• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Worst CPUs ever, now with poll!

Page 5 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

What's the worst CPU ever? Please explain your choice.

  • Intel iAPX 432

  • Intel Itanium (Merced)

  • Intel 80286

  • IBM PowerPC 970

  • IBM/Motorola PowerPC 60x

  • AMD K5

  • AMD family 15h

  • AMD family 10h

  • Intel Raptor Lake


Results are only viewable after voting.
I never had a problem with my Northwood based P4s and I had two of them. I also had Willamate and Prescott based P4s that ran well too.
 
I know Tulatin came out after P4, and I mentioned that in my post, but my point was that releasing Tulatin after the P4 essentially punished the early adopters who ditched their S370 systems for S423, which was a doomed socket, when essentially S370 had just as much longevity. Is there anything a 1.7 Ghz Willamette can do that a 1.4 Ghz Tulatin can't?

Tualatins wouldn't fit in most socket 370 boards. They used a different package than Coppermine. I think Coppermine-T used the same package as Tualatin.

Anyway, look at it like this:

The FX-8150 performs near identically to a Core 2 Quad Q8400 - a CPU that's now several generations old. The FX-8150 is slower than a Q9750 in almost all benchmarks. A 3770K - a CPU that Bulldozer is expected to compare to - performs around twice as fast as a Q9750. It just isn't feasible for AMD to refine Bulldozer enough to gain >100% performance.

Netburst couldn't compete with K8 - absolutely no one in their right mind will deny that. Top of the line Netburst CPUs, however, were more than half as fast as the AMD CPUs they were expected to rival.

A Core i3-3240 should be a pretty decent match for an FX-8150 when it gets released. A mid-range CPU can keep up with AMD's "best." A Sempron 3800+ would have a hard time managing against a 3.46 EE, 3.73 EE, or 3.8 E.

Netburst was bad, but it was not a complete disaster on the scale of Bulldozer. Netburst did what it was intended to do - in its life, it went from a 1.5 GHz unit that struggled against Tualatin to a 3.8 GHz unit that was slightly competitive (though decidedly on the losing end) with AMD's CPUs. That's more than a 250% increase in clock speed. Maybe I'm wrong, but I don't expect AMD to begin pushing out 7.2 GHz units any time soon.
 
I don't get all of the hate for the entire P4 family. I distinctly remember the Northwood 2.4c or whatever it was being a very popular CPU on this board. So much so that I actually built a computer in 2003 based on one. It used DDR memory, had decent MB options and OC'ed well.
Maybe it was popular and good value for money. But clearly, Pentium 4 was not the gamers choice. In fact, AMD was much faster per clock and used less power at the same time.

 
Last edited:
Tualatins wouldn't fit in most socket 370 boards. They used a different package than Coppermine. I think Coppermine-T used the same package as Tualatin.
Wrong, they fit just easily as Coppermines. Tualatins could work in most 370 boards with a simple hard mod, or with the help of converters. I had it working in so-called "incompatible" boards as well. They did not require any special microcode to begin with (incl. the "S" parts).

Netburst was bad, but it was not a complete disaster on the scale of Bulldozer. Netburst did what it was intended to do - in its life, it went from a 1.5 GHz unit that struggled against Tualatin to a 3.8 GHz unit that was slightly competitive (though decidedly on the losing end) with AMD's CPUs. That's more than a 250% increase in clock speed. Maybe I'm wrong, but I don't expect AMD to begin pushing out 7.2 GHz units any time soon.
In the first iteration, Netburst was expensive and slow-performing, requiring RAMBUS. Zambezi isn't great either, but at least it works with the same memory and some the higher-end AM3 motherboards would allow you to use them. IMO, Bulldozer has had a much better start (both in terms of the price and performance) than Netburst. And like I said earlier, the Bulldozer architecture hasn't seen its best yet. So it's only valid to compare it with the first Pentium 4s.
 
Last edited:
This thread is an obvious troll nest.

Wow, people voting for Bulldozer have a very short attention span.
Pentium 4 was not great, but it was good enough to be Intel's only retail x86 offering for many years and a lot of people still use them.

Where is the good old VIA C3? Yes I know it is a renamed Cyrix III.
Lots of people used them for low power machines before Intel launched the Atom processor and the C3 was horrible at everything.
The only arguments for buying a C3 was that it was passive and the only mini-itx choice.
 
In the first iteration, Netburst was expensive and slow-performing, requiring RAMBUS. Zambezi isn't great either, but at least it works with the same memory and some the higher-end AM3 motherboards would allow you to use them. IMO, Bulldozer has had a much better start (both in terms of the price and performance) than Netburst. And like I said earlier, the Bulldozer architecture hasn't seen its best yet. So it's only valid to compare it with the first Pentium 4s.

Pentium 4 wasn't three generations behind when it was released. The future of Bulldozer is bleak because there's no way it can make up am absurd amount of ground with tweaks. Willamette was overpriced, but it was best described as "redundant."

As for the Pentium 4 vs. AMD benchmark posted: Northwood C was out before Athlon 64. Athlon XP was competing well with it and rated fairly until the 3200+ was released. That one almost never managed to keep up with a 3.2C.
 
people who bash bulldozer like its the worst thing ever forget there was a core before core2. core didn't beat a P4 in single thread either, but it was much more power efficient design. Bulldozer was basically a complete rebuild and while it didn't hit its perf targets AMD are still all the better for it. They now have a micro processes architecture designed around modern foundry processes not 130nm.

trinity being a perfect example, minimal changes to the core, a very solid IPC increase with higher clocks at the same power(edit: on the same process!). Now lots of people say 32nm SOI hasn't been the greatest, the move to 28nm Bulk will give as some kind of idea how true that is.

the mistakes of bulldozer come from two areas
1. there design goal turned out to be the wrong one
2. they missed there performance targets.

that doesn't mean in the slightest that the stuff inside of bulldozer is rubbish.
 
Last edited:
The P4 hate is certainly warranted, not only due to the lackluster performance, but the choice of platforms Intel offered for purely political reasons. It's bad enough that the P4 initially couldn't beat a late model P3, but why force every first year P4 buyer to get the disaster that was RDRAM? Then later on, you had a choice between the ridiculously expensive RDRAM or the piss-poor SDRAM platform, while the Athlon was cruising along using fast and inexpensive DDR RAM. Intel even took to legally threatening third party chipset makers who took the initiative to make DDR RAM P4 boards. Way to screw your customers, Intel.
 
well intel p4 and amd bulldozer is the same mistake
1)both made a cpu that was worse than the other
2)bulldozer signle thread perfomance is not that bad compared to p4 vs p3 perfomance of -30%
3)bulldozer may fell in single thread but in multi thread can score a tiny better than thuban 6 true cores (p4 era was single core cpu)
4)the power envelope is the same for thuben to bulldozer but compared to intels amd is so far away
5)intel pushed for expensive rambus or lacking perfomance sdram
bulldozer is using the same ram

the thing is that even that bulldozer does not have the same limitations as intel is so wrong for amd

p4 gave amd the opportunity to make a name and gain marketshare from intel
amd now is behind from intel and bulldozer will cause her to loose more ground
intel could afford that mistake
amd can not afford it
 
These poll options are so trollish it's maddening. Classifying PPC 60x chips as one processor makes about as much sense as counting all chips under the Pentium umbrella as one.
 
I had a Cyrix 133 machine that was nothing but headaches and crashes. Put up with it for about a year and then dumped it to get a P1 233 MMX that was so much better.
 
Is this poll a troll attempt or what?Classifying family 10h(seriously?) or even 15h(with all its weak points) as a "worst CPUs ever" is borderline insane.
 
wait what was wrong with the 286? it was a perfectly good cpu... 16bit bus, 16bit operations ... very fast for its time. and you could even get a 16mhz or maybe 20mhz one from some of the clone companies like harris semiconductor at the time.

i'd say the cyrix 6x86 was horribly poor especially with its fairly exaggerated PR ratings, and 10000 variants with slightly different bus speeds. national semiconductor mediaGX might be the worst.


and most people here probably won't remember (unlike most sane children i spent my 8 year old years reading pc mags like byte) the 80186 was an actual cpu that only a few companies adopted...

The 80186 was mainly for embedded systems. My company still uses them in its products!
 
Why would they? Cacheless celerons were the bomb. Where else could one go from 226 mhz to 448 mhz by simply upping the bus speed and get amazing performance in games of the era. PII processors of that era rarely could do above 75 or 83 mhz FSB, nowhere near the 112 mhz FSB that could be done with the ASUS P2B and a cacheless celeron 266.

The celerons with cache were fantastic overclockers and got great performance in games, but weren't the cacheless versions like the 266 and 300 gutless piles that were quickly done away with after giving the whole celeron brand a bad image right out of the gate?
 
and most people here probably won't remember (unlike most sane children i spent my 8 year old years reading pc mags like byte) the 80186 was an actual cpu that only a few companies adopted...

As a kid I was fascinated by the Tandy TRS-80.

Wikipedia... said:
"
Processor: 8-bit Intel 80C85, CMOS, 2.4 MHz
Memory: 32 kB ROM, 8, 16, 24, or 32kB static RAM. Machines with less than 32 kB could be expanded in 8 kB increments of plug-in static RAM modules.
Display: 8 lines, 40 characters LCD with 240 by 64 pixel addressable graphics. The screen was not backlit.
 
Made an interesting wPrime comparison, the other day. This is one test, where a 90nm Prescott [Intel Pentium 4 506 2.66 GHz] totally blows to a 130nm Tualatin [Intel Pentium III 1400, tb1 stepping]. I had it slightly overclocked.. but still, a very respectable result. Consider just, that Tualatin was released in 2001 and that Pentium 4 in 2005. And still, looks like the original poster has found no room for it in the poll? Oh dear.


 
Last edited:
Simple, Any AMD unless you like to cook your eggs on your PC.. 😉

Meh.. AMD hasn't always been bad. Back in the XP and X2 days, AMD was a stellar choice that offered a fair bit of performance over Intels P4. It really wasn't until the C2D that AMD started to fumble and stutter. It has taken them several generations just to catch up with Intel's C2D performance.
 
@ Cogman,

Yeah, that's where the original AMD FX processor showed its real strength. It was the 3960X of its time. Most people here, are too young to have such a short memory 😛
 
Meh.. AMD hasn't always been bad. Back in the XP and X2 days, AMD was a stellar choice that offered a fair bit of performance over Intels P4. It really wasn't until the C2D that AMD started to fumble and stutter. It has taken them several generations just to catch up with Intel's C2D performance.
They're still not there yet...
 
What about the Pentium Overdrive?

I voted Itanium for obvious reasons but the 486 drop in upgrade to Pentium performance was the worst thing I ever tried to use myself.

In the 20 hours I tried to use the thing it crashed after booting within 20 seconds and the manual was nothing more than a list of reasons it wont work with your system in particular. (It was a long time ago and my memory is fuzy.. plus I was relatively young at the time.)
WHAT! How very dare you sir.

Pentium overdrive was awesome!

http://forums.anandtech.com/showthread.php?t=2250863&page=3&highlight=overdrive
 
Back
Top