AMD FX is AMD's Pentium 4

Disclaimer: I am not an Intel fanboy, to the contrary I've only bought AMD CPUs (owned Intel but never bought).

AMD has been dropping the ball in it's CPU division ever since they bought ATI. They are a shadow of their former self, their CPUs have stagnated and their GPUs have blossomed, go figure! The FX reminds me of the Pentium 4, why? Lets go back in time.... Intel was late in releasing the Pentium 4 constantly pushing it back, and when it was released performance was sub par the higher clocked Pentium IIIs of the time were able to beat the Pentium 4 in most applications/benchmarks (sound familiar?). So the original Pentium 4 based on the Willamette core sucked but after, Intel was able to release a more competitive Pentium 4, the Northwood based Pentium 4, but with that CPU came high power consumption (remember how you needed a motherboard with a 4-pin CPU power connecter to use the higher clocked Pentium 4s?) and increased heat output (does this all sound familiar to the FX? Because it should). In the end the Pentium 4 was a hot and IPC inefficient CPU based on clock speed marketing just like how the FX is based on core count marketing. This is why I think the FX is AMD's Pentium 4. Does anybody agree?

im not saying FX isnt a dissapointment, but lets wait to see if windows 8 really does improve on it

Now we are to moderators watching this. Please keep it most clean and civilized as possible.

Improvements to the scheduler in Windows 8 won't save Bulldozer. The design isn't necessarily a bad idea, but it is ahead of its time, the vast majority of software simply isn't threaded enough to take advantage of the higher core count. That combined with what in this day and age is almost abysmal single threaded performance, and you have a flop.

Its only saving grace as far as gaming goes is the fact that most games are GPU bound these days, so its poor performance is only really apparent in a handful of CPU heavy titles, or when it is paired with a really high end SLI/Crossfire setup. It also performs okay in a small number of content creation applications that are heavily multithreaded. However, the fact that a quad core Intel CPU without hyperthreading performs very close to an 8 core AMD CPU is quite telling. There are definitely some major flaws AMD has to work out with this design, you shouldn't need to overclock your CPU to 4 or 5 GHz to get good single threaded performance.

i think voting options are too limited.
p4, disappointing and garbage do not usually mean the same thing. so the first option could come off as 'bulldozer-bashing'. amd fanboys could take offence and start up some verbal fights with ipcsomething or multithreadedthingamazig.
for instance, i am disappointed with bulldozer. but i don't think it is garbage. the idea behind bulldozer is pretty good. the implementation, execution and marketing were bad. still voted the first option because i could not fully agree with the other one.
imo if the voting options were a bit more diverse, people's responses would be more accurate.

Maybe its cause you are overpricing them...anyways this thread is about fx not amd..

they do great in integer heavy benchmarks, and those that really use 8 threads.
they are not garbage, though i did select that option in the poll.

would i recommend it to anyone ? never.
would i ever buy it myself? never.
will win8 make it better? no.
but it is not garbage.

Its a well known fact that the Higher clocked PIIIs performed better than the original (Willamette) P4s in all but a handful Applications at the time of release - In a sense, It was ahead of its time (Most applications still used older instruction sets and code more suited to the PIII or the Athlon) - Which again is a testament to his claim about how Bulldozer is a kin to the P4.

I like the FX, keeps my table with one shorter leg very stable, AMD!

Back then, i was still assembling computer for a company that built workstations and "boutique" computers. The original willamette (1.3 to 1.5 ) was weaker than most athlons, tualatin P3, and later on Intel launch a celeron 1200 model that also wiped their ass. Even tomshardware had charts reflecting that. P4 was a fluke. And using RIMMs increased alot the price of the computer.

Athlon XP/TB vs P4

[link]http://www.tomshardware.com/reviews/athlon-xp-meets-p4,383.html[/link]

Another Willamette Review
[link] http://www.tomshardware.com/reviews/intel,264-23.html [/link]

Quotes from the second review

"Let's get serious now however. We have learned that Pentium 4 has got a rather exciting and interesting brand new design that comes with a whole lot of potential. However, the benchmark results might seem a bit sobering to the majority of you. Whatever Pentium 4 is right now, it is certainly not the greatest and best performing processor in the world. It's not a bad performer as well though."

"Pentium 4 at 1.4 GHz goes for $644, Pentium 4 at 1.5 GHz costs $819 right now. It's not exactly a bargain, but, hey, who really cares about price if it really is all about style?"

Can't find any reviews under the codename Tualatin, that was the last iteration of PIII. It reached a maximum of 1.4GHz (stock) and we used it widely for our boxes. We could use 166 Mhz SDRAM at the fraction of the COST of RIMMS and usually with more than double capacity. Tualatins capabilities were always hidden from consumer because they were really good and Intel wanted to sell those extremely overpriced P4.

If you search a bit you will see, that P4's were crap. Too expensive for their own good. Any P3 or Athlon were literally half the price. And with more performance.

Well Im an AMD fanboy, but well Intel is doing a great job with their strong processors...

I really hope that AMD will come with a better processor. At least the so much talked "improvement" that Win8 will give to FX should help, but what I am really waiting is a new architecture or at least the next FX gen... I´d would change my crappy processor by there. AMD! PUT YOUR EYES IN THIS KIND OF FURUMS!

Does FX have onboard GPU, is FX available in a laptop?

And to the original question, yes it might be ahead of its time, but you have design for the market that exists, not one that might exist in a 2-3 years.

wow op what a load of crap. amd made 1 cpu that managed to beat intel every 1 since has been slower and under performed against intels comparative cpu...
amd have always been a generation behind but got lucky in the early days and came up with a cpu that punched above its weight.
bulldozer was a gamble that didnt pay off. they were hoping to create another athlon but failed... to say that bulldozer is amds pentium suggests there was some sort of too and fro of the title of best cpu, when we all know ever since athlon, intel have held the crown...
amd had the budget cpu against intels premium for performance , but since sandy bridge and the 2100 intel have made inroads into amds territory in a big way...
no bulldozer isnt amd's pentium coz every cpu since pentium has been amd's pentium... and yes you are a fanboy....
(you gave it away by sayin you owned but never bought intel. like it left a bad taste in yer mouth.... grow up.... buy the best not a brand.)

If you need the help...

This error is what stopped be in the original post as well. The p3s kept up with or were faster then the higher clocked P4s.

There are two ways to be the fastest chip. Either perform more instructions per clock, or have more clocks. Both BD and the P4 seem to be going or went the way of more clocks. But for me if you are going to talk about how things are and can't get your chips straight, step back and try again. P3s had better IPC then the P4s, and were clocked slower. I too remember the 1.3GHz P3s beating 1.6GHz P4s. It's not exactly news.

You Misunderstand.

What we (or at least I) meant was that the PIIIs with the higher clocks (1.0Ghz Coppermine, Coppermine T and Tualatin cores as opposed to 500Mhz) were faster than the early Willamette P4s, I am fully aware that P4s had terrible IPC and were shipped with higher clocks. Please don't confuse Chip Knowledge with your inability to understand a simple argument.

If thats the case you should use "later P3s" or the chip names like Coppermine. The way it was written lead to confusion.

IMO, the poor power consumption, etc can be blamed on upper management's decisions from the time they spent too much money on the ATI acquisition, forcing them to sell off their fabs and thus be dependent on GloFlo's yields and process problems.

However, BD is also the latest example of AMD occasionally badly misjudging the current market. It's mundane IPC performance, having to rely on a future OS and software capable of using all those cores, is very similar to AMD's insistence on 'native quad-core' for Barcelona, instead of going with the MCM approach like Intel did. That decision gave Intel a solid year to market quad-core CPUs before the TLB-fixed Barcie came out, and marked the beginning of AMD shedding many billions in value up until the end of 2010.

If AMD got more involved in software development, like Intel does, they'd understand the present market and where it's going a bit better..

It's also each companies first forte into "hyperthreading". Intel didn't get it 100% right the first time, not surprised AMD is having issues as well.

AMDs Bulldozer should not be confused with SMT.

I think they should have let Bulldozer ferment a little longer and just release an improved 32nm Phenom III X8 or something, it would've been better then what we have now.

The original Phenom was an undoubted flop, but it was also an evolution for AMD. They took their original Athlon Dual Core ideas and were able to translate that to quad core.

When I look at Bulldozer, I see the original Phenom. Not a stellar performer when compared to the competition, but at the same time, AMD is really the only company offering ANY competition to Intel right now. At least BD is able to stay in the same heat.

But when I look back at the original Phenom, I also realize that the Phenom II was a huge advance in both performance and capability over the Phenom. So I have my hopes set for Piledriver. I don't expect anything phenomenal, but I do have hopes that it will still be in the same heat as Intel yet again.

The future of mobile computing will always be tighter integration of resources into the same die. SoC for mainstream computing is already here in some cases, and for gaming it really isn't that much further off. And I think AMD has a better chance of getting it right than Intel does, given the lack of improvement Intel has had with graphics performance.

Bulldozer - a good candidate, but lackluster when compared to the Intel competition.

Piledriver - hope in the wings, but we will see what comes of it.

Llano - really enthused over this budget performer. Not necessarily an enthusiast platform, but this APU rocks hard for HTPC builds and budget laptops.

Trinity - Next gen APU...enthused to see PD with Radeon Cores on the same die. Even if PD is as lackluster as BD, the APU integration will only improve budget performance without taking anything away from the enthusiasts.

Same concept. The heart of SMT is to keep the inactive parts of the execution unit busy by allowing more tasks to be run at once. At a bare minimum, this requires a duplication of the register stack, which is essentially all HTT does. AMD CMT basically throws in extra Integer units, which would help in certain loads, but its the same concept overall. [FYI, the MIPS architecture was the grandfather of the "keep the CPU pipeline busy and ramp up clocks" approach]

The biggest issue is that p4 had no hope

VS

It took over 500 mhz for a p4 to = Athlon XP in increase of 33%
BD is capable of overtaking even a i7 2600k with a mere 200 mhz or 5%

Think about this, this is a synthetic test. This test pretty much ignores the cache curse that follows BD around like the plague. but it shows that BD is more than capable of performing once that issue is taken care of. P4 had no hope at all, internally it was slow as hell. The only thing that helped P4 was the rambus speeds vs ddr 1 memory.

Again, keep in mind this is the cpu's potential in a benchmark, not real world. BD != P4

Cache curse http://www.anandtech.com/show/4955 [...] -tested/6.

That's how you go from the top synthetic test to moderate performance.

i read that when zambezi launched. the cache design looks like a conscious design decision. i mean, from amd's perspective it must have looked like a design worthy of implementing (disregarding how it really turned out). it's not a curse.

Correct. They did the same thing when they moved from S939 to AM2. They had to increase the latencies a bit to handle any extra L2 Cache they might need to add later, and to handle the faster DDR2 speeds. They have done this throughout all their CPUs that I can remember. I have no idea what he's talking about but I'm always/usually in the mood for education/more info on something.