ABOUT BENCH FORUMS PODCAST LOGIN REGISTER
PC Components
Smartphones & tablets
Systems
ENTERPRISE & IT
GUIDES

Understanding the Cell Microprocessor

 by Anand Lal Shimpi on March 17, 2005 12:05 AM EST
70 Comments | Add A Comment
70 Comments

Cell’s In-Order Architecture

We have mentioned that both the PPE and SPEs are in-order cores, but in order to understand the impact of an in-order core on performance, there’s a bit of background knowledge that we have to go over first.

Dependencies, Instruction Ordering and Parallelism

What are Dependencies?

In many of our past CPU articles, we’ve brought up this idea of dependencies as seen by the CPU.   At the very basic level, a CPU is fed a stream of instructions that are generally of the form:

OP destination, source1, source2, ... , source n

The instruction format will vary from one CPU ISA to the next, but the general idea is that the CPU is sent an operation (OP), a destination to store the result of the operation, and one or more sources on which to get data to perform the operation.   Depending on the architecture, the destination and sources can be memory locations or registers.   For the sake of simplicity, let’s just assume that for now, all destinations and sources are registers.

Let’s take a look at an example with some data filled in:

ADD R10, R1, R2

The above line of assembly would be sent to the CPU, telling it to add the values stored in R1 (Register #1) and R2 and store the result in R10.   Simple enough.   Now, let’s give the CPU another operation to crunch on:

MUL R11, R10, R3

This time, we’re multiplying the values stored in R10 and R3, and storing the result in R11.   As a single line of assembly, the above code is easily executed, but when placed directly after our first example, we’ve created a bit of a problem:
  1.      ADD R10, R1, R2
  2.      MUL R11, R10, R3
  3.      ADD R9, R11, R4
Line 1 writes to R10, while Line 2 reads from R10.   Under no circumstances can the CPU begin executing line 2 before line 1 completes - the same goes for lines 3 and 2.   What we’ve created here is what is known as a RAW dependency, Read After Write.   There are many more types of dependencies, but understanding this basic example is more than enough to take us to the next topic at hand - the impact of such dependencies.

The problem with a dependency is that it limits what can be executed in parallel.   Take the Athlon 64, for example.  It has three integer execution units, all of which are equally capable of executing the code (in a slightly revised, x86 assembly format, of course) that we used above.   In theory, the Athlon 64 could execute three lines integer operations in parallel at the same time - assuming that no dependencies existed between the operations.   In executing the above code, two of the Athlon 64’s integer execution units would go idle until the first line of code was executed.

Dependencies, such as the simple one that we talked about above, hinder the ability of modern day microprocessors to function to the best of their abilities.   It’s like having three hands, but only being able to clean your room by picking up one item at a time; frustratingly inefficient.

Cell's On-Die Memory Controller In-Order Architectures
PRINT THIS ARTICLE
Comments Locked

70 Comments

View All Comments

  • scrotemaninov - Thursday, March 17, 2005 - link

    #23: True, but I believe that when the SPE's access the outside memory they go through the cache. Sure it's a lower coherancy than we're used to but it's not much worse.

  • Houdani - Thursday, March 17, 2005 - link

    18: Top Drawer Post.
    20: Thanks for the links!

  • fitten - Thursday, March 17, 2005 - link

    "Given the speed of the interconnect and the fact that it is cache-coherant,"

    Only the PPC core has cache. The individual SPEs don't have cache - they have scratchpad RAM.

    #22: I believe the PPC core is a dual issue core that just happens to be 2xSMT.

  • AndyKH - Thursday, March 17, 2005 - link

    Great article.
    Anand, Could you please clarify something:
    I had the impression that the PPE was a SMT processor in the sense that it had to be executing 2 threads in order to issue 2 instructions per clock. In other words: I didn't think the PPE control logic could decide to issue 2 instructions from the same thread at any given clock tick, but rather that it absolutely needed an instruction from each thread to issue two instructions.

    After reading the article, I don't assume my impression is right, but a comment from you would be nice.

    As I come to think about it, my impression is rather identical to 2 seperate single thread in-order cores. :-)

  • Koing - Thursday, March 17, 2005 - link

    Cell looks VERY interesting.

    Any of you guys seen Devil May Cry 3 on the PS2? Looks great imo same with T5 and GT4.

    Cell at first will be tough like most consoles. BUT eventually THE developers will get around it and make some very solidly good looking games.

    Lets hope they are innovative and not just rehashed graphics and nothing else.

    Thanks for the great article.

    Koing

  • scrotemaninov - Thursday, March 17, 2005 - link

    I really hate just dumping loads of links, but this basically is the available content on the CELL.

    http://arstechnica.com/articles/paedia/cpu/cell-1....
    http://arstechnica.com/articles/paedia/cpu/cell-2....
    http://realworldtech.com/page.cfm?ArticleID=RWT021...
    http://www.blachford.info/computer/Cells/Cell0.htm...

    http://www.realworldtech.com/page.cfm?ArticleID=RW...

    http://www.hpcaconf.org/hpca11/papers/25_hofstee-c...
    http://www.hpcaconf.org/hpca11/slides/Cell_Public_... (slides)

  • mrmorris - Thursday, March 17, 2005 - link

    Brilliant article, there are few places for in-depth hardcore technology presentations but Anandtech never fails.

  • scrotemaninov - Thursday, March 17, 2005 - link

    Real concurrency is hard to do for the programmers. It's a real pain to get it right and it's hard to debug. Systematic analysis just gets too complex as there are just too many states, you end up with a huge graph/markov-model and it's just impossible to solve it tractably.

    Superscalar and SMT just try to increase ILP at the CPU level without burdening the programmer or compiler-writer. However, we've pretty much come to the end of getting a CPU to go faster - at 5GHz, LIGHT travels 6cm between clocks, and an electic PD will travel slower. As it is, in the P4 pipeline, there are at least 2 stages which are simply there to allow signals to propogate across the chip. Clearly, going faster in Hz isn't going to make the pipeline go faster.

    So the ONLY thing that they can do now is to put lots of cores on the same chip and then we're going to have to deal with real concurrency. IBM/Sony are doing it now with CELL and Intel will do it in a few years. It's going to happen regardless. What we need is languages which can support real concurrency. The Java Memory Model is an almost ideal fit for the CELL, but other aspects don't work out so well, maybe. We need Pi-calculus/Join-calculus constructs in languages to be able to really deal with these cpus efficiently.

    Your comments about CELL not being general purpose enough are a little wrong. IBM /already/ has the CELL in workstations and are evaluating applications that will work well. Given the speed of the interconnect and the fact that it is cache-coherant, I think we'll be seeing super-computers based on many CELLs, it's an almost ideal fit (as it is, you've almost got ccNUMA on a single chip). Also, bear in mind that this is IBM's 5th (or 6th?) generation of SMT in the PPE - they've been at it MUCH longer than Intel - IBM started it in the mid-90s around the same time that the Alpha crew were working on the EV8 which was going to have 8-way thread-level parallelism (got canned sadly).

    Also, if you look at IBMs heavy CPUs - the POWER5, that has SMT and dispatches in groups of 8 instructions, not the 3/4 that AMD/Intel manage.

    What I'm saying here, is that sure, the SPEs don't have BPTs of BTBs, they're all 2-way dispatch and not greater, but, they all run REALLY fast, they have short pipelines (so the pain of the branch misprediction won't be so bad), and, IBM have had software branch prediction available since the POWER4, so they've been at it a few years and must have decided that compilers really can successfully predict branch directions.

    Backwards compatibility doesn't matter. Sure, Microsoft took several years to support AMD64 but that didn't stop take up of the platform - everyone just ran Linux on it (well, everyone who wanted to use the 64bit CPU they'd bought). It'll only be a few months after the CELL is out that we'll have to wait until Linux can be built on it. 100quid says Microsoft will never support it.

    Frankly, considering that it's far more likely to go into super-computer or workstation environments, no one there gives a damn about backwards compatibility or Windows support. No one in those environments /wants/ a damn paper clip.

  • Reflex - Thursday, March 17, 2005 - link

    #14: Replace 'lazy developers' with 'developers on a budget' and you will have a true statement. Its not an issue of laziness, its an issue of having the budget to optimize fully for a platform.

  • GhandiInstinct - Thursday, March 17, 2005 - link

    Wow Super CPU and SUPER RAMBUS? AHHHH!

    This will replace my computer. PS3 that is.
Copyright © 2024. All rights reserved.
BENCH
TOPICS
FOLLOW
ABOUT

Log in

Don't have an account? Sign up now