Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. ×
AMD The Courts

AMD Sued Over Allegedly Misleading Bulldozer Core Count 311

An anonymous reader writes: A class action suit accuses AMD of misleading buyers about the number of cores in its Bulldozer-based CPUs. The complaint claims that the chips effectively had only four cores, while AMD claims there are eight. According to Ars: "AMD's multi-core Bulldozer chips use a unique design that combines the functions of what would normally be two discrete cores into a single package, which the company calls a module. Each module is identified as two separate cores in Windows, but the cores share a single floating point unit and instruction and execution resources. This is different from Intel's cores, which feature independent FPUs. The suit claims that Bulldozer's design means its cores cannot work independently, and as a result, cannot perform eight instructions simultaneously and independently. This, the claim continues, results in performance degradation, and average consumers in the market for a CPU lack the technical expertise to understand the design of AMD's processors and trust the company to give accurate specifications regarding its CPUs."
This discussion has been archived. No new comments can be posted.

AMD Sued Over Allegedly Misleading Bulldozer Core Count

Comments Filter:
  • Intel has a lot to learn from its smaller rival's marketing department :)
    • by Entrope ( 68843 ) on Sunday November 08, 2015 @12:01PM (#50887867) Homepage

      AMD's CPU architecture has a similar purpose as hyperthreading -- to share hardware resources between what looks to the OS like independent cores -- but the tradeoff is different. Intel's hyperthreading approach only works to cover memory latency, because the hyperthreads share so many physical resources (I think basically everything except register files and hyperthreading-related state). AMD's is somewhat different in that each "module" has two independent integer ALUs, register files, and L1 data caches. The module has one L1 instruction cache, one L2 data cache, one FPU, and one instruction fetch/decode unit.

      But AMD has always been pretty up-front about this architecture. There is maybe a cause of action against resellers who package the AMD chips into systems and do gloss over which aspects each "core" shares with another core, but AMD publicly presented the core-vs-module distinction well before the chips were released.

      • by gbnewby ( 74175 ) * on Sunday November 08, 2015 @12:54PM (#50888115) Homepage

        Concur. The design of the bulldozer modules was abundantly clear. The fact that the "cores" share an FPU was clearly disclosed and part of any diagram of the parts. The shared FPUs played a huge role in assessing bulldozer-based CPUs for high performance computing workloads. For HPC, the usual benchmark is HPL (a.k.a., high performance LINPACK), which is a measure of double precision floating point performance for a particular matrix operation called DGEMM. The fact that the FPU was doing double-duty for two "cores" on a module meant that the peak theoretical performance was limited by the number of FPUs in a CPU, not the number of cores or modules or anything else.

        As others have noted, hyperthreading via Intel can have exactly the same impact: the threads share various components, including the FPU.

        Another aspect that can have a major impact on performance is the number of memory channels, and how things like cache coherency is handled. Among other things, AMD's hyper-transport exhibits different scalability characteristics depending on the number of sockets. In a four or eight-socket configuration, latency due to cache coherency operations can have a big impact on performance.
            - gbn

        • It really wasn't AMD who was misleading customers, though. Retailers were the ones who were describing the PC's with these processors as something along the lines "blazing fast 8 core gaming monsters" when in reality they were being outperformed by most Core i5's that were available at the time.

          • They get those quotes from AMD.com
            "AMD FX Processors unlock maximum, unrestrained processing performance for extreme responsiveness you can see and feel."
            "The industry's first and only native 8-core desktop processor for unmatched multitasking and pure core performance with "Bulldozer" architecture."

          • So, I actually have an AMD FX 8320E eight core as the processor on my personal desktop.

            This is what AMD [amd.com] says:
            "The industry's only 8-core desktop processor", or "The industry's first and only native 8-core desktop processor for unmatched multitasking and pure core performance with "Bulldozer" architecture".

            Now, I bought it knowing there was likely some behind the scenes tricks, and because I don't strictly need a high-level of sustained CPU intensive tasks. For me it was as much about letting multiple progr

        • Sigh, does nobody know how the BD/PD arch works here? It CAN share the FPU but absolutely does NOT "have to". Allow me to quote from Wikipedia: "."Two symmetrical 128-bit FMAC (fused multiplyâ"add capability) floating-point pipelines per module that can be unified into one large 256-bit-wide unit if one of the integer cores dispatches AVX instruction and two symmetrical x87/MMX/SSE capable FPPs for backward compatibility with SSE2 non-optimized software."

          So each core DOES have an FPU, they simply used a simpler 128bit FPU that CAN be joined together to make a single 256bit FPU for use with AVX instructions and ya know what? The majority of users will NEVER NOTICE as every program and game Joe Average runs will run great on an AMD chip. I myself run an FX-8320E which I upgraded from a Phenom II X6 (which according to the lawsuit is a "full" core compared to mine) and the FX blows through transcodes and multieffects layering a good 40% faster than the P II X6. And if his BS was true how does he explain the fact that the 6350 and 1100T are equal in performance [cpuboss.com] despite the fact that according to him its a triple core versus a hexacore?

      • It's not quite so clear cut. Both vendors' chips are superscalar and so have multiple pipelines. Intel chips have around 5 independent integer pipelines, plus the AVX ones. Very few workloads will saturate them all. This was the main motivation for IBM's SMT: if you have two threads then you can get a lot closer to saturating the execution units than with one. Of course, you're still likely to suffer a bit because of contention on a few units, which is why Hyperthreading often isn't an overall performa

      • They're not upfront about it now

        Have a look at their website
        The FX series is "The industry's only 8-core desktop processor" and "The industry's first and only native 8-core desktop processor for unmatched multitasking and pure core performance with "Bulldozer" architecture"

        http://www.amd.com/en-us/produ... [amd.com]

        Trying searching their website for information on Bulldozer. There's bugger-all.

    • The problem is in the details. When you buy a I7 Intel is selling you a processor with 4 cores and hyperthreading. When you buy a bulldozer from AMD it is selling it as 8 cores but it is actually a 4 cores with something similar to hyperthreading. AMD is deceiving the consumer because her design have not really 8 cores, is a 4 cores with hyperthreading.
      • by tepples ( 727027 )

        Where was AMD dishonest that a Bulldozer contains four modules?

    • by gweihir ( 88907 )

      It is not hyperthreading at all. It is two full integer CPUs sharing one FPU. Hyperthreading gives you one additional brain-damaged CPU at maybe 40% the main speed. For integer loads, AMD gives you the performance of 8 full cores.

  • by Anonymous Coward on Sunday November 08, 2015 @11:59AM (#50887861)

    http://www.extremetech.com/extreme/217672-analysis-amd-lawsuit-over-false-bulldozer-chip-marketing-is-without-merit

  • Pretty Laughable (Score:5, Insightful)

    by TooManyNames ( 711346 ) on Sunday November 08, 2015 @12:08PM (#50887895)

    Why is this even reported? This suit isn't going to go anywhere (unless AMD's lawyers are extremely incompetent, and the judge is extremely incapable of understanding basics about computer architecture and ISAs).

    The AMD cores shared an FPU, sure, but sharing a resource doesn't mean that cores cannot execute simultaneously. The AMD cores still have independent integer-based execution units (instruction registers, register files, ALUs, branch counters, etc.), after all, and are fully capable of executing integer instructions simultaneously (which accounts for the vast majority of instructions under typical loading).

    • Re:Pretty Laughable (Score:4, Interesting)

      by Ramze ( 640788 ) on Sunday November 08, 2015 @12:37PM (#50888017)

      I agree. Also, if I understand correctly, that shared FPU can be used either by one CPU core as a 256 bit FPU, or by both simultaneously as 2 independent 128 bit FPUs.

      So, shared, but not likely to be a bottleneck.

      Also, since when do cores have to include all the "extras" ? I recall when 486's had a math co-processor and there were no mmx instructions or other such multimedia or physics sets. This guy is going to have a really tough time explaining how exactly AMD's architecture doesn't provide exactly the number of cores listed -- even if the architecture has its limitations due to sharing resources.

      • by jimbo ( 1370 )

        Well, I imagine bigger issues are the single instruction decoder and shared L1 cache. However, as mentioned before here, AMD was completely open about this architecture and this suit should go nowhere.

    • Re: (Score:3, Informative)

      by beelsebob ( 529313 )

      They share *everything* except for an ALU. When 90% of the functionality of the core is not duplicated, it becomes pretty damn difficult to assert that it's actually 2 cores. Instead, it's 1 core with 2 ALUs.

      There are only 4 instruction fetch units on the chip - that suggests 4 cores
      There are only 4 instruction decode units on the chip - that suggests 4 cores
      There are only 4 L1 caches on the chip - that suggests 4 cores
      There are only 4 floating point units on the chip - that suggests 4 cores
      The only thing

    • and Juries are really, really unpredictable. Even if the outcome is legally obvious it doesn't mean the jury won't get dazzled into blaming AMD.
  • by rbrander ( 73222 ) on Sunday November 08, 2015 @12:12PM (#50887915) Homepage

    I was shopping for VCRs about 20 years back and asked the Future Shop guy how much better it was for having (quoting from the card beside the VCR) a "19 micron tape head". Turns out they ALL had 19 micron tape-heads (whatever the hell that *meant*) as it was the spec for a VCR tape head, at the time, at least. It was just another bit of science-y sounding technobabble to put on the card.
    Buying based on core count is like buying for the 19-micron thing; it's either a fast machine for your purposes or not. Absolutely the only way to tell that for sure is a test. The only thing that was ever useful with, say, "megahertz" was that it had for a decade or so there a correlation with the performance you'd get in real use. I've never found "cores" to have anything of the sort.

    • The state of consumer CPUs today is that measuring by a single thing like gigahertz or number of cores is not useful. Even from generation to generation of AMD chips, you can't use those to gauge different chips as newer gen 6 core AMD might outperform an older 8 core AMD. There are however a number of CPU benchmarks that can gauge relative performance and the buyer should beware. That being said, for most consumers, they don't even come close to use all the potential of modern CPU.
    • by Ramze ( 640788 )

      Cores mattered when most systems had only 1 core and upgrading to systems with 2 cores or 2 cores with hyperthreading (4 virtual cores) made a huge difference in performance. That was a long time ago, though. The multiple cores meant less latency -- especially for multimedia tasks as the CPU didn't have to stop what it was doing to schedule time for *random background service* or other apps.

      Nowadays, it's as you say -- market droid speak. CPU makers hit a wall with Mhz and needed a new marketing ploy,

    • Good point. It's one thing to base it on MHz, which usually was a good indicator of relative speed within a CPU architecture (even though a 150MHz Alpha was slower than a 150MHz Pentium). It's another thing to base it on #cores, since it's well known that parallelism in CPUs is hard to extract, in which case, beyond 4 CPUs, one is quickly hitting the point of diminishing returns.

      Also, to what extent have the loads on Floating point units changed? I recall that in the 90s, FPUs used to be heavily de-emp

    • by jetkust ( 596906 )
      The amount of cores is extremely important when you working with multiple processes or multiple threads. And the reason isn't terribly complicated. Though it's possible that a single core processor could be faster than a multi-core processor, in the real world benchmarks of multi-core processors of similar product lines are always faster. In other words, double the cores, and roughly double the performance. Same reason video cards are fast. The amount of cores matters. There's a reason pretty much al
    • by wwalker ( 159341 )

      Why did *I* buy based on the number of cores?! Because I needed a CPU that can run as many processes in parallel as possible. Why else would you need multiple cores? No, it's not a web server, and not for bitcoin mining. 8 cores means I should be able run 8 processes in parallel, for 8 times speed-up over 1 core.

      Yes, doing a testing would've been ideal, but you can't return a CPU, as far as I know. Plus, I thought I can trust AMD to sell an 8-core processor when, you know, it says "8 cores" right on the box

    • When you do audio processing, for instance, the number of real cores makes a huge difference. If you use an Intel chip.

  • A judge is likely to ask : Were there 8 cores ? If the answer is yes, which it seems to be , then AMD is in the clear.

    No multi-core CPU box ever came with a statement that all 8 cores would be capable of processing instructions in parallel at the same time. It does however mean, that AMDs 8-core is significantly worse than Intel's 8-core.

    • In fact, the people suing will have a tough time convincing a judge. "But, your honor, there are only 4 FPU units." Is the judge gonna adjudicate whether 2 cores sharing an FPU unit merges them into 1?
    • by gnupun ( 752725 )

      Well two cores are not supposed to share things like integer and floating point units. So those two cores could be considered as one core by the court, and rightly so.

      Anyone know if each so-called core has separate L1, L2 caches or not? TFA is very light on details.

      • by hjf ( 703092 )

        according to another comment here: yes. unlike intel core i series, which share cache.

    • by swalve ( 1980968 )
      Way back when dual core processors came out, the implication was that the processor was two independent processors in one package. That logic follows onto 4 and 8 core machines. The example being that Intel doesn't try to pass off hyperthreading as separate cores.
  • I mean, if you bought a 3.5GHz chip and it didn't perform as fast as an Intel 3.5GHz it must be misleading marketing, right? We all know that clock speed is all that matters when you compare a chip. And, I've been told, that they may be reducing clock speed dynamically when the processor isn't fully loaded - basically cheating you out of the speed you PAID for. I hear AMD also ran over your cat.

  • by Cowclops ( 630818 ) on Sunday November 08, 2015 @12:30PM (#50887983)
    Regardless of how many cores you have, 90% of the crap that 90% of people do doesn't even benefit from having more than two cpu cores, and it only benefits from that because you can use one core for "the app you're running in the foreground" and one core for literally all other background tasks so your system doesn't lock up if you really need to use 100% of a core, AMD can throw more cores at their CPUs but end users won't see real world performance benefits except in limited scenarios. Intel's super cheap Pentium G3220 at 3ghz (and with only two cores) beats everything AMD sells except for hand-picked benchmarks where AMD's core advantage overcomes its brutal per-clock disadvantage. Single threaded performance is whats king and you'd have to run an AMD cpu at like 5ghz to overcome its poor performance per clock to beat a 3ghz Intel (core-architecture) CPU. And its sad for me to say this because I bought nothing but AMD CPUs from like 2000 until 2014. I wish they could be competetive but the situation looks pretty bleak. I can't use their stuff when their 8 core 4.5ghz cpu is only "a little bit faster" than my 4 core 3ghz phenom from 2009, but my i7 4790k is like, a solid 3 times faster than that 2009 phenom in almost anything you throw at it.
    • Your observations on the AMD vs Intel performance gap seem to match my own observations.

      I have a 3.6 ghz fx-8150 and my friend bought a Intel i7 also at 3.6 ghz. On most desktop operations you can't tell the difference. When we start operations like video encoding the difference in performance is staggering. On the average he gets over a 40% performance increase over my AMD system.

      That is a hell of a performance gap.

      • by swalve ( 1980968 )
        The PCs I support are mostly AMD now, and I've noticed the same thing. The numbers can say all they want, but they underperform compared to much older/slower/cheaper Intel CPUs.
  • Is this the same AMD "8-core" device that's in both PS-4 and Xbox One consoles?

    If so, the consequences may go much wider than this story makes it appear!

        -- Steve

    • Re: (Score:2, Informative)

      by Anonymous Coward

      Cant be, as the PS4 and Xbox One use Puma derived cores, a completely different Architecture than the Buldozer derived Desktop CPU.
      The Puma Cores do NOT share the FPU. Puma CPU use a special low power Design that doesn`t scale up well above about 2 GHz

    • I don't see how this issue can affect PS-4 and Xbox One consoles. Both those are marketed with out disclosure to the public on how many cores their processors have. Most people that buy games consoles don't care about that as long as their a radical performance increase over their last console.

      I doubt that ether microsoft or sony will have anything to say ether. Both their engineering teams crawled all over the cpu designs they where planning to put in the console. They didn't just point at a cpu an

  • Honestly they need to stop being allowed to advertise "cores" if they are not honest to goodness real separate processor cores.

    I hope AMD loses HARD on this and forces the industry to stop being misleading assholes.

    • by Luthair ( 847766 )
      As explained elsewhere what AMD is doing is not the same as HT.
      • by Lumpy ( 12016 )

        Correct, it's actually a little bit better.

        but it's still marketing bullshit.

        • The alternative is to try to call them "dual module" processors, and then go through a big long explanation to customers who really don't care what "module" actually is.

          Can it execute two separate threads simultaneously? Yes. At full performance? Mostly, although with a shared frontend, when both units are running at full load, instruction disp---- Is someone trying to make a practical judgement of a chip's performance based solely on its core count going to have a clue what any of that means? No. Then it

  • They've got nothing on NVIDIA, who advertise the GTX 980 as having 2048 "cores", when by any standard definition it only has 16 (or if you're really generous, you could maybe argue it has 64, but that's pushing it). They count every lane of their vector unit as a separate core. By that standard, AMD (and Intel) should multiply all their core counts by 8, since each AVX unit can do 8 int or float operations at once.

    • by tempmpi ( 233132 )

      They've got nothing on NVIDIA, who advertise the GTX 980 as having 2048 "cores", when by any standard definition it only has 16 (or if you're really generous, you could maybe argue it has 64, but that's pushing it). They count every lane of their vector unit as a separate core. By that standard, AMD (and Intel) should multiply all their core counts by 8, since each AVX unit can do 8 int or float operations at once.

      I think 64 cores in 16 modules would be a fair assessment of a GTX980. Each SMM contains 4 register files and dedicated execution units, 4 instruction fetch and scheduling units. Not too many things are shared within each SMM: mostly caches and shared memory.

  • "The suit claims that Bulldozer's design means its cores cannot work independently, and as a result, cannot perform eight instructions simultaneously and independently."
    If the suit really makes this claim it is easy for AMD to defend, because an AMD Bulldozer with 4 modules / 8 cores can actually execute 8 independent floating point instructions per cycle.

    The two cores in each module share the floating point units, but each module contains 2 independent 128-bit FMAC units. Floating point throughput could be

  • I run at least 5 VMs on it. And they all run all the time (in addition to the host OS). It also took some initial effort to unpark every other core. Which means cores 2,4,6,8 can be parked independently of their cousins. AMD chips are better for running VMs, while Intel chips are better at running games. I couldn't run 5 VM's on Intel 5 (4 cores with hyperthreading). The actual CPU count is 8 on the AMD chip. You don't need an FPU to run basic operating system threads themselves. It also has more sy
  • by glsunder ( 241984 ) on Sunday November 08, 2015 @06:36PM (#50889691)

    If you're a consumer, you can't base your buying decision based on a single specification. You have to look at your needs vs what is important to you. For some people performance/$ is important. For others, performance $/watt is more important. You have to compare based on the applications that are important to you. If you, as a typical desktop/laptop user, mostly use application A and price is the main consideration, it doesn't matter if the CPU runs at 3 or 4 GHz, 4 threads or 8 threads, etc. What matters is performance/$. If you have $200 to spend on a cpu, it really doesn't matter who makes the better $700 cpu.

    There are plenty of resources available to help people make decisions. Only relying on marketing department information is just plain dumb.

  • by friedmud ( 512466 ) on Sunday November 08, 2015 @11:17PM (#50890635)

    The fact that, years later, _WE_ are still arguing about this proves that the case has merit.

    If WE can't come to a consensus about this... then how is Joe Scmoe supposed to figure it out?

    The fact is: this was _misleading_ advertising. They could have easily come up with another name for it (like Intel did with Hyperthreads)... instead they consciously chose to call the extra ALUs _cores_... which does have a meaning to the typical consumer. They did this, on purpose, to muddy the waters... and they REALLY did.

    Does that mean that people shouldn't be more careful about what they buy? Sure. But that doesn't absolve AMD from putting out misleading advertising.

Disclaimer: "These opinions are my own, though for a small fee they be yours too." -- Dave Haynie

Working...