Intel Did Not Invent Hyperthreading

owenrutz · Mar 31, 2005

In: Dueling Multicores: Intel and AMD Fight For the Future. Part 1
It says: HT is actually one of the many efficiencies that Intel engineers invented as a way to combat the perennial problem of latency .

That is an incorrect statement. A software engineer testing CPU performance measurement software on a test motherboard stumbled on to hyperthreading. Intel provided a test motherboard that simulated Semetric Multi-Processing to OEMs so that they could start comming up to speed on SMP technology that would be provided for Pentium I CPUs. The motherboard simulated SMP using one processor because the SMP technology had not yet been built. A software engineer who was testing some CPU performance mesurement software was startled to see a very significant improvement in CPU performance using this simulation. He excitedly told Intel about his discovery only to find out the Intel was horrified at what he had discovered. They spent a tremendous amount of effort trying to hide this discovery. Intel was concerned that this knowledge would provide performance improvements that Intel would not be able to control. They were concerned that it would provide a new over-clocking technology.

fishmahn · Mar 31, 2005

Umm, based on what you said, this software engineer discovered something that Intel had not invented on purpose. Sounds like they did invent it - just someone else discovered that it worked that way.

It could be a case of me reading too precisely what you wrote, however. Do you have a link or something to the real story?

Mike.

owenrutz · Mar 31, 2005

Mike,

I think your description of the situation is somewhat accurate. I'm not a patent lawyer, but I'm sure that is how Intel would explain it.

This happened so long ago, it might be hard for me to turn up a web link to back up my story. A bet Tom Pabst remembers some details about it

I'll do a little Googling and see what I find...

owenrutz · Mar 31, 2005

United States Patent 5,944,816
AMD owns the patent for Hyperthreading?
OK, it may be wierder than I remembered.
Still reseaching...

P4Man · Mar 31, 2005

You are right, intel did not "invent" it. SMT is concept that has been researched since the mid 80's by various university professors.

Intel was also not the first to implement it, as the DEC Alpha 21164 can claim that AFAIK.

Intel was however, the first to implement it in an x86 cpu, or any high volume cpu for that matter.

BTW, the hard part about SMT is not "inventing" it, the concept is pretty simple, it is implementing it and making it work that is pretty damn difficult.

= The views stated herein are my personal views, and not necessarily the views of my wife. =

fishmahn · Mar 31, 2005

Probably nothing because of cross-licensing agreements.

Also, I read that that patent borrowed a bit from a 1994 patent filed by Intel about multi-processing (didn't read what it borrowed, so it may be immaterial to HT).

Mike.

Mephistopheles · Apr 1, 2005

I found that article from THG extremely badly-written. I found several mistakes, and I also thought that putting HT versus dual-core is simply silly and stupid. HT has been around since 2002, and dual core isn't even reality yet for the average user. HT is no direct competitor against dual-core tech, that is simply bullsh1t!

I was so honestly disappointed at that article that I simply refused to read part II of it until now. It sounds too much like flaming and too little like reason to me.

Spitfire_x86 · Apr 1, 2005

Who reads THG anymore?

I'll check these two articles later, though.

------------
<A HREF="http://www.mozilla.org/products/firefox" target="_new">Rediscover the web</A>

Mephistopheles · Apr 1, 2005

Hey, good, I was beginning to get afraid I was the only one thinking that that article was sh!t!

Personally, I've been reading much more xbitlabs than THG for a rather long time now. And I usually take THG with a grain of salt anyway.

mozzartusm · Apr 1, 2005

If he left why do they still call it Toms's?

Intel P4 550(3.4)@4.2 posted 4.8
ASUS P5AD2-E-Prem
Ballistix PC2 5300@DDR2 780
ATI Radeon X800XL 459/609
TT 680W PSU

P4Man · Apr 1, 2005

>AMD owns the patent for Hyperthreading?

No. they own *a* patent regarding a specific SMT implementation (one that doesnt not run 2 software threads like HT, but one regular software thread, and one independant interrupt processing thread). Basically, a simpler version of HT than what intel implemented, since you don't need to duplicate or switch contexts.

These days, patents are such nonsense anyway. You'll find find patents describing SMT from NEC, IBM, HP and SGI (as well as intel and AMD). Intel has filed over 30 (could be more, too bored to go through them) patents which basically describe the same thing (CMT). While browsing, I even found someone that patented a remote control for a drill.

= The views stated herein are my personal views, and not necessarily the views of my wife. =

apesoccer · Apr 1, 2005

Being new to the forum...and the site in general...what happened there? Did he leave from some internal conflict, or did he just look for something new to do?

Current machines running F@H:
AMD: [64 3500+][64 3000+][2500+][2000+][1.3x1][366]
Intel: [X 3.0x3][P4 3.0x2][P4 2.4x5][P4 1.4]

"...and i'm not gay" RX8 -Greatest Quote of ALL Time

slvr_phoenix · Apr 1, 2005

I think Tom left THG because he was too intelligent for people to take. He actually wrote technical articles that were worth reading. :O

<pre>

Yes I'm insane, but it's a good kind of crazy.

</pre>@ 184K -> 200,000 miles or bust!

Ned_Flanders · Apr 1, 2005

What?! There is no Tom at Tomshardware guide?!

______________
Who's the man with the master plan?

slvr_phoenix · Apr 1, 2005

If there is, it's only in an exceedingly silent way. I do wonder though, what Tom's official role in THG is these days. Does he own part (or all) of the company still? Or is he out entirely and they just didn't feel like renaming it?

<pre>

Yes I'm insane, but it's a good kind of crazy.

</pre>@ 184K -> 200,000 miles or bust!

slvr_phoenix · Apr 1, 2005

And what do I have to do to get rid of this <explitive never entered> Eternal Poster title? I'm tired of hanging on someone's wall.

Can't I do something to get past these poster upgrades? At this point even the title of Eternal Loser would be an improvement. When they say eternal, they aren't kidding. I've had this title forever.

(Okay, that's used up my whining allotment for the day. **ROFL**)

<pre>

Yes I'm insane, but it's a good kind of crazy.

</pre>@ 184K -> 200,000 miles or bust!

jg38141 · Apr 1, 2005

Maybe this has something to do with why he left
http://www4.tomshardware.com/blurb/20020825/index.html

My Jesus is whiter than your Jesus.

slvr_phoenix · Apr 1, 2005

<A HREF="http://www.tomshardware.com/blurb/20020825" target="_new">clicky</A>

<pre>

Yes I'm insane, but it's a good kind of crazy.

</pre>@ 184K -> 200,000 miles or bust!

endyen · Apr 2, 2005

Thanks to wusy, I dont know how many posts you are at.
It takes 4600 to get into the forums. I'm over 4550 now, and am looking forward to a time when i wont be a poster child any more. Expect to see a few more pointless posts.

mozzartusm · Apr 2, 2005

The truth is "Im the one that invented Hyperthreading"

Intel P4 550(3.4)@4.2 posted 4.8
ASUS P5AD2-E-Prem
Ballistix PC2 5300@DDR2 780
ATI Radeon X800XL 459/609
TT 680W PSU

slvr_phoenix · Apr 4, 2005

I hear you slvr_phoenix. It's people like you and me who've been here for ages, but don't get fame for being the oldest forum members around.

I never really thought of it that way. I'm just tired of being a 'poster' of one sort or another. It feels like I've been one forever, and I for one, like occasional change. **ROFL**

But yeah, when you put it that way, I have been here an awful while. And it's not like I haven't helped countless people over those years. That itself should count for something. :\

Oh well. Such is life.

As for Tom, yep he left us good alright. I'm 99% sure he has got nothing to do with THG right now.

Except maybe a paycheck from the advertising if he still owns a piece of the pie.

<pre>

I express to you a hex value 84 with my ten binary 'digits'.

</pre>@ 185K -> 200,000 miles or bust!

slvr_phoenix · Apr 5, 2005

Hey. I'll be damned. Either Fredi noticed my whining or I actually managed to get past the poster title at a weirdly coincidental time. :O

<pre>

I express to you a hex value 84 with my ten binary 'digits'.

</pre>@ 185K -> 200,000 miles or bust!

Intel Did Not Invent Hyperthreading

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Splendid

Distinguished

Splendid

Distinguished

Distinguished

Splendid

Splendid

Splendid

Splendid

Distinguished

Splendid

Splendid

Splendid

Splendid

Splendid

Share this page