And the belief in such a device is harmful
David Chisnall, University of Cambridge
There is an increasing trend in computer architecture to categorize processors and accelerators as "general purpose." Of the papers published at this year's International Symposium on Computer Architecture (ISCA 2014), nine out of 45 explicitly referred to general-purpose processors; one additionally referred to general-purpose FPGAs (field-programmable gate arrays), and another referred to general-purpose MIMD (multiple instruction, multiple data) supercomputers, stretching the definition to the breaking point. This article presents the argument that there is no such thing as a truly general-purpose processor and that the belief in such a device is harmful.
Many of the papers presented at ISCA 2014 that did not explicitly refer to general-purpose processors or cores did instead refer to general-purpose programs, typically in the context of a GPGPU (general-purpose graphics processing unit), a term with an inherent contradiction.
A modern GPU has I/O facilities, can run programs of arbitrary sizes (or, if not, can store temporary results and start a new program phase), supports a wide range of arithmetic, has complex flow control, and so on. Implementing Conway's Game of Life on a GPU is a fairly common exercise for students, so it's clear that the underlying substrate is Turing complete.
Here is one definition of a general-purpose processor: if it can run any algorithm, then it is general purpose. This is not a particularly interesting definition, because it ignores the performance aspect that has been the driving goal for most processor development.
It's therefore not enough for a processor to be Turing complete in order to be classified as general purpose; it must be able to run all programs efficiently. The existence of accelerators (including GPUs) indicates that all attempts thus far at building a general-purpose processor have failed. If they had succeeded, then they would be efficient at running the algorithms delegated to accelerators, and there would be no market for accelerators.
With this in mind, let's explore what people really mean when they refer to a general-purpose processor: the specific category of workloads that these devices are optimized for and what those optimizations are.
The Operating System
A common requirement for a general-purpose processor is that it can run a general-purpose operating system—meaning an operating system that is either Unix or (like MS Windows) has a similar set of underlying abstractions to Unix. For example, most modern processors lack the ability to cleanly express the memory model found in Multics, with fine-grained sharing and transparent virtualized access to memory-mapped I/O devices.
The ability to run an operating system is fundamental to the accepted definition. If you remove the ability to run an operating system from a processor that is considered general purpose, then the result is usually described as a microcontroller. Some devices that are now regarded as microcontrollers were considered general-purpose CPUs before the ability to run a multitasking, protected-mode operating system became a core requirement.
The important requirements for running an operating system (again, with the caveat that it must be efficient and not simply running an emulator) are similar to those enumerated by Gerald Popek and Robert Goldberg for virtualization.6 This is not surprising, as an operating-system process is effectively a virtual machine, albeit with a very high-level set of virtual devices (file systems rather than disks, sockets rather than Ethernet controllers).
To meet these requirements, a processor needs to distinguish between privileged and unprivileged instructions and the ability to trap into privileged mode when a privileged instruction is executed in unprivileged mode. Additionally, it needs to implement protected memory so that unprivileged programs can be isolated from each other. This is somewhat weaker than Popek and Goldberg's requirement that memory accesses be virtualized, because an operating system does not necessarily have to give every process the illusion that it has an isolated linear address space.
Traditionally, all I/O-related instructions were privileged. One of the advantages of memory-mapped I/O regions, as opposed to separate namespaces for I/O ports, is that it makes it possible to delegate direct access to hardware to unprivileged programs. This is increasingly common with network interfaces and GPUs but requires that the hardware provide some virtualization support (e.g., isolated command queues that cannot request direct memory access to or from memory controlled by another program).
In MIPS, the privileged instructions are modeled as being implemented by CP0 (control coprocessor). This dates back to the pre-microcomputer model where the CPU consisted of a small number of chips and the MMU (memory management unit) was usually placed on a different chip (or even collection of chips).
In a modern multicore system, it's fairly common to allocate some cores almost purely to user-space applications, with all interrupts routed away from them so that they can run until a scheduler on another core determines that they should stop. It's therefore not clear whether every processor in a multicore system needs to be able to run all of the kernel code. This model was explored in the Cell microprocessor,3 where each SPE (synergistic processing element) was able to run a small stub that would message the PowerPC core when it needed operating-system functionality. This design is relatively common in supercomputers, where each node contains a slow commodity processor to run an operating system and handle I/O, as well as a much faster processor optimized for HPC (high-performance computing) tasks, over which the application has exclusive control.
The Barrelfish research operating system1 proposes that, as the transition is made from multicore to many-core architectures, time-division multiplexing makes less sense for virtualization than space-division multiplexing. Rather than one core running multiple applications, one application will have exclusive use of one or more cores for as long as it runs. If this transition takes place, it will change the requirements considerably.
The heuristic used in the RISC I processor,5 discovered by examining the code produced by the Portable C Compiler on the Unix code base, was that general-purpose code contains approximately one branch instruction for every seven instructions. This observation indicates why branch predictors are such an important feature in most modern superscalar pipelined CPUs. The Pentium 4, for example, could have around 140 instructions in-flight at a time and thus (if this observation is correct) needed to predict correctly 20 branches into the future to keep the pipeline full. With a 90 percent branch predictor hit rate, it would have only a 12 percent probability of correctly predicting the next 20 branches.
Fortunately for the Pentium 4, not all branches are of equal difficulty and the most common branch in CPU-dependent code is the backwards branch at the end of a loop. The Pentium 4 included a static heuristic that this branch is always taken on first use, and a complex loop predictor would determine which iteration would be the last one.
Consider the simple expression a+b. In C, you statically know the types of both a and b. The operation is either an integer or floating-point addition. Now consider the same line of code in C++. If a is a primitive type, then this is a simple arithmetic operation (although there may be some complex code including function calls to coerce b to the same type). If a is a class that has a virtual overload for the addition operator, then it is an indirect branch via a vtable (virtual table) lookup. The compiler will insert a load of the function pointer at a fixed offset in the vtable, followed by an indirect jump to the loaded address.
It's worth noting that SPARC architecture has a tagged integer type and add and subtract instructions specifically designed for dynamic languages.7 A SPARC tagged integer is 30 bits, with the lowest two bits in a 32-bit word reserved for type information. The tagged operations set a condition code (or trap) if either of the operands has a nonzero tag or if the result overflows.
Locality of Reference
One feature shared by most modern CPUs is their large caches. In the case of Itanium and POWER, this trend has gone beyond the realm of propriety, but even commodity x86 chips now have more cache than a 386 or 486 had RAM.
Cache, unlike scratchpad memories, is (as its name implies) hidden from the software and exists to speed up access to main memory. This ability to stay hidden is possible only if the software has a predictable access pattern. The strategy most commonly used for CPU caches optimizes for locality of reference. The caches are split into lines of typically 64 or 128 bytes. Loading a single byte from main memory will fill an entire cache line, making accesses to adjacent memory cheap. Often, caches will fill more than one adjacent line on misses, further optimizing for this case.
This design is a result of the working-set hypothesis, which proposes that each phase of computation accesses a subset of the total memory available to a program, and this set changes relatively slowly.2 GPUs, in contrast, are optimized for a very different set of algorithms. The memory controller on an Nvidia Tesla GPU, for example, has a small number of fixed access patterns (including some recursive Z-shapes, commonly used for storing volumetric data); it fetches data in those patterns within a texture and streams it to the processing elements. Modern Intel CPUs can also identify some regular access patterns and efficiently stream memory but with a lower throughput.
Jae Min Kim et al. showed that some programs run faster on the slow cores in ARM's big.LITTLE architecture.4 Their explanation was that the slow Cortex A7 cores have a single-cycle access to their level-1 caches, whereas the fast Cortex A15 cores have a four-cycle penalty. This means that, if the working-set hypothesis holds—the working set fits into the L1 cache, and the performance is dominated by memory access—then the A7 will be able to saturate its pipeline, whereas the A15 will spend most of its time waiting for data.
This highlights the fact that, even within commodity microprocessors from the same manufacturer and generation, the different cache topologies can bias performance in favor of a specific category of algorithm.
Models of Parallelism
Parallelism in software comes in a variety of forms and granularity. The most important form for most CPUs is ILP (instruction-level parallelism). Superscalar architectures are specifically designed to take advantage of ILP. They translate the architectural instruction encodings into something more akin to a static single assignment form (ironically, the compiler spends a lot of effort translating from such a form into a finite-register encoding) so that they can identify independent instructions and dispatch them in parallel.
Although the theoretical limit for ILP can be very high, as much as 150 independent instructions,8 the amount that it is practical to extract is significantly lower. VLIW (very long instruction word) processors simplify their logic by making it the compiler's job to identify ILP and bundle instructions.
The problem with the VLIW approach is that ILP is a dynamic property. Recall that the RISC I heuristic was that branches occur, on average, every seven instructions. This means that the compiler can, at most, provide seven-way ILP, because it cannot identify ILP that spans basic blocks: if the compiler statically knew the target of a branch, then it would most likely not insert a branch.
VLIW processors have been successful as DSPs (digital signal processors) and GPUs but not in the kind of code for which "general-purpose" processors are optimized, in spite of several attempts by Intel (i860, Itanium). This means that statically predictable ILP is relatively low in the code that these processors are expected to run. Superscalar processors do not have the same limitation because, in conjunction with a branch predictor, they can extract ILP from dynamic flow control spanning basic blocks and even across functions.
It's worth noting that, in spite of occupying four times the die area and consuming four times the power, clock-for-clock the ARM Cortex A15 (three-issue, superscalar, out-of-order) achieves only 75-100 percent more performance than the (two-issue, in-order) A7, in spite of being able (theoretically) to exploit a lot more ILP.
The other implicit assumption with regard to parallelism in most CPUs is that communication among parallel threads of execution is both rare and implicit. The latter attribute comes from the C shared-everything model of concurrency, where the only way for threads to communicate is by modifying some shared state. A large amount of logic is required in cache coherency to make this efficient, yet it is relevant only for shared-memory parallelism.
Implementing message passing (as embodied both by early languages such as Occam and Actor Smalltalk and by newer ones such as Erlang and Go) on such processors typically involves algorithms that bounce cache-line ownership between cores and involve large amounts of bus traffic.
The general-purpose processors of today are highly specialized and designed for running applications compiled from low-level C-like languages. They are virtualized using time-division multiplexing, contain mostly predictable branches roughly every seven instructions, and exhibit a high degree of locality of reference and a low degree of fine-grained parallelism. Although this describes a lot of programs, it is by no means exhaustive.
Because processors optimized for these cases have been the cheapest processing elements that consumers can buy, many algorithms have been coerced into exhibiting some of these properties. With the appearance of cheap programmable GPUs, this has started to change, and naturally data-parallel algorithms are increasingly run on GPUs. Now that GPUs are cheaper per FLOPS (floating-point operations per second) than CPUs, the trend is increasingly toward coercing algorithms that are not naturally data parallel to run on GPUs.
The problem of dark silicon (the portion of a chip that must be left unpowered) means that it is going to be increasingly viable to have lots of different cores on the same die, as long as most of them are not constantly powered. Efficient designs in such a world will require admitting that there is no one-size-fits-all processor design and that there is a large spectrum, with different trade-offs at different points.
1. Baumann, A., Barham, P., Dagand, P.-E., Harris, T. L., Isaacs, R., Peter, S., Roscoe, T., Schüpbach, A., Singhania, A. 2009. The multikernel: a new OS architecture for scalable multicore systems. In Proceedings of the ACM SIGOPS 22nd Symposium on Operating Systems Principles: 29-44.
2. Denning, P. J. 1968. The working-set model for program behavior. Communications of the ACM 11(5): 323-333.
3. Gschwind, M., Hofstee, H. P., Flachs, B., Hopkins, M., Watanabe, Y., Yamazaki, T. 2006. Synergistic processing in Cell's multicore architecture. IEEE Micro 26(2): 10-24.
4. Kim, J. M., Seo, S. K., Chung, S. W. 2014. Looking into heterogeneity: when simple is faster. In The 2nd International Workshop on Parallelism in Mobile Platforms.
5. Patterson, D. A., Sequin, C. H. 1981. RISC I: a reduced instruction set VLSI computer. In Proceedings of the 8th Annual Symposium on Computer Architecture: 443-457.
6. Popek, G. J., Goldberg, R. P. 1974. Formal requirements for virtualizable third-generation architectures. Communications of the ACM 17(7): 412-421.
7. SPARC International Inc. 1992. SPARC Architecture Manual: Version 8. Upper Saddle River, NJ: Prentice-Hall Inc.
8. Wall, D. W. 1993. Limits of instruction-level parallelism. Technical report, Digital Equipment Corporation Western Research Laboratory.
LOVE IT, HATE IT? LET US KNOW
David Chisnall is a researcher at the University of Cambridge, where he works on programming-language design and implementation. He spent several years consulting in between finishing his Ph.D. and arriving at Cambridge, during which time he also wrote books on Xen and the Objective-C and Go programming languages, as well as numerous articles. He also contributes to the LLVM, Clang, FreeBSD, GNUstep, and Étoilé open source projects, and he dances the Argentine tango.
© 2014 ACM 1542-7730/14/1000 $10.00
Originally published in Queue vol. 12, no. 10—
see this item in the ACM Digital Library
Mohamed Zahran - Heterogeneous Computing: Here to Stay
Hardware and Software Perspectives
Hans-J Boehm, Sarita V. Adve - You Don't Know Jack about Shared Variables or Memory Models
Data races are evil.
Satnam Singh - Computing without Processors
Heterogeneous systems allow us to target our programming to the appropriate environment.
Dorian Birsan - On Plug-ins and Extensible Architectures
Extensible application architectures such as Eclipse offer many advantages, but one must be careful to avoid "plug-in hell."
(newest first)I genuinely hope to be able to prove this wrong at some point. I'm not looking for today's or tomorrow's claimed raw performance, that's a limiting perspective - hardware can and will improve and be optimized for any purpose. It is good to give hardware people something to do.
The software and microarchitecture is the limiting factor here. We are too lazy to re-implement an entire platform from scratch (Itanium anyone? or all the countless obscure or failed chipsets?) .. we want hardware that will run our existing applications. So we virtualize, translate, emulate, etc.. because we're lazy. But that's fine - what works, works.
But if your applications were all just libraries in your environment, and errors created data, and there was no such thing as compiling...
poxix.com/eal/docWhen I think "special purpose", I think "provides an order of magnitude or more advantage at the particular thing it's useful for." That doesn't exclude using it for other purposes. For example, GPUs are very, very good at graphics, easily a couple orders of magnitude beyond the host CPU in terms of efficiency.
Can I compile almost arbitrary programs and run them on the GPU these days? Sure. But they won't be nearly as efficient. Compute-intensive codes might enjoy a speedup vs. the host CPU, but unless they fit a certain profile, you won't see the same orders-of-magnitude efficiency gains. (Efficiency includes energy efficiency as well as raw performance.) To me, that makes the GPU special purpose.
What's a general purpose processor then? It's a processor that doesn't really stand out as orders of magnitude more efficient for any particular task. Some general purpose processors are more efficient than others (A7 more efficient at pointer chasing than A15, for example), but they're all within small factors of each other. Some are tuned for peak raw performance, others for peak energy efficiency, but they don't have a particular application they're driving to extreme efficiencies.
I also agree with Benoit's point above that at the SoC level, an SoC is easily special purpose even if one or more processors on board are general purpose, if the peripheral mix is specialized. In the business I'm in, we combine ARMs, DSPs, and custom accelerators for FFTs, networking, etc. because we're specializing those chips for a certain set of markets. The same ARM on a different chip could go into a general purpose Linux server too.Very interesting article. It goes through a lot of concepts in this "short" article. I think it is long enough to give a good overview of the situation but far too short to explain in details every aspect of the problem. A lot of interesting points like an history of the chip/computer companies strategies (from specifically purpose built main frames to of the self computers in clusters...),HPC consideration & benchmarks,operating system and compiler design could have given more insight/arguments but they are way out of the scope of this article.
One last note about the definition "war". For me the separation between a micro-controller and a microprocessor lies more in the addition of extra built-in peripherals in a micro-controller than in the internal architecture. Taking into account the peripherals, a micro-controller is never a general purpose processor even it it contains one.
Industry trends to leave special purpose architecture (transputer,Freescale DSP, PIC...) to more standard 32 bits "multi purpose" architecture based most of the time on ARM core. I think we lost a lot of diversity in the process.
Finally, an article has to be a little be catchy and only represent the point of view of the author.General Purpose means that the processor can be used for any purpose that can be fit into it's available memory. If it is Turing Complete, then it is General Purpose. This is as opposed to a Spcial Purpose processor. Such things can be built perform specific functions much faster than any General Purpose Processor. to Some are built including both general and special purpose processor The math co-processor implemented it's math functions in gate logic, to speed up the mathematical operations. It did not, however include some of the functions required for Turing completeness. Early Linux systems often included a library to mimic the math co-processor for systems which didn't include one. From the 80486 onward, Intel processors have included the math co-processor on the same chip with the general purpose chip. To me, the Author here is pointing out that most of the current 'special purpose' processors are actually also general purpose. The distinction thus becomes meaningless. @ET3D: In the essay form of writing, it is not uncommon for the title to be merely the starting point for a possibly wide-ranging discourse. This can result in a discursive ramble, but when done well, it can lead readers to a point of view they had not considered before. If you approach such essays with the attitude of grading them on how closely they adhere to the issue raised by the title, you will miss some interesting reading. That's an awfully long article to write on a definition. You say that the definition of "general purpose processor" has changed over time, and now doesn't include microcontrollers. I feel that it's your own definition and you're trying to press a point that you yourself have created. If you look at "general purpose" in the more usual sense of "not geared toward a narrow field of operation", then there is a lot of sense in general purpose processors. That is why microcontrollers, which are general purpose processors, are so successful. "General purpose" doesn't mean "one size fits all".
While this article offers some interesting analysis, I feel that it would be stronger if you dropped the argument with the "general purpose" definition. It might make for a catchy title, but otherwise is off-putting.