Question Anybody remember the Pentium III 1133?

Hulk

Diamond Member
Oct 9, 1999
3,362
859
136

AMD had Intel on the run with the Athlon. There was a race to 1GHz, then there was a 1100 Athlon and then a Pentium III 1133 appeared. But it was unstable and they all had to be recalled. Intel is rarely behind but when it does happen they tend to get desperate and the knee jerk reaction seems to be to pump up the clocks. The Pentium III 1133 and then later the fire breathing Prescott Pentium 4's.

The upcoming Comet Lake parts are supposed to go to 5.3GHz with over 300W right? Kind of makes you wonder.
 
  • Like
Reactions: lightmanek

DrMrLordX

Lifer
Apr 27, 2000
19,155
7,915
136
Tom's covered it pretty extensively (back when I used to read them regularly):

 

Topweasel

Diamond Member
Oct 19, 2000
5,414
1,606
136
Tom's covered it pretty extensively (back when I used to read them regularly):

It always bugged me that Tom got credit for it. It took a lot of work from Kyle Bennett and another guy before Tom was convinced. He was just the biggest at the time to write an article.

Maybe its some bias on my part because of the obvious bias in his 24/7 webstreamed torture test.
 

AnandThenMan

Diamond Member
Nov 11, 2004
3,920
414
126
Maybe its some bias on my part because of the obvious bias in his 24/7 webstreamed torture test.
I remember that but am fuzzy on the details is that where the Intel motherboard blew up complete with a smoke show? IIRC the AMD system never skipped a beat the Intel box was down several times.
 

Avalon

Diamond Member
Jul 16, 2001
7,547
110
106
Yep, I remember the recall. That was wild. Some time shortly after, they introduced the Tualatin core, which was a die shrunk Coppermine, and those were able to reach those speeds and more, up to 1.4Ghz I believe? I had one for a short while because I was interested in the fact that it was an uncommon chip and nearly as fast as the best Pentium 4/Athlon chips out at the time. It was almost a very early preview of Core.
 
  • Like
Reactions: lightmanek

Topweasel

Diamond Member
Oct 19, 2000
5,414
1,606
136
Yeah the math in my head isn't perfect. But if I recall Toms set up the test in a really weird fashion.

They rightfully tried to minimize differences. But the goal off the bat was obvious what they were trying to do. So the hardware was the best EE and FX at the time just can't remember if it was a P4 w HT and a FX A64 or if it was a P4D EE and A64x2 FX. Both running Nvidia Nforce boards because it was available on both sides. A marginally large PSU. 2 Video cards.

So the two things Tom was trying to prove was 1. That the Pentium with HT was going to be better over time because it was superior, so they loaded more tasks then AMD had cores for. Think he was tired of everyone thinking that just because AMD would win a benchmark or two, HT was the savior that made "actual computing" better. 2. That a strained AMD setup was going give up the ghost and prove that AMD wasn't near Intel in terms of a stable platform.

The first thing was apparent. The AMD setup was staying ahead and gaining ground every minute in the test.

The second thing that happened was that within a day, the P4 nuked a PSU. (that was the onscreen explodey thingy). They swapped it out with a real oversized PSU for sake of finishing the test after that.

After that the P4 saw significant downtime. Eventually they swap out the Nforce board for a more limited Intel board (with no SLI support), pulled the extra card. Reset the test for "fairness" sake.

The conclusion was that while AMD got more done there wasn't much to infer from that because it wasn't a realistic use case. That both platforms once running on proper platform were good reliable platforms. But AMD being at the mercy of Nvidia was proof because of the Nforce board failure on the Intel system that they didn't have a proper platform.

This was the same site that made a big deal about how easy it was to kill an AMD CPU, 1 gen into Intel having some basic temp protection. Recording themselves removing a heatsink from and Athlon and Pentium IV while it was running and having the AMD CPU pop. It was my final straw with the site and probably adds a bias towards that heat test and the PIII 1.13GHz recall. If my memory serves me correctly it, one site couldn't get the CPU to compile a linux kernal. He got in touch with Kyle who said he had a few oddities with his. Kyle went a little more indepth and identified as a CPU that was overclocked to much (since this is HardOCP, he had lots of experience with that). Tom had been pretty bad about calling a fourth reviewer out (also can't remember) about just maybe getting a bad chip and maybe not knowing how to setup for the tests. Kyle not liking a fellow reviewer making an idiot of himself without all the information got a hold of Tom, told him hey before you write anything more on the issue I think there is really something wrong with them. To prove the point Kyle other guy (linux kernal guy) sent Tom their CPU's. Tom tested them wrote the famous article and published it ahead of the other 2 and was by far the biggest site.
 

DrMrLordX

Lifer
Apr 27, 2000
19,155
7,915
136
It always bugged me that Tom got credit for it. It took a lot of work from Kyle Bennett and another guy before Tom was convinced. He was just the biggest at the time to write an article.

Maybe its some bias on my part because of the obvious bias in his 24/7 webstreamed torture test.
Kyle and Anand had a hand in it. At a minimum.

@Topweasel

I thought the Pentium III 1.13 article at Tom's was funny since so many people (at the time) saw Tom's as being biased against AMD. And yet he really gave it to Intel over that chip.
 

Topweasel

Diamond Member
Oct 19, 2000
5,414
1,606
136
Kyle and Anand had a hand in it. At a minimum.

@Topweasel

I thought the Pentium III 1.13 article at Tom's was funny since so many people (at the time) saw Tom's as being biased against AMD. And yet he really gave it to Intel over that chip.
I put it on him being made to look like a fool. You see it from a tuber from time to time. Say they aren't bias, say they are always nuetral. But when the one side they support has an issue that really erks them they explode because they had all the faith in them and it was destroyed. My point with tom was the build up, I think the fourth one, the review that Tom went after behind the scenes might have been the Anand one. Pretty sure it wasn't them that sent the cpu to Tom's (the linux compile tester). What's odd is usually when this happens they take a step back. I am sure Tom thought Intel was a cut and dry better business, like a Lockheed martin or something, something so rigorous, so careful, so good like a personification of the whole no one gets fired for purchasing an Intel CPU slogan. It's that way because Intel is just that good of a company. So I don't think he was a shill like we think of them so much, he was to attached to Intel just this entity and I think the other half that fed his issues is that while he respected AMD's performance, they were still this company that at its heart was duking it out with Cyrix's and Centuars, that barely survived by buying a company for the K6 design. They just couldn't be as professional as Intel (I would say that was a good thing). But on that, he "found" the issue, talked to Intel, they recalled, and he gave them a tongue lashing. To him it might have been a kink, but I think by immediately recalling a CPU that only had shipped to reviewers at that point, it actually helped cemented that they were this super professional company that just had a QC lapse.

To me is was like in a 3 year period. His defense of a product that had to be recalled till he was hit over the head with how bad it was. The bashing of AMD for not having thermal protection on their CPU's 1 month into Intel first offering a CPU that did. Then the Torture test. The last being the last time I read anything of his and completely missed when the company was sold. CPU's are the one tech, I don't get to deeply into it though, that feeds my desire for tech knowledge. I just couldn't look at them the same. I got more attached to Anandtech after that.
 

zir_blazer

Golden Member
Jun 6, 2013
1,084
322
136
This was the same site that made a big deal about how easy it was to kill an AMD CPU, 1 gen into Intel having some basic temp protection. Recording themselves removing a heatsink from and Athlon and Pentium IV while it was running and having the AMD CPU pop.
Around a decade ago (When people STILL said that AMD Processors overheated and burned. That damn video did years-lasting damage) I looked around for info about that matter, and on contemporary Hardware forums, the most knowledgeable users claimed that the video was a fake set up. On the famous video with Quake 3 running, when they removed the Heatsink on the Pentium 4, it throttled with Q3 massively slowing down, then they plugged the Heatsink back on it again and it resumed proper operation instantly. When they removed the Heatsink, they passed over it an infrared thermometer tool that measured the Willamatte to be around 29°C or so, which is physically impossible. Besides, thermal throttling required to be around 100°C or so before kicking in. There was no way that the temperature was real.
They also did a later test (I think that they did both an Athlon Thunderbird and later an Athlon XP Palomino) where an AMD Processor burned after they removed the Heatsink. I think that the Athlons XP introduced a better shut down system than previously available, but it required Motherboard support (The only one available was a Fujitsu-Siemens one) yet Tom's Hardware used a random one.
 

lightmanek

Senior member
Feb 19, 2017
332
607
136
Around a decade ago (When people STILL said that AMD Processors overheated and burned. That damn video did years-lasting damage) I looked around for info about that matter, and on contemporary Hardware forums, the most knowledgeable users claimed that the video was a fake set up. On the famous video with Quake 3 running, when they removed the Heatsink on the Pentium 4, it throttled with Q3 massively slowing down, then they plugged the Heatsink back on it again and it resumed proper operation instantly. When they removed the Heatsink, they passed over it an infrared thermometer tool that measured the Willamatte to be around 29°C or so, which is physically impossible. Besides, thermal throttling required to be around 100°C or so before kicking in. There was no way that the temperature was real.
They also did a later test (I think that they did both an Athlon Thunderbird and later an Athlon XP Palomino) where an AMD Processor burned after they removed the Heatsink. I think that the Athlons XP introduced a better shut down system than previously available, but it required Motherboard support (The only one available was a Fujitsu-Siemens one) yet Tom's Hardware used a random one.
Very true!
I personally have gamed for two weeks on an Athlon or Duron CPU running at 105C without realising it. Only by accident, one day I went to BIOS to check settings when I noticed my CPU temp to be hovering around 100C! Took radiator off and realised I left protective film on the radiator base. CPU was fine!
 

Hulk

Diamond Member
Oct 9, 1999
3,362
859
136
Intel has twice resorted to increasing clockspeeds to unrealistic levels to respond to AMD. First with the PIII 1133 against the Athlon and later with P4 against the Athlon. It seemed desperate.
Ryzen is more of a threat than the Athlon because Intel has no "Conroe" response as far as we know. Desperate times require desperate measures. If the tide start to move in AMD's direction it's hard to make it swing back.
We are living in interesting times.
 
  • Like
Reactions: whm1974

Topweasel

Diamond Member
Oct 19, 2000
5,414
1,606
136
@Topweasel
You still have a link for that HT vs AMD benchmarking? I don't remember that.
Had to hunt it down. A few things are a little different then I remember it.

 

Insert_Nickname

Diamond Member
May 6, 2012
4,580
1,163
136
Some time shortly after, they introduced the Tualatin core, which was a die shrunk Coppermine, and those were able to reach those speeds and more, up to 1.4Ghz I believe?
Yes, it topped out at a 1.4GHz model.

Tualatin. That was a beastly CPU back then. As fast or faster then the early P4s, but completely let down by the Socket370/i815 platform required.

Best illustration was the simple fact Intel resurrected it for mobile applications. Then developed the Pentium M from it... the rest is history...
 

richaron

Golden Member
Mar 27, 2012
1,357
329
136
Intel has twice resorted to increasing clockspeeds to unrealistic levels to respond to AMD. First with the PIII 1133 against the Athlon and later with P4 against the Athlon. It seemed desperate.
Ryzen is more of a threat than the Athlon because Intel has no "Conroe" response as far as we know. Desperate times require desperate measures. If the tide start to move in AMD's direction it's hard to make it swing back.
We are living in interesting times.
Everyone credits Conroe with the paradigm shift of that era. But in my head it was Yonah which was the harbinger of change. Note my memory is not the best but I recall it being the result of an almost ground up rebuild of a pentium mobile CPU by a somewhat independent Indian branch of Intel. I doubt that combination of disjointed development and revolutionary tthinking will ever happen again, and I guess that limits options to things like clocking to the sky...
 

DrMrLordX

Lifer
Apr 27, 2000
19,155
7,915
136
@richaron

Conroe really finds its roots in Banias and Dothan. Pretty sure they came out of Haifa, along with Yonah. And some of what went into Yonah goes back to Timna, a failed product experiment also conducted at the Haifa facility.
 
  • Like
Reactions: lightmanek

richaron

Golden Member
Mar 27, 2012
1,357
329
136
@richaron

Conroe really finds its roots in Banias and Dothan. Pretty sure they came out of Haifa, along with Yonah. And some of what went into Yonah goes back to Timna, a failed product experiment also conducted at the Haifa facility.
Israel makes sense. I said my memory is bad... The point was more about the separate development and different design methodology. That is an "isolated" rebuild of an efficient core to make it big vs tweaking a big core to make it better (like usual). And up until this point Big vs efficient CPUs had been more separate (efficient CPUs being more in their infancy). So again I think this particular situation was pretty unique.

I recall Yonah being a big deal when pushed in a proper motherboard (again random facts about being first to 4Ghz overclock and/or under 17 seconds in some prime or pi benchmark are coming to me but I doubt my reliability). Random imagined or not stats aside, it was clear at the time that Yonah was the revolutionary leap and Conroe was more evolutionary (but also better known).
 
Last edited:

Insert_Nickname

Diamond Member
May 6, 2012
4,580
1,163
136
I recall Yonah being a big deal when pushed in a proper motherboard (again random facts about being first to 4Ghz overclock and/or under 17 seconds in some prime or pi benchmark are coming to me but I doubt my reliability). Random imagined or not stats aside, it was clear at the time that Yonah was the revolutionary leap and Conroe was more evolutionary (but also better known).
Back then we played with fitting Pentium M into desktop Socket 478* mainboards. When using desktop grade gear, the performance you could get out of them was amazing. At a fraction of the P4's thermal output. Intel actually made 130W single core P4 desktop models back then, it was really embarrassing having them matched by a lower-clocked 37W laptop CPU.

*ASUS CT479 adaptor. There was also a few select Socket 479 desktop boards. Among them the AOpen i915GMM-HFS, which powered my proto-HTPC for many years.

Smiles at his 6C/12T 4.2GHz Ryzen 3600. What a way we've come.
 

Topweasel

Diamond Member
Oct 19, 2000
5,414
1,606
136
@richaron

Conroe really finds its roots in Banias and Dothan. Pretty sure they came out of Haifa, along with Yonah. And some of what went into Yonah goes back to Timna, a failed product experiment also conducted at the Haifa facility.
Everything from Banias to and including Nehelem was out of there. Main Intel development center took over with Sandy Bridge.
 

ASK THE COMMUNITY