Announcement

Collapse
No announcement yet.

How A Raspberry Pi 4 Performs Against Intel's Latest Celeron, Pentium CPUs

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #61
    Originally posted by vladpetric View Post
    Unfortunately the microarchitecture of the RPi4 is still a bit of a joke.

    I've had a lengthy (and peppered with the occasional insults, of course) discussion earlier on phoronix forums.

    It's not just the clock speed. The IPC of the RPi4 is mediocre.

    The RPi4 chip (Cortex A72) doesn't have a proper Load Store Queue, to allow it to issue loads out-of-order with respect to earlier stores (so all prior stores need to complete before issuing a load). This would require machinery to detect and correct potential misspeculations, of course.

    It can only decode 3 instructions per cycle, and issue 5.

    Actually the venerable Alpha 21264 from 1996 had better microarchitecture. Granted, that was a trailblazer chip, but still ... 1996!

    So, this is not about nm. It's about a toy microarchitecture.
    Let's be realistic here, we are paying toy prices. It's built for a price and not for sheer performance. You want the fastest ARM then you are more than welcome to pay Apple prices and good luck breaking out of their garden.

    Comment


    • #62
      Originally posted by starshipeleven View Post
      Is it really?
      ARM is an architecture that is supposed to scale up into the high-performance too.

      The Raspi and embedded devices in general are designed to be efficient, but there are high-end ARM CPUs that really need a real heatsink.
      Yeah, old ARM32 has some limitations and made it less than ideal, but no worse than the crap i386 was hampered with, but AArch64 is better designed. And instruction sets doesn't matter that much in the end with high end processors with decoding, queuing and out-of-order execution.

      Comment


      • #63
        Originally posted by vladpetric View Post

        Apple processors only use the instruction set - they designed their cpus from scratch, and didn't use the toy reference designs from ARM.
        That is just a marketing lie. They still had the reference designs and components. They have built up their new overall architecture from scratch, but they have all the pre-built components that they bought from ARM. So a new architecture "from scratch", but not a new CPU design from scratch. Or about as new as a new architecture from AMD or Intel.

        Comment


        • #64
          Originally posted by schmidtbag View Post
          According to what?
          According to the fact that it actually does and there are real-life examples of high-performance ARM CPUs?

          I'm aware, but they're practically a wholly different product, as are Apple's CPUs. So it doesn't make sense to compare a low-end SBC like that, which is what's going on here.
          In your posts so far you were sayin "ARM this" and "ARM that", you were talking about a whole instruction set or ISA, not specific low-power implementation in Raspi.

          There is a difference between the instruction set and the actual implementation. Raspi CPU implements ARM instruction set in a way to make it low power. But there is nothing in ARM instruction set that limits it to low power. Apple and Amazon CPUs aren't crappy low-power SBCs

          we're talking about a $35 computer here.
          No we are not. You said "ARM isn't built to compete with desktop performance." and "ARM isn't built to be performant. It's built to be efficient." and more along these lines, which is bullshit as I already said. There are ARM CPUs that are built to be performant too.

          Maybe you wanted to say "Raspi" instead of "ARM" in those statements, but that's your own mistake.
          Last edited by starshipeleven; 08 August 2020, 05:17 AM.

          Comment


          • #65
            Originally posted by CommunityMember View Post

            Also don't forget the Ampere 80 core processor (designed for the hyperscale cloud providers). It was supposed to ship around now in volume (but, again, only to the hyperscalers, and I would not be surprised the schedule has been "adjusted").
            I'm counting only stuff that is confirmed "real" so to speak. Apple and Amazon and higher-end embedded from Marvell and NXP are very real, the Ampere is still in limbo for now.

            Comment


            • #66
              Originally posted by wizard69 View Post

              Actually I think a lot more is going on here than Marketing. Apple has a vision for its computing hardware and ARM is only a small partner here. I'm pretty much ocnvinced after viewing a few WWDC videos this year that ARM is just an convenience to them on their journey to far more interesting platforms.
              Yeah, that makes sense. I wouldn't be too surprised if ARM is something they went with because it's ready here and now while their R&D teams work on getting some sort of low powered OpenPOWER design ready for mass use for their laptop, phone, and portable lines. It makes sense for them to unify the CPU architecture used across all their product lines so it'll be interesting to see if they stick with ARM long-term or if it's just an intern stop-gap measure on the road to something else.

              Comment


              • #67
                Originally posted by starshipeleven View Post
                According to the fact that it actually does and there are real-life examples of high-performance ARM CPUs?
                What a terrible argument. Just because something can be a certain way, doesn't mean it was ever intended to be. Unlike you, I take my facts from ARM themselves:
                https://www.arm.com/products/silicon-ip-cpu
                "Supreme performance at optimal power". "Reliable mission-critical performance" (something where more power isn't a priority). "Powering the most energy-efficient embedded devices".
                Neoverse is scalable, in the sense that you can add more cores or add more to the cores; that doesn't mean the cores are designed to be pushed harder. And even then, they tout it's efficiency, because y'know... that's their priority.
                SecuCore is basically just a rebrand of their M series; nothing worth noting and certainly not built to be performant.
                Ethos again has a heavy focus on efficiency.
                Nowhere does it mention things like IPC, clock speeds, or how the performance can be compared, because they know it's not impressive, and that's not why you buy them.
                For companies like Amazon, Apple, and Nvidia, they don't push the ARM cores very hard, because they know it won't serve them well. They just either add more cores or they add instructions. Qualcomm is, from my recollection, the first company to readily breach 3GHz, and I wouldn't say their relatively high-clocked CPU is very impressive, because it's also relatively power-hungry. Because like I've been saying - the architecture doesn't scale up very well in clock speed.
                The Cortex-A78 is, to my knowledge, the best ARM has to offer. Again, all they're focusing on is efficiency:
                https://www.arm.com/products/silicon...x-a/cortex-a78
                The scalability they mention isn't about pushing high clock speeds.
                I wasn't able to find it, but I do remember reading an article about how ARM tries to make sure that they only add performance if they can remain within the same power envelope. That doesn't mean the licensees care to follow that trend, but my point is ARM themselves, y'know, the core designers of the architecture, do not imply the cores are built to scale up. They can but they're not meant to and they're not typically good at it.
                There is a difference between the instruction set and the actual implementation. Raspi CPU implements ARM instruction set in a way to make it low power. But there is nothing in ARM instruction set that limits it to low power. Apple and Amazon CPUs aren't crappy low-power SBCs
                They're not power hungry beasts either. They're still super efficient and not clocked very high, they just have a lot of cores and instructions. ARM actually becomes more efficient when you add more cores (do not confuse efficiency with power draw).
                No we are not. You said "ARM isn't built to compete with desktop performance." and "ARM isn't built to be performant. It's built to be efficient." and more along these lines, which is bullshit as I already said. There are ARM CPUs that are built to be performant too.
                Well, I have sources and you don't.
                Last edited by schmidtbag; 08 August 2020, 09:24 AM.

                Comment


                • #68
                  Originally posted by bnolsen View Post

                  oof i've seen so,e yotube reviews on the pi and the 64 bit clearly outperforms 32 bit.
                  The Cortex-A53/57 (armv8) were this first 64 bit arm chips, so 64bit distro's use that as a minimum instruction set when compiling.

                  The 32bit disto's use armv6.

                  If you compile a 32bit disto to use the Cortex-A53 as minimum, then it will be just as fast or even faster that the 64bit OS.

                  This is done on Alpine linux 3.12.0 32bit on a RPI 3-B (non-plus):

                  zpm:~# clang -O3 prime.c -o prime
                  zpm:~# time ./prime
                  664580
                  real 3m 36.44s
                  user 3m 36.42s
                  sys 0m 0.00s

                  zpm:~# clang -O3 -mcpu=cortex-a53 prime.c -o prime
                  zpm:~# time ./prime
                  664580
                  real 0m 39.16s
                  user 0m 39.16s
                  sys 0m 0.00s

                  Last edited by Raka555; 08 August 2020, 10:25 AM.

                  Comment


                  • #69
                    The other thing that is holding the RPIs back, is the bus. They use USB vs PCIe on AMD/Intel.
                    No doubt the price will skyrocket if they give the RPI an PCIe bus, but so will some of the performance.

                    Comment


                    • #70
                      Originally posted by schmidtbag View Post
                      What a terrible argument. Just because something can be a certain way, doesn't mean it was ever intended to be.
                      ARM instruction set was born 35 fucking years ago in a completely different world, how can you make these bold claims about what it was "intended to be" based on PR releases of current ARM CPU designs (ARM CPU designs are specific implementations of ARM instruction set) and extrapolate that ARM instruction set as a whole was ever intended to be only for low power?

                      Current ARM CPU designs have been tailored towards power efficiency because that's the market niche they managed to get into, and what anyone will want to buy from ARM Holdings company, but this does not tell us anything about the architecture's ability or original goals.

                      While the fact that high-performance ARM CPUs tells us that the instruction set does indeed allow high performance designs too.

                      For companies like Amazon, Apple, and Nvidia, they don't push the ARM cores very hard, because they know it won't serve them well.
                      Which is why for example Amazon's ARM CPUs have similar performance to high end server AMD and Intel CPUs on single-core tests https://www.anandtech.com/show/15578...ntel-and-amd/5 while having a TDP that is 50-100w lower and the reviewer said "Amazon was able to deliver on its promise of 40% better performance per dollar, and it’s a massive shakeup for the AWS and EC2 ecosystem." Intel would throw fucking infants into a big fire to get a 40% better "performance per dollar" ratio on their high end CPUs.

                      Meanwhile, Apple has decided to drop x86 and migrate to their ARM CPUs, after years of benchmarks that showed how their ARM CPUs were more or less on par with Intel's 15w laptop parts.

                      What about NVIDIA? Oh they just want to buy ARM whole https://www.zdnet.com/article/nvidia...ks-to-buy-arm/

                      Plus there are Ampere and Nuvia companies that are designing new ARM CPUs for high performance segment.

                      Yeah, they are totally not pushing or betting on ARM CPUs. It's all an hallucination.

                      They just either add more cores or they add instructions.
                      Or design their own custom ARM CPU implementations when they want actual performance because ARM CPU designs licensed from ARM are designed to be sold to embedded.

                      Qualcomm is, from my recollection, the first company to readily breach 3GHz, and I wouldn't say their relatively high-clocked CPU is very impressive, because it's also relatively power-hungry. Because like I've been saying - the architecture doesn't scale up very well in clock speed.
                      The Cortex-A78 is, to my knowledge, the best ARM has to offer. Again, all they're focusing on is efficiency:
                      https://www.arm.com/products/silicon...x-a/cortex-a78
                      More and more proof that you don't understand the difference between CPU core IMPLEMENTATIONS that ARM is selling, and ARM instruction set.

                      ARM Holdings designed ARM CPU cores, and sells the license to use these CPU designs to embedded SoC manufacturers. They also design ARM cores for microcontrollers, and sell the license to use these to microcontroller manufacturers. Qualcomm is just buying ARM CPU designs and slapping them into their products. A Cortex-A78 is the same thing, be it in a Qualcomm, Huawei or Broadcomm or NVIDIA SoC.

                      But ARM Holdings also sold much more expensive ARM instruction set licenses, that allows anyone to make their own CPU design using the ARM instruction set. Similar to what Intel did with AMD and VIA. AMD and VIA CPUs are completely different from Intel CPUs even if they use the same x86 instruction set. Different implementations. Limits of one implementation are not limits of another, which is why for example VIA CPUs were so good at low-power back in the day when Intel had only blast furnaces.

                      Apple and Amazon's ARM CPUs aren't using Cortex or whatever CPU licensed from ARM. They designed their own CPUs using ARM instruction set, because they have the ARM instruction set license.

                      Same for Ampere and Nuvia startup.
                      Well, I have sources and you don't.
                      Yeah, sources that clearly explain how you can't tell the difference between a CPU implementation and a CPU instruction set.

                      Comment

                      Working...
                      X