9 Years of AMD CPUs: From AMD FX to Ryzen 5000 Series, Tested

[quadibloc]

It's true that the people suing AMD won a settlement.
However, vector instructions like MMX, SIMD, and AVX are supplements to the core instruction set of a CPU, and so even if AMD decided that it was worth settling the lawsuit instead of fighting it, that doesn't mean that the advertised core counts of Bulldozer didn't correspond to the generally accepted definition of what a CPU core is.
So counting Bulldozer processors in your survey as having half their actual number of cores doesn't improve the accuracy of the comparisons.

 

[maxxcool7421]

"Some might still argue that the FX-8350 is an 8 core CPU, but it’s not, and it’s certainly not according to the California Consumers Legal Remedies Act."

I don't care what some court says about tech, especially in some absurd civil case in a country where you can legally sue ANYBODY for ANYTHING. The CPU core was defined as an integer computing core since the original Intel 4004. According to the "California Consumers Legal Remedies Act", Intel, AMD, Motorola, Texas Instruments (Cyrix), SPARC, IBM and VIA could ALL be sued for selling CPUs that didn't have FPUs attached. The FPU isn't even x86 architecture, it's x87 architecture and used to be (pre-486DX) sold separately as a "math co-processor" for floating-point operations.

Steve, I'm surprised at you that you'd let some court in another country make you ignore all the knowledge that you have about CPUs. Since when are lawsuits in the USA determined by who is right as opposed to "who just doesn't want to deal with it anymore"?

No court in some US state is going to define what a CPU core is for me when I've known damn well what a CPU core is for more than 20 years before that frivolous lawsuit case was brought forward. I wouldn't be surprised if it turned out that Intel had something to do with it.

Hahahahahahahahahahahah so true

 

[maxxcool7421]

Except that's not what the issue was (and you're wrong regardless). They were whining about the fact that two integer cores (the x86 CPU cores) were sharing a single floating point core and so they tried to argue that it wasn't a "true" 8-core CPU despite the fact that Windows still recognised it as such. It was a case of computer illiterates who thought that they were geniuses and so they screwed up royally but tried to blame AMD for it.

I HAVE an FX-8350 and used it for five years. I know that it's an 8-core CPU because it functioned EXACTLY as an octocore CPU should. Do you or did you actually OWN one of these things or are you just assuming that you know these CPUs better than people who actually own them?

Calling an FX-8350 a quad-core CPU is nothing that an intelligent person well-versed in tech does. It's what someone says when they have no clue but WANT it to be the case.

Never owned a faildozer. Stuck with my (3) 1090-Thubans@3.9ghz. Saw the initial reviews for the 8350 and held onto the older parts I had until Intel Devils canyon came out and clocked it to a mild boring 4.8ghz for 3 years.

If it cannot execute 8 threads, all day long, regardless of fpu / int data types ... it is not an 8 core cpu it is just a efficient 4 core. Except it wasn't even good at that in some cases.

 

[Theinsanegamer]

For the time, the AM3+ FX series CPUs were solid performers, especially at the price point they were offered.

"solid" is arguable. Even back int he day they were getting beaten by intel parts, and said intel part shave aged much more gracefully then FX did.

You guys forget a few things why the FX still was very populair:

- Multithread power, it would beat the i7 on encoding and such
- Overclocking headroom, I mean buy a 3.2Ghz model and push towards 5Ghz if you had the right motherboard, cooling and such.

Overclocking did yield more out of it, not just the raw core clocks but also the internal bus like the FSB or CPU/L3 Northbridge etc.

If I remember well, my 8320 at 4.8Ghz did around 767CB which was equal to a Ryzen 1600x.

They could take a punch, overvolting all the way through 1.6V and nothing was wrong. Unlike ryzens today. They degrade in weeks if overclocking was done improper. With the birth of Ryzen, the OC'ing kind of died at that point. PBO etc is far more advanced then all core.

"popular"? Nah mate. Bulldozer was fun to OC, yeah, but it was never popular, it lost a lot of market share t0o intel, this was the era when AMD hit rock bottom in marketshare, share price, had to sell their HQ, ece.

My friend and I had a blast tweaking his 6100 and later 8370 platform, watercooling and such, but everyone else (including myself) had sandy/ivy bridge or nehalem at the time.

Wrong.

GTA V was capable of using, and receive a benefit from, 6 core CPU's back in 2013. Your facts are all wrong as usual, and as usual again you are fanboying Intel hard.

Oh wow, one game! You sure showed me!

You forgot battlefield, that too benefitted from more then 4 cores. Of course intel chips were outperforming AMD's 8 core back in 2012, so ehh........

sorry that facts trigger you so hard, but facts remain, games disnt show benfit from having more cores until the late 2010s, at which point intel had a 6 cor emodel on the market. Is intel blameless? Nope. Was there an alternative? Yeah, intel offered 6 and 8 cores for years, but people whined because they cost more.

 

[Theinsanegamer]

It's true that the people suing AMD won a settlement.
However, vector instructions like MMX, SIMD, and AVX are supplements to the core instruction set of a CPU, and so even if AMD decided that it was worth settling the lawsuit instead of fighting it, that doesn't mean that the advertised core counts of Bulldozer didn't correspond to the generally accepted definition of what a CPU core is.
So counting Bulldozer processors in your survey as having half their actual number of cores doesn't improve the accuracy of the comparisons.

Well, by that definition then the pentium IV was techincally a dual core CPU. It had 2 ALUs after all. And let's not get started on apple's G5, which had 2 FPUs and 4 ALUs for a single "core"!

The best description for faildozer was that AMD built a core with hardware assisted hyperthreading that shared a single FPU. The CPU could run 2 executions at once on two sets of hardware, but anything FPU related had to go in order, technically not a true dual core nor a true single core, hence, quad core with hardware assisted hyperthreading. Anything floating point related (like games), the bulldozer was a quad core, not an octo core.

Either way, it also royally sucked and lost to its own predecessor, the phenom II. an OCed 1100t could go toe to toe with a 8370.

 

[maxxcool7421]

Well, by that definition then the pentium IV was techincally a dual core CPU. It had 2 ALUs after all. And let's not get started on apple's G5, which had 2 FPUs and 4 ALUs for a single "core"!

The best description for faildozer was that AMD built a core with hardware assisted hyperthreading that shared a single FPU. The CPU could run 2 executions at once on two sets of hardware, but anything FPU related had to go in order, technically not a true dual core nor a true single core, hence, quad core with hardware assisted hyperthreading. Anything floating point related (like games), the bulldozer was a quad core, not an octo core.

Either way, it also royally sucked and lost to its own predecessor, the phenom II. an OCed 1100t could go toe to toe with a 8370.

My 1090T's all ran 3.9ghz with a basic hypercool 212 cooler for 7 years before giving them away. Stayed under 50c while folding for Folding@Home for TechReport the entire time. Yeah, as soon as the 8150 benchs came out R ran the same bench's and saw that in many cases like heavy FPU the classic cores were miles ahead still I just walked away and waited until someone made a cpu that could beat the 1090t's in a cost affordable way.

Current rig is 1950x @4ghz 24x7 without hyperthreading enabled. So I do love me some AMD ... but not THAT AMD .... ugh

 

[Theinsanegamer]

My 1090T's all ran 3.9ghz with a basic hypercool 212 cooler for 7 years before giving them away. Stayed under 50c while folding for Folding@Home for TechReport the entire time. Yeah, as soon as the 8150 benchs came out R ran the same bench's and saw that in many cases like heavy FPU the classic cores were miles ahead still I just walked away and waited until someone made a cpu that could beat the 1090t's in a cost affordable way.

Current rig is 1950x @4ghz 24x7 without hyperthreading enabled. So I do love me some AMD ... but not THAT AMD .... ugh

I wanted a phenom II x6 in high school, as they were much more affordable then intel to a part time job haver. the semphron, athlon, and phenoms had great looking boxes too, compared to today.

By the time I saved up enough money, the bulldozers had come out, and so had piledriver. It was getting obvious that AMD was in a rut, and they stated they would not be releasing any more AM3+ parts and were dropping from the high end market. so I built my ivy bridge build (that ran at 4.6 GHz on the same old hyper 212+ cooler) that has served well ever since.

Looking back I should have kept it as my main for a few more years, instead of juming on the $250 9700k I have now.

 

[MasterMace]

Let's talk about FX, but we're not talking about 'FX', nobody wants to talk about 'FX', that would show how bad it can get, we're talking about "FX".

 

[Avro Arrow]

Never owned a faildozer. Stuck with my (3) 1090-Thubans@3.9ghz. Saw the initial reviews for the 8350 and held onto the older parts I had until Intel Devils canyon came out and clocked it to a mild boring 4.8ghz for 3 years.

If it cannot execute 8 threads, all day long, regardless of fpu / int data types ... it is not an 8 core cpu it is just a efficient 4 core. Except it wasn't even good at that in some cases.

Except that if you were to use that same logic on any CPU that came before the 486DX, they would also all fail that test. The fact that it had 8 integer cores meant that, it had 8 CPU cores. Now, to be fair, the fact that floating point operations were integrated into software did slow the FX down remarkably. However, that doesn't mean that AMD used false advertising. I've seen advertising that's FAR more dishonest than that without people batting an eyelid. I think that AMD did what it did because it was desperate at that point.

If nobody had bought any FX CPUs, there's a good chance that we wouldn't have Ryzen today so I'm glad that I did my part.

 

[Ephebus]

I'm 100% sure that the people behind the "California Consumers Legal Remedies Act" are experts in microprocessor design and architectures . **** them.

 

[ZedRM]

"solid" is arguable.

No, it isn't. Back then I regularly built systems based on the 8 & 6 core versions of the FX. They ran perfectly stable and were an excellent budget/value option which allowed people to focus their money on a GPU. They were very solid performers, even surprisingly so at times.

 

[Puiu]

Stardock also got pay from AMD to support their first Ryzen processors. Nothign shady with either

Paying for support isn't shady, but the way they implemented that support is weird. Limiting the thread count to exactly match Intel is just shady.

 

[trieste1s]

Of course it does. It turns out that Intel paid Stardock for the updated version that used 16 cores but ONLY 24 threads! (Someone posted the evidence of this in the comment section of the actual article about the AOTS test specifically). Stardock would never have done that on their own because before Alder Lake, there had never before been a 16-core 24-thread CPU. All 16-core CPUs had 32 threads. That leak is what is known in the industry as irrelevant.

It's about as trustworthy as SysMark, Intel's Compiler, LoserUserBenchmark and liquid cooling using a secret water chiller.

Of course it does. It turns out that Intel paid Stardock for the updated version that used 16 cores but ONLY 24 threads! (Someone posted the evidence of this in the comment section of the actual article about the AOTS test specifically). Stardock would never have done that on their own because before Alder Lake, there had never before been a 16-core 24-thread CPU. All 16-core CPUs had 32 threads. That leak is what is known in the industry as irrelevant.

It's about as trustworthy as SysMark, Intel's Compiler, LoserUserBenchmark and liquid cooling using a secret water chiller.

To think we have so-called classic capitalistic corporates in good ol Murica actually working against the market economy.
"May the best man win" doesn't mean best performance, it means best loopholing, best undermining biz tactics, best cap in hand to Congress for handouts.

 

[Theinsanegamer]

Except that if you were to use that same logic on any CPU that came before the 486DX, they would also all fail that test. The fact that it had 8 integer cores meant that, it had 8 CPU cores. Now, to be fair, the fact that floating point operations were integrated into software did slow the FX down remarkably. However, that doesn't mean that AMD used false advertising. I've seen advertising that's FAR more dishonest than that without people batting an eyelid. I think that AMD did what it did because it was desperate at that point.

If nobody had bought any FX CPUs, there's a good chance that we wouldn't have Ryzen today so I'm glad that I did my part.

I've seen this point brought up before and I will add a counterargument: CPUs from the 486 era are an entirely different beast. Modern software is not written around CPUs that needed calculator chips added on to do algebra. Every modern CPU, from x86 to ARM to RISC, has a FPU. It is an expected part of a CPU core today, thus you cannot measure what counts as a modern CPU "core" by what passed as standard in 1989.

Yes, AMD was beyond desperate, but that didnt justify their weasle wording on what constitutes a CPU "core". The FX was no more an 8 core then a pentium IV was a "dual core" or a pentium D a "quad core", or indeed a mac G5 a "6" core. Both the industry and consumers hav ecollectively agreed that modern CPu cores need to have a FPU to count as a full core, and the government eventually agreed with that as well. If your CPU has cores that cannot run modern instructions without a different core helping them, they should not be considered full cores.

No, it isn't. Back then I regularly built systems based on the 8 & 6 core versions of the FX. They ran perfectly stable and were an excellent budget/value option which allowed poeple to focus their money on a GPU. They were very solid performers, even surprisingly so at times.

When your 8 core can reguarly be beaten in gaming FPS tests by a dual core i3, it isnt really arguable at all. The FX bulldozer chips were slower then the phenoms they replaced, and that is beyond inexcusable. The piledriver 8 cores could only match phenom II x6 CPUs by clocking to the moon.

Could you build a system on them? Well yeah, same as you could build a rig on a prescott pentium IV. We had no issues calling the pentiums hot garbage, but for some reason people wanted to defend AMD's FX series. I'd argue that the FX being a "solid performer' simply isnt possible when everything else is whopping it at every opportunity. It was functional, fun to play with, but it wasnt a solid performer unless all you played was cinebench.

 

[Avro Arrow]

To think we have so-called classic capitalistic corporates in good ol Murica actually working against the market economy.
"May the best man win" doesn't mean best performance, it means best loopholing, best undermining biz tactics, best cap in hand to Congress for handouts.

You're 100% on-point with that statement. That's why capitalism is failing now. Too many people have learnt ways to rig the system. It's like when Steinbrenner bought all of those Yankee World Series wins. Watching baseball became irrelevant because it was no longer a sport, but an auction. Whoever bid the highest payroll won the World Series. Sure, it didn't ALWAYS happen but it did happen 8 out 10 times which means that most teams didn't have a ghost of a chance.

The NeoLiberal idea of Laissez-Faire Economics doesn't mean more freedom, it means more freedom for the rich and less freedom for the masses. This is why I'm a strong admirer of the Nordic countries. Their heavy regulation of the market doesn't make the market unprofitable (as many right-wingers would have you believe), it just makes sure that it's not just those at the top who benefit.

The starting pay at McDonald's in Denmark is like, $20USD per hour plus 4 weeks vacation. Yet, McDonald's in Denmark is still profitable (because if it wasn't, they'd have closed their doors) despite the fact that they also pay much higher corporate taxes than they do in the USA. I know that it has little to do with tech but it just makes you realise the kind of lies that corporations try to propagate to change the narrative. It's not just hurting marketplace competition (which they claim is so important) but it's hurting people directly as well.

Governments like those of the USA, Canada, the UK and Australia are enabling these bad conditions. Macron in France tried the same thing but it resulted in a nationwide strike which forced him to back down. The French know that it's damaging society as a whole and will loudly oppose and stop it. How long has it been since several generations have been unable to buy a home and raise a family? The last generation to be commonly able to do so was the Boomers. After them, X, Y and Z have all been unable to with things getting worse with each generation (X'ers maybe could, Y not really and Z no chance).

 

[Avro Arrow]

I've seen this point brought up before and I will add a counterargument: CPUs from the 486 era are an entirely different beast.

Actually, they really weren't. If you were willing to wait forever, you could run Windows XP on a 386.

If you were to look at the architecture of the 486:

And the Pentium-4:

You would see that they're really not different beasts. Regardless of this, the definition of CPU core is that of an integer core (ALU) because it has to be applicable to ALL microarchitectures and not all microarchitectures have a 1:1 ratio of CPU(ALU):FPU cores. If you have an IBM Power CPU or an ARM CPU, the number of CPU cores has nothing to do with the number of FPU cores. The number of based on the number of ALUs and always has been.

That's why, despite this lawsuit outcome, AMD has never been charged with false advertising, not even in the much more strictly regulated EU.

Modern software is not written around CPUs that needed calculator chips added on to do algebra.

I think you missed a part of my post:

Now, to be fair, the fact that floating point operations were integrated into software did slow the FX down remarkably. However, that doesn't mean that AMD used false advertising. I've seen advertising that's FAR more dishonest than that without people batting an eyelid. I think that AMD did what it did because it was desperate at that point.

That does sound exactly the same as what you just said. In any case, there was (and still is) software that doesn't use floating point operations nearly as much as integer operations. Where AMD dropped the ball was failing to implement a system where excess FPU operations would be offloaded to the GPU because, while not perfect, it would have improved performance.

However, this does not mean that it's not an 8-core CPU because it still has the 8 ALUs that, using the standard (and long accepted) definition of a CPU core, make it an 8-core CPU. It's a crappy 8-core CPU, but it's an 8-core nonetheless.

Every modern CPU, from x86 to ARM to RISC, has a FPU. It is an expected part of a CPU core today, thus you cannot measure what counts as a modern CPU "core" by what passed as standard in 1989.

Yes, every modern CPU does have an FPU, but no, they don't necessarily have the same number of CPU cores (aka ALUs) as FPU cores.

Yes, AMD was beyond desperate, but that didnt justify their weasle wording on what constitutes a CPU "core". The FX was no more an 8 core then a pentium IV was a "dual core" or a pentium D a "quad core", or indeed a mac G5 a "6" core. Both the industry and consumers hav ecollectively agreed that modern CPu cores need to have a FPU to count as a full core, and the government eventually agreed with that as well. If your CPU has cores that cannot run modern instructions without a different core helping them, they should not be considered full cores.

When your 8 core can reguarly be beaten in gaming FPS tests by a dual core i3, it isnt really arguable at all. The FX bulldozer chips were slower then the phenoms they replaced, and that is beyond inexcusable. The piledriver 8 cores could only match phenom II x6 CPUs by clocking to the moon.

Could you build a system on them? Well yeah, same as you could build a rig on a prescott pentium IV. We had no issues calling the pentiums hot garbage, but for some reason people wanted to defend AMD's FX series. I'd argue that the FX being a "solid performer' simply isnt possible when everything else is whopping it at every opportunity. It was functional, fun to play with, but it wasnt a solid performer unless all you played was cinebench.

The reason that FX gets defended is the fact that most of AMD's situation wasn't self-inflicted, it was a result of illegal practices by Intel. That makes people feel sympathetic towards AMD (we're human after all). Now, I never ONCE stated that I thought the performance of the FX-8350 was as good as even the i7-2600K for most tasks, quite the contrary.

What I did say is that while I know the Sandy Bridge architecture was faster (and it was, no doubt), I also said that I got my FX-8350 for $170CAD, a price that was incredible since the i7-2600K cost about to $475CAD and the i5-2500K cost about $400CAD. Add the fact that I paid WAY less for my Gigabyte 990FX motherboard than I would have for a motherboard based on the Intel Z77 chipset (I think it was Z77 at the time).

Despite the fact that the i7-2600K (and, at the time, the i5-2500K) were clearly faster chips for gaming, I never felt like I was suffering with my FX-8350 when it came to gaming. Sure, maybe I wasn't getting the super-high fps rates that the Intel CPUs would have given me, I get that. The truth of CPU gaming performance though has always been more complicated than that. The truth has always been that as long as your CPU is fast enough to keep out of your GPU's way, it's "good enough" and the FX-8350 was "good enough". As a matter of fact, the FX-6300 hexacore was the real sweet spot for gaming and that's what I was going to buy initially but Tiger Direct had a sale on the FX-8350 that made it only $30 more. I actually only paid $20 more.

I printed the coupon from their site for $20 off (it was selling for $200CAD normally) but they insisted that I was to get $30 off. I asked them if they were certain (I'm honest that way) and showed them the coupon that I printed but they showed me on their screen that it said $30. Well, I said "Ok" and bought it for only $20 more than the FX-6300. When I got home, I went back to my PC and looked at TD's site again, even refreshing it to see if anything had changed but nope.

So, now I thought I was going insane but since I had worked there previously it dawned on me that they might have been looking at tigerdirect.com instead of tigerdirect.ca. Sure enough, the US site had $30 off while the Canadian site had only $20 off. I had to chuckle over that. It was a new girl serving me and if she had done it by herself I would have gone back to save her some grief but her manager was there training her and he was fooled as well.

I loved my FX-8350 because it gave me the performance that I needed for less than half the price of what I would have paid for an Intel system. Believe me, if my FX system's performance was terrible, I wouldn't have used it for almost 5 years and I definitely wouldn't be speaking highly of it now. Looking back, I can see that it was definitely the right decision for me and probably would have been for most people if gaming was the most hardcore thing that they used their PCs for. That pretty much is the definition of most home PCs.

Having said that, you won't see me defending the FX mobile parts because those are things that I won't touch. As a mobile CPU, the FX was absolutely terrible because it used too much electricity and was seriously detrimental to battery life if you wanted decent performance out of it. If AMD had stuck with the stars architecture like Llano, they would've been much better off in the mobile space. FX was total garbage when it came to craptops.

I defend the FX-8350 because I truly believe that it is a true octocore CPU because it has 8 ALUs, no other reason than that. If Intel was the company in question, I'd be defending them as well with the exact same arguments.

To me, saying that the FX-8350 is NOT an 8-core CPU because it has fewer FPUs than ALUs is like saying that the new Intel 16-core CPU is actually only a 12-core (or less) because it only has 24 threads and some of the cores are weaker than others. I wouldn't say that because it's not true. It has 16 ALUs and to me, it doesn't matter if it has SMT or an equal number of FPUs as ALUs. A CPU is, and always has been, defined by the number of ALUs that it possesses, nothing more, nothing less.

Please note that this is all about a frivolous nuisance LAWSUIT in ONE US STATE, not a federal investigation into wrongdoing. As I pointed out earlier, the European Commission would have nailed AMD's hide to its office door if AMD had lied in the way that they've been accused. The question was never raised in ANY country by their governments (and what happened to Intel shows that they weren't shy to do so) and that, more than anything, should tell you what a crock this is.

I still can't believe that Steve Walton, of all people, would buy in to this crap.

 

[Avro Arrow]

I used to really hate AMD because I had to live with an FX6300 for almost 10 years

Well jeez, if I had to use the same CPU for 10 years, I'd probably also hate whatever CPU it was. Ten years is WAY too long to have the same CPU if you're a gamer. Funnily enough, in the early 2010s, the FX-6300 was considered the sweet-spot CPU for gamers because it gamed well and was dirt-cheap.

If you're using it now though, don't hate AMD because their CPU didn't stay relevant for ten years because I've never seen a CPU that can stay relevant that long.

 

[ZedRM]

When your 8 core can reguarly be beaten in gaming FPS tests by a dual core i3, it isnt really arguable at all.

Not one i3 in that generation matching the FX-8xxx CPU's, nor any of the following gens for 6 years. You are seriously silly if you really believe a dual core will beat out those CPUs. So please quit with your meritless nonsense.