance benchmarks</font color=red>" mentioned of a P3 or Athlon compared to a P4 in bandwidth-intensive apps.

<font color=red>
Quote:
Due to the Pentium 4's long pipe architecture, performance has lagged megahertz for megahertz compared to the Pentium 3 and the Athlon for most office applications

</font color=red>Again, nonsense. Yes, there is a performance hit <b>during branch mispredictions</b> because of the P4's long pipe. However, again, compared to horrendous FPU and bit-shifting of the P4, that performance hit from the long pipeline is minimal.

<font color=red>
Quote:
Could HT be Intel's Trojan Horse that will render the Pentium 4's performance commensurate with its clock-speed?

</font color=red>Does Bruce Gain even <i>know</i> what a Trojan Horse is? A Trojan Horse is historically a dangerous object (in the original case a wagon holding armed troops) that is disguised and presented as something friendly (in the original case a statue of a horse). The 'sucker' recieves the 'gift' only to end up impaled later by the 'true' nature of the Trojan Horse. So how in the world could HT be a 'Trojan Horse'? It's true nature most definately isn't being hidden behind some innocent-looking facade. In fact, Intel isn't even trying to be sneaky about it at all. Further, it's beneficial (in theory), not harmful. So how could anyone with half a brain associate HT with a Trojan Horse? There are <b>no</b> similarities.

<font color=red>
Quote:
By definition, parallelism boosts performance by performing independent tasks simultaneously.

</font color=red>Again, close, but no cigar. Parallelism boots performance <b>when running multiple <i>simple</i> tasks</b> by performing independant tasks simultaneously. Parallelism doesn't do jack for single complex tasks that use most (or all) of the available resources, and it doesn't do anything for running several complex tasks that all want to use most (or all) of the available resources. All that it does is make simplistic tasks run quicker when several simplistic tasks are running simultaneously.

<font color=red>
Quote:
The first logical processor can track one software thread, while the second logical processor can track another software thread simultaneously. Because the two threads share one set of execution resources, the HT can use resources that would be otherwise idle if only one thread was executing.

</font color=red>With one massive flaw that is conveniently left out. With HT as Intel currently has implemented, whenever a process is idle because the on-die cache didn't contain the information that it needed and it has to get that information from the system memory, then HT can really screw things up because now you can have TWO idle processes waiting for a new page of memory a lot more often instead of just one as they keep fighting with each other. And with two processes that can be completely unrelated, that's going to happen A LOT because they won't have any memory in common between them.

<font color=red>
Quote:
Users only stand to benefit while the transition to an HT architecture will remain largely transparent to developers.

</font color=red>Clearly Bruce Gain has <i>never</i> written multi-threaded code, as the transition to a multi-threaded architecture is <b>NEVER</b> even <i>minutely</i> transparent to developers. It's a <b>lot</b> of work, and often without benefit unless you know that the code will actually be run on a multiple-CPU system. Otherwise, just using a profiler to maximize your usage of the CPU's resources will be a <b>lot</b> easier <i>and</i> largely transparent to the coder.

<font color=red>
Quote:
However, only WindowsXP recognizes HT in a Windows environment, so while existing multi-threaded applications may be able to take advantage of HT, older version of Windows will not.

</font color=red>To my knowledge, Windows 2000 and Windows NT4, as well as various flavours of Linux and Unix both will all recognize an HT processor as two processors and utilize them (it) correctly. So that means that WinXP is <i>not</i> the only Windows OS, and certainly not the only OS overall.

<font color=red>
Quote:
In summary, with so many performance gains that HT offers independent of clock-speed, perhaps Intel is on the verge of ending its emphasis on the megahertz race with AMD. At the very least, Intel has begun to attempt to convey to the average user that there is more to life than clock speeds alone.

</font color=red>In summary, Intel hasn't pushed clock speed as the <i>only</i> performance factor for <b>years</b>, and only a biased antiquated dinosaur would even suggest as much. Perhaps Intel's flaunting of HT will finally be a loud enough yell that even the deafest most closed-minded of AMD-fanboys will finally have to concede that Intel doesn't push clock speed alone.

And only a complete loser who can't let go of the now distant past would even try to deface Intel with such a lamely written article. Why THG would publish the article of such a loser is only testament to the ever increasing decline of the THG website's credability and professionalism.

<b>To any THG staff:</b>I used to love THG articles. What the hell happened? How could you let things get <i>this</i> bad?

<A HREF="http://forumz.tomshardware.com/community/modules.php?na..." target="_new"><font color=red>Join</font color=red> <font color=blue>the</font color=blue> <font color=green>THGC</font color=green> <font color=orange>LAN</font color=orange> <font color=purple>Party</font color=purple>!</A>

More about : intels pentium performance hangs hyper thread

December 2, 2002 6:08:04 PM

I couldn't even read the whole article...and I noticed that Trojan Horse thing too. I thought maybe there was something I wasn't getting but after reading what you just said I guess I was getting it. I was getting that this guy is an a$$bag.

Those dogs can smell anything. That's why you gotta kick em in the throat!!!
December 2, 2002 6:13:02 PM

Concerning the "megahertz is everything" thing, Intel indeed never claimed that megahertz equals "absolute performance". Intel just never saw any need of wasting time on a "real-world performance rating" which would never be a real-world rating but more likely a real-biased rating.

Years ago people didn't need a performance rating since everyone knew how a P60 would perform compared to a P200 or a PIII 450MHz to a PIII 800MHz. There was absolutely no need for a different (performance) rating other than pure MHz until AMD came in the game and though they needed to change everything. This whole Megahertz myth is of course real but above all a marketing trick used by AMD to be able to sell their lower clocked but higher IPC performing processors. And to get this trick working they invented the megahertz claim of Intel which never really took place...
Related resources
December 2, 2002 6:23:27 PM

I got about 2 paragraphs into the piece before I realized it was a pile o' crap. Fortunately these articles are still more rare on THG than on other sites. Hopefully THG doesn't let the author write-up anything else. Of course, with the obvious flaws/biases, letting him run a benchmark comparison could be humourous at least.
December 2, 2002 7:31:59 PM

Intel spends a significant amount of money educating users that bigger numbers means bigger performance. HT changes the tack the company takes, and it is subtle, and it is a Trojan Horse in that regard.

For the last 4 IDFs that I have been present at, Intel's push has been higher clock speeds.

Intel's tack with HT impacts the way they will market processors through channels and OEMs, and that's where the big bucks are. Bruce addresses these issues in his article.

And, how you can call Bruce's article biased is beyond me. He is a respected writer who has written for Electronics Buyers News, and other CMP electronics publications so, I get a little annoyed with people just trying to tar and feather an author because they don't like what he says.

Omid Rahmat
General Manager
Tom's Guides Publishing LLC
December 2, 2002 8:01:35 PM

Slvr I dunno but I think you were not informed on the fact that ONLY WindowsXP supports and fully manages HT as it should. In Windows 2000 you can go as much as lose 50% of performance, as proven in an article out there in the past weeks from some website.

--
The THGC Photo Album project,<A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&..." target="_new">CLICK HERE!</A>
December 3, 2002 1:03:57 AM

I've heard a lot about it in the past while, and have come to believe that their is no performance difference on single CPU systems.

But,

Win2000 doesn't differenciate between physical processors and logical processors. Since Win2000 (non-server) will allow up to two processors, it's happy. However, on a Xeon DP system there are FOUR logical processors and so Win2000 only uses the first two logical ones it finds (which will be the two logical processors that reside within physical processor #0). Physical processor #1 will get completely ignored by the OS, and you will lose 50% of your performance.

WinXP knows the difference between physical and logical processors and will allow up to two physical processors, and (IMHO) an unlimited number of logical processors. So both your Xeon DPs get used to their full extent.

*Dual PIII-800 @900 i440BX and Tualeron 1.2 @1.74 i815*
December 3, 2002 1:10:22 AM

Yes however the benchmark lost a huge ~50% in performance while HT was enabled, and that's on a 3.06GHZ single setup system!
It proves clearly that there is no other OS out there at the moment that can manage HT like WinXP.
I would assume MS can have a service pack for Win2k which will tweak it and optimize it for that, except I'd ask you all to switch CAN to WILL, and turn that into a question. You'll answer yourself.

--
The THGC Photo Album project,<A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&..." target="_new">CLICK HERE!</A>
December 3, 2002 1:58:22 PM

Quote:
Intel spends a significant amount of money educating users that bigger numbers means bigger performance.

Really? Intel spends money for the sole purpose of educating users that bigger numbers means bigger performance? They just go out of their way to host classes for Joe Schmoe off of the street to educate him that his Pentium 3 733EB is slower than Bob's (his neighbor) Celeron 750? Or perhaps you mean the stupid alien commercials that don't even talk about clockspeeds? Please do tell more.

Quote:
HT changes the tack the company takes

HT is just one of <i>many</i> things that Intel has done to state that MHz alone isn't an end all in performance factors. How else can you explain Celerons? How else can you explain cache size differences? How else can you explain FSB speed differences? I'm sure that Intel would most definately agree that a 750MHz Celeron is slower than a P3 733EB, so please do explain how HyperThreading is the very first indication that Intel has <i>ever</i> given that clock speed alone is not in fact the end all of performance.

Quote:
and it is subtle

Really? Please also do explain how Intel is being in the lest bit subtle about HyperThreading.

Quote:
and it is a Trojan Horse in that regard.

Are you familiar with the phrase 'Trojan Horse'? If so, please do support your statement.

Quote:
For the last 4 IDFs that I have been present at, Intel's push has been higher clock speeds.

No offense Omid, but I do believe that half of the the <i>point</i> of IDF is for Intel to flaunt the fastest CPUs that they desire to. Would you even want to go to an IDF where Intel says "You know, we <i>could</i> have demoed an uber-fast Pentium for you to salivate over, but instead we decided to bake cookies. So the only Pentium we have is the same one you saw at the last IDF. So sorry." and then doesn't even hand out the cookies?

Quote:
Intel's tack with HT impacts the way they will market processors through channels and OEMs, and that's where the big bucks are. Bruce addresses these issues in his article.

Please read Bruce's article and then give me even two quotes where "<font color=red>Bruce addresses these issues in his article.</font color=red>" or show me even once mention of marketting "<font color=red>processors through channels and OEMs</font color=red>".

Quote:
And, how you can call Bruce's article biased is beyond me.

Because it most definately, without a doubt, was. Just read everyone elses replies. Obviously I'm not the only one who saw it. If that is beyond you...

Quote:
He is a respected writer who has written for Electronics Buyers News, and other CMP electronics

That's nice. However, I don't care if he's a respected writer who has written for Forbes or the NY Times. Crap writing is crap writing, and that was a horrible article that he wrote, end of story. Everyone has a bad day. To err is human. Maybe he normally writes much better. I don't know, nor do I really care. My point is that particular article was biased poorly-researched badly-written crap. So why is THG publishing this kind of stuff?

If you can give a counterpoint to each and every point that I made at the start of this thread that would stand up in a debate for even a minute, then I'll gladly state that I was wrong. If, however, you won't (or can't) then I suggest you give the article an honest read without prejudice before you tell me that it isn't just plain crap.

Quote:
so, I get a little annoyed with people just trying to tar and feather an author because they don't like what he says.

It has nothing to do with not liking what he says. It has everything to do with the simple facts that he:
1) Specifically picked and chose what information he included in the article to put Intel into a much worse light than they deserve. (In other words, he wrote a very biased article.)

2) Didn't do a single bit of research. Many of his claims are completely unfounded and dare I even suggest whimsical, not fact. Simple research proves many of his statements completely wrong.

He could have talked about things that I don't like until he's blue in the face (in the fingers?) for hours on end for all that I care, so long as his information was accurate and he didn't intentionally skew facts. It isn't about opinions. It is about professionalism and the quality of the article.

As far as I can see, all that his article really did was poorly explain HT amidst a huge stinking pile of biased dung. Had the article been filed away under the 'Columns' section, I might have even let it pass without comment, since that section is just for ranting anyway. (As you no doubt obviously know.) However, it was placed under the CPU section as though it was a professional article.

If this is THG's new idea of professionalism, then I might as well go to <A HREF="http://www.amdmb.com" target="_new">AMDMB</A>.

<A HREF="http://forumz.tomshardware.com/community/modules.php?na..." target="_new"><font color=red>Join</font color=red> <font color=blue>the</font color=blue> <font color=green>THGC</font color=green> <font color=orange>LAN</font color=orange> <font color=purple>Party</font color=purple>!</A>
December 3, 2002 2:14:21 PM

Eden, I know that months ago I had read reviews on HyperThreading that showed both Win2K and WinNT running HT Xeons just fine so long as your license was for as many <i>logical</i> processors that are in the system.

Further, since HT is a strictly hardware issue, the <i>only</i> software requirements that it has is to identify the difference between physical and logical processors so that software doesn't have to be licensed for twice as many processors as you physically have installed. There should be absolutely no code needed to run HT 'properly', as it is handled all in hardware. <i>Any</i> multi-threaded applications (including OSes) should run both normal processors and HT processors just fine. (So long as the software is compiled for that particular processor type anyway. Obviously a multi-threaded Mac app won't run on an x86 CPU.)

All of the research that I have done on the issue has said thus. However, if you can provide links to material proving otherwise, I would love to read it. :) 

[EDIT]Of course the possability does exist that when running single-threaded applications on a Win2K dual-HT system the OS could be trying to apply the first process to the first CPU and the second process to the logical clone of the first CPU <i>if</i> the hardware/firmware was designed to hand out the processors in the order of PHYSICAL - CLONE - PHYSICAL - CLONE instead of PHYSICAL - PHYSICAL - CLONE - CLONE. Again though, that's more a hardware/firmware issue than a software issue because a software workaround <i>shouldn't</i> be specially coded to work around that as it <i>should</i> be handled in hardware/firmware. And anyone who designed a motherboard that hands out the processors in the order of PHYSICAL - CLONE - PHYSICAL - CLONE did something really stupid.[/EDIT]

<A HREF="http://forumz.tomshardware.com/community/modules.php?na..." target="_new"><font color=red>Join</font color=red> <font color=blue>the</font color=blue> <font color=green>THGC</font color=green> <font color=orange>LAN</font color=orange> <font color=purple>Party</font color=purple>!</A><P ID="edit"><FONT SIZE=-1><EM>Edited by slvr_phoenix on 12/03/02 10:39 AM.</EM></FONT></P>
December 3, 2002 4:38:20 PM

But software is, and indeed does need to be specially coded. WinXP recognizes the difference between physical and logical processors and allocates threads accordingly. Also, when a second logical processor is working on a thread, it takes up half the re-order buffer and several other queues which can hamper the first thread. It's the OS's responsibility to issue a HALT command to the second thread whenever it feels that the first thread needs priority (or the second thread is idle). That is the main part of "optimized for HT".

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
December 3, 2002 5:07:38 PM

They will take care of it see <A HREF="http://biz.yahoo.com/bw/021203/30076_1.html" target="_new">Here</A>

-Jeremy

<font color=blue>Just some advice from your friendly neighborhood blue man </font color=blue> :smile:
December 3, 2002 6:28:52 PM

It still sounds to me like software doesn't <i>need</i> HT optimization, but it can <i>benefit</i> from it. The fact that code would even benefit from special consideration for HT though means that Intel could have implemented it better.

<A HREF="http://forumz.tomshardware.com/community/modules.php?na..." target="_new"><font color=red>Join</font color=red> <font color=blue>the</font color=blue> <font color=green>THGC</font color=green> <font color=orange>LAN</font color=orange> <font color=purple>Party</font color=purple>!</A>
December 3, 2002 6:41:28 PM

Although I was about to go look for URL, I was hoping you might come in and lay out your explanation. For now this is what I've been trying to say, though unprofessionally :wink: !
I'll go check for the benchmarks under Win2000.

--
The THGC Photo Album project,<A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&..." target="_new">CLICK HERE!</A>
December 3, 2002 6:52:05 PM

Intel did a whole powerpoint presentation IIRC just to show people that coding with HT optimization can yeild very interesting results. I don't remember the link to it, but it was on Intel's website for sure, in PDF format.

--
The THGC Photo Album project,<A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&..." target="_new">CLICK HERE!</A>
December 3, 2002 6:58:17 PM

I like your spin on multi-threading and programming. Anyone who has ever tried to step over a "<font color=red>lock</font color=red>" instruction, knows exactly where you're coming from.

Dichromatic for your viewing plesure...
December 3, 2002 7:12:37 PM

Ok there's the site, yes translated so it has MANY faulty words and translations, but you get the benches: <A HREF="http://babelfish.altavista.com/babelfish/urltrurl?lp=fr..." target="_new">http://babelfish.altavista.com/babelfish/urltrurl?lp=fr...;/A>
If it does not open, you'll need to manually scroll through the first pages until you reach COMPATIBILITY HARDWARE SOFTWARE: <A HREF="http://babelfish.altavista.com/babelfish/urltrurl?lp=fr..." target="_new">http://babelfish.altavista.com/babelfish/urltrurl?lp=fr...;/A>

--
The THGC Photo Album project,<A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&..." target="_new">CLICK HERE!</A>
December 3, 2002 7:30:34 PM

*snoooooooooooooooooooooooooooooooooooze*

Dude, that's one hell of a rant.


"I personally think filesystems should be rewritten from scratch every 5 years..." --- Hans Reiser
December 3, 2002 7:37:42 PM

That's disturbing that special software consideration is needed, but again, not surprising. I've said from the beginning that Intel's implementation of HT is not impressive. HT is a nice enough of a concept I suppose, but it still needs work.

What's also disturbing is that in the middle of the page, the text went from bad English to (I believe) French right in the middle. What's weirder is how much of it I could read anyway, even though I only know English. Heh heh.

Buy anywho, that aside, it still sounds like the only software changes that have to be made at all are in the OS, not in the applications themselves. And a simple patch would fix Win2K or NT4. Heck, I didn't even see mention of a Win2K patch number, so I'd say that the possability exists (however unlikely) that MS has already fixed Win2K with a patch.

And even still, HT still runs on Win2K (and no doubt NT4). It just runs badly. Heh heh.

Thanks for the link though. :)  It figures that you have to leave the English language just to find information like that. Heh heh.

<A HREF="http://forumz.tomshardware.com/community/modules.php?na..." target="_new"><font color=red>Join</font color=red> <font color=blue>the</font color=blue> <font color=green>THGC</font color=green> <font color=orange>LAN</font color=orange> <font color=purple>Party</font color=purple>!</A><P ID="edit"><FONT SIZE=-1><EM>Edited by slvr_phoenix on 12/03/02 03:43 PM.</EM></FONT></P>
December 3, 2002 7:55:01 PM

That's what I've been suggesting.
MS could do a patch to properly support HT on Win2K. But then I ask you to replace COULD by WILL, put it before MS, and turn that into a question sentence. You'll answer your own self lol.

--
The THGC Photo Album project,<A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&..." target="_new">CLICK HERE!</A>
December 3, 2002 11:42:49 PM

and slvr_phoenix, No one cares about you.

hehe j.k<P ID="edit"><FONT SIZE=-1><EM>Edited by ra5555 on 12/03/02 08:47 PM.</EM></FONT></P>
December 4, 2002 12:39:50 PM

Considering the number of corporations using Win2K, you would think that MS <i>would</i> release a Win2K patch. Then again, considering that MS wants everyone to migrate to XP, you would think that they wouldn't. So really, I think it's a toss up that depends entirely on how much Intel wants to push MS. Maybe the toss up is not quite a 50/50 balance, but at least 40/60.

<A HREF="http://forumz.tomshardware.com/community/modules.php?na..." target="_new"><font color=red>Join</font color=red> <font color=blue>the</font color=blue> <font color=green>THGC</font color=green> <font color=orange>LAN</font color=orange> <font color=purple>Party</font color=purple>!</A>
December 4, 2002 12:46:29 PM

Quote:
I like your spin on multi-threading and programming. Anyone who has ever tried to step over a "lock" instruction, knows exactly where you're coming from.

Not just that (though that's a massive reason in and of itself) but also anyone who has ever had timing problems that sometimes happen and sometimes don't. Or worse, bugs that suddely appear when run on a faster or slower machine than you wrote/debugged the code on that just don't happen on your machine because, again, of timing problems. Writing multi-threaded code is just not something done lightly. <i>Anyone</i> who has ever worked on coding a serious multi-threaded app would know this. Bruce Gain appearantly (observation made from his own words) doesn't know this.

The very first rule of writing is write about what you know. We can clearly deduce that he <i>didn't</i> know about coding multi-threaded applications. So why did he write about it?

That's what bugs me. He wrote an awful lot of comments on things he seemed (as observed by his writing) to have had no actual knowledge of. A little research could have made a world of difference to that article.

<A HREF="http://forumz.tomshardware.com/community/modules.php?na..." target="_new"><font color=red>Join</font color=red> <font color=blue>the</font color=blue> <font color=green>THGC</font color=green> <font color=orange>LAN</font color=orange> <font color=purple>Party</font color=purple>!</A>
December 4, 2002 1:04:27 PM

Quote:
and slvr_phoenix, No one cares about you.

Heh heh. Which is fine. :)  I'm not so great of an ego-maniac that I think people <i>must</i> care about me.

However, I would <i>think</i> (and maybe I'm just nuts here) that THG <i>should</i> care about the quality of it's articles. This particular article's quality was <i>very</i> poor. Yet the one THG staffer, Omid Rahmat, instead of being appalled by the poor quality of the article, or instead of expressing concern that so many THG readers all agreed that the quality was so bad that they couldn't even stand to finish reading the article, he actually tries (and lamely I might add) to <i>defend</i> the article. <i>Worse</i>, he tries to twist it into just a personal attack against the author of the article instead of an expression of problems with the article itself and the fact that THG published it.

As an official representative of THG, is this what we are to believe THG's policy towards articles is? That they publish and defend badly-written and poorly-researched pieces of crap? Are we to expect more and more articles from THG to be <i>that</i> bad? Has THG gone down the tubes <i>that</i> much? If Omid is any indication, appearantly so.

<A HREF="http://forumz.tomshardware.com/community/modules.php?na..." target="_new"><font color=red>Join</font color=red> <font color=blue>the</font color=blue> <font color=green>THGC</font color=green> <font color=orange>LAN</font color=orange> <font color=purple>Party</font color=purple>!</A>
December 4, 2002 1:59:38 PM

Though I agree that the article wasn't exactly of the highest quality, I think you guys may be taking it a little bit out of context. The "Megahertz Myth" is not especially something produced my intel as much as it is by people new to computers who only see the GHZ number as an indication of speed. I think the author is trying to suggest that intel's marketing may shift and not focus quite as much on the actual clock speed. Highest clock speed has really been their selling point to the average user for years, but now they may need to seguay into other things. Without trying to nitpick specific situations, we all know that the P4 is a somewhat inefficient processor. Basically, the author is saying that intel is moving away from trying to just achieve high clock speeds, and if focussing on efficiency as well. Please don't hit me with a bunch of numbers showing about how intel is efficient in this case and this one and that one, or their exact spendings on certain types of marketing. I know you have that capability, but please refrain. :wink:

Overall, I agree with nearly all of your arguments against the background information in the article, but I think perhaps we should not try to over analyze it. Just take it for what it is (whatever you may think of it).

<font color=blue> If it ain't broke, don't fix it...tweak it.</font color=blue>
December 4, 2002 8:00:02 PM

I must agree with you here on the MHZ thing. By looking at the roadmaps as well, every clock speed ramp for the P4 has become very stretched on the date. It is clear HT is there for a reason.
I have always seen Intel as MHZ means all, and that is IMPLICITLY saying.

However true, no add has ever said it, I say once more as proof by the consumers, that they fall for MHZ and Intel when it comes to new computer. Otherwise Intel could have slowly ramped speeds. We've had in one year, what, a full 1GHZ ramp if not more!
Athlon has also climbed up 1GHZ, but not even 1GHZ to be exact.

However I should say one thing, if anything, Intel is finally being more honest about the REASONS to get a P4.
Did you guys see the new YES commercials on TV?
The one with the Moby song playing? I must say this is the best Intel add yet and I have in fact a lot of respect for it. It shows all usage that is intensive such as image and video encoding, music editing and such. All things that P4s do indeed excel at. And the question in the end "Does a better computer change your life"? The answer is YES by them and me, and that is why I think Intel has really improved by showing the real usage and power of the P4, not just by Aliens modifying a picture of someone...

So while the MHZ myth was never explicitly indicated by Intel (ALTHOUGH, I do recall once where they put Sanders in front of an Intel cronie and the guy supposedly spewed major FUD that MHZ is everything), it is certainly acquainted by the community of the consumers, and that, nobody can defend the fact that it isn't associated to Intel.

--
The THGC Photo Album project,<A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&..." target="_new">CLICK HERE!</A>
December 5, 2002 9:27:46 PM

I'm glad I'm not the only one who thought this article was totally biased and poorly written. I thought you expressed your arguments very eloquently.

And Mr. Rahmat, maybe you should be a little more open minded to what is being stated here instead of getting so defensive.
December 6, 2002 3:22:08 PM

Completely agree with slvr_phoenix...this is a terrible article, definitely not professional quality, looks more like something taken from Van's Hardware than from Tom's.

Someone should check Bruce Gain's his articles before posting them on the webpage. If he doesn't know what a Trojan Horse is, then he shouldn't even mention it.

__________________________________________________
It's not important to know all the answers, as long as you know how to contact someone who does.
December 6, 2002 6:59:28 PM

I somehow thought the Trojan Horse reference was adequate. HT is quite a surprise nevertheless, the multitasking power it provides is phenomenal and can be up to 5 times faster in the OS than a non-HT CPU that is higher clocked.
I dunno guys but I think the TH reference was ok, didn't think it was out of place.

--
The THGC Photo Album project,<A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&..." target="_new">CLICK HERE!</A>
December 6, 2002 7:32:23 PM

Quote:
I dunno guys but I think the TH reference was ok, didn't think it was out of place.

Okay then Eden, explain how Intel yelling Hyper Threading from the rooftops, a technology which (theoretically) improves performance, is in any way akin to a bunch of soldiers sneakily hiding in a fake statue with the false pretense of being a tribute to superior foes, waiting for the enemy to pull them inside of the city walls so that they can kill sack the town from the inside.

<A HREF="http://forumz.tomshardware.com/community/modules.php?na..." target="_new"><font color=red>Join</font color=red> <font color=blue>the</font color=blue> <font color=green>THGC</font color=green> <font color=orange>LAN</font color=orange> <font color=purple>Party</font color=purple>!</A>
December 7, 2002 1:54:26 AM

Words DO change their meaning over time now!
Some words had a totally different meaning back in the sixties than now.
Several proper names have been used in a phrase despite not being too related.

--
The THGC Photo Album project,<A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&..." target="_new">CLICK HERE!</A>
December 7, 2002 4:51:15 AM

That was quite a detailed response. I do not have that much to say other than I was a bit disappointed. First, I left with more questions then anything else. Second, what was the darn point of this article? I don't necissarily think that this was a dilliberate defacing of intel though, slvr_phoenix. Personally I think this was nothing more than a poorly written article. I would like to see some sort of follow up written to this article.


-=-Sean-=-
I hate picking fights... but it's so fun :smile:
December 7, 2002 1:25:03 PM

Lets pretend that I (the consumer) represent the Trojans and that Intel represents the Greeks. Now Intel wants my money represented by Hellen (she's sooooo hot!!!). So what does intel do? They hook me up with the so called Trojan Horse aka a P4 with Hyper-Threading. Now Hyper-Threading looks really great, so being the kickass Trojan that I am I say, sh!t, bring that dope-a$$ Hyper-Threading all up in here (here being my computer). Then I get totally trashed and pass out and what happens? If Hyper-Threading were a Trojan horse, at this point it would suck and not work well at all but WAIT, Hyper-Threading doesn't want to pillage, burn, and rape the village that is my PC, it wants to make it better.

So exactly how is Hyper-Threading a "Trojan Horse?"

Those dogs can smell anything. That's why you gotta kick em in the throat!!!
December 7, 2002 7:39:09 PM

If it meant anything else it wouldn't make sense. A Trojan Horse is a Trojan Horse. The story doesn't change. So to use the term in any other way is incorrect and misleading.

Those dogs can smell anything. That's why you gotta kick em in the throat!!!
December 7, 2002 11:04:31 PM

Ok so drinking a Bloody Mary must remind you of the history of a woman who was quite ravenous?
Why do we call them Deviled eggs if there is no devil-relation or 666 surrounding them?

The point is words or names don't have to relate to their initial departure and that's how I see that expression used as inoffensive and rather good.

--
The THGC Photo Album project,<A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&..." target="_new">CLICK HERE!</A>
December 8, 2002 12:20:02 AM

so what does Trojan Horse mean these days Mr. Webster? I'd like to hear the new definition.

Those dogs can smell anything. That's why you gotta kick em in the throat!!!
December 8, 2002 12:24:23 PM

To me it sounded like it's their new "stealth" weapon. Does the average consumer hear about it or know it's there?
A consumer can know he has SSE2 on his Pentium 4, as it can be seen in the apps he might use, as an option.
It is their Trojan Horse because they've done something that's quite smart too, they tried to push more parallelism into the x86 without taking much die space.
So it's smart, it is quite stealthy, and surprises even the fastest CPU out there in multitasking.
That's how I see it.

--
The THGC Photo Album project,<A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&..." target="_new">CLICK HERE!</A>
!