POWER guys commented on this years ago ...

By: anon (anon.delete@this.anon.com), August 17, 2014 2:40 am
Room: Moderated Discussions
Aaron Spink (aaronspink.delete@this.notearthlink.net) on August 17, 2014 12:42 am wrote:
> anon (anon.delete@this.anon.com) on August 17, 2014 12:10 am wrote:
> > I'd like to see actual evidence of this. Of course we've
> > all seen graphs showing the opposite -- that release
> > consistency (weak ordering with release barriers) is more performant than strong and processor ordering.
> >
> Most of the evidence is anecdotal in nature. Some is this is because processors that are strongly ordered
> tend to spend some extra hardware resources to increase performance (most x86), second because the programmers
> tend to be extra cautious on weakly ordered architectures because they've been burned too much.
>
> > On the other hand, those have tended to be on academic papers, and the most recent one I've seen
> > is quite old. So I don't put a lot of stock in those, for modern CPUs and real applications.
> >
> As I said, this tends to be one of those areas where theory and practice
> don't exactly mesh. Academic papers tend to me more theory.

Right, but anecdotes and practice are notoriously unreliable as well. Even from experts, with no offense intended to you or Linus.

>
>
> > That said, I've never seen anywhere evidence (except anecdotal and not very convincing) evidence
> > of the opposite. Keep in mind that implementations can always choose to implement stronger ordering
> > than specified (I think I've read that many ARM implementations do this, for example). Possibly
> > that is done for compatibility and/or ease of hardware implementation. But if it were a performance
> > advantage, you would expect to see IBM going that way. Because they actually *do* support the option
> > for x86-like ordering in recent POWER CPUs, but they keep their weak ordering around.
> >
> Even if they implement stronger ordering than specified, it is just laying a mine field in the future. One
> example is the EV5 to EV6 transition, where EV5 was much more strict than the actual architecture which resulted
> in numerous hard to track down bugs when EV6 released. EV6 was WaA but the software wasn't.

The point is that there is no minefield in going the *other* way. Maybe they create a minefield if they ever want to switch back, but if it is unquestioningly better to go with stronger ordering, then they have no barrier to make the transition.

> > It would actually be an interesting study to run native powerpc code under strong ordering mode vs weak,
> > and examine the performance difference. Obviously it would just be a single data point, not evidence of
> > general case, because different levels of optimization probably went into it (Intel probably spent far
> > more effort to optimize processor ordering than IBM did!). But still it would be interesting, IMO.
> >
> One effect that contributes to the issue is that architects for weakly ordered
> MOMs tend to not devote resources to having the corner cases be fast.

Right, but (at least ideally), that's an opportunity cost. They have more resources to spend on having the regular cases be fast.

> > I'm not sure if I agree. It may be true that they tend to introduce more bugs, but really
> > if they don't know what they are doing, they really should just use synchronization libraries
> > and language features, and those should generally get the ordering details correct.
> >
> IIRC, Linus has commented on this topic in the past on this very board. They tend to be more conservative
> with weakly ordered architectures. AKA, they'll put in extra barriers than possibly needed just in case
> or because trying to figure every case is far too complicated. So when in doubt, use a barrier.

Not... really. I mean, in practice sometimes, but that should not really happen.

In core code (core kernel/compiler/library) doing that indicates that you do not understand what the data races are, and therefore your code is probably buggy *anyway*. Even if you hit it with a big barrier, there could easily be another race or synchronization bug lurking somewhere.

If peripheral (e.g., driver) or application code is doing that, then it indicates the synchronization APIs are inadequate.

I know Linux's driver APIs and potential memory ordering semantics were notoriously shit, which is why, indeed, you would get drivers failing to work on powerpc, or being utterly riddled with wmb() between every statement.

That could have largely been avoided with better APIs. Don't tell driver writers, "err, stores to cacheable and MMIO regions are unordered with respect to one another, so use barriers". Give them APIs and calls that match the steps of their process-flow when setting up driver operations. Begin DMA initiation. Release a lock protecting MMIO registers. etc.

> > I'm not saying that this always happens, but I've never heard of anybody programming
> > portable code and getting worried about a weakly ordered CPU and then declining to attempt
> > some optimization that they otherwise would have. People who don't understand this probably
> > don't understand either that x86 can reorder loads ahead of stores too.
> >
> It manifests itself generally as more barriers in code around anything that could be a corner
> case combined with weakly ordered CPUs generally handling barriers poorly. Strongly ordered
> CPUs effectively have barriers everywhere so the hardware tends to handle them much better.

mfence is heavyweight on x86 too, just saying. If they don't understand what they are doing, and the spectre of the weakly-ordered boogyman is haunting them and making them put barriers everywhere, then they probably did the x86 port a favor too, although they slowed it down in the process, they got rid of some obscure bugs in it too.

> > > So far, the
> > > performance/strictness relationship with MOMs has been a case of Theory != Practice.
> >
> > I'm really not sure that's the case, because it's very hard to know the details. We may not know whether
> > Intel spends 1% additional power on a larger store queue because in-order retirement of stores effectively
> > increases the amount of time stores have to remain. We don't know if speculative reordering of loads
> > costs them a small additional amount of complexity or power due to failed speculation, etc.
> >
> > We (for most values of 'we' that comment on this topic in this forum) see the software
> > side of it. Which *appears* like memory barriers are slow and annoying to use.
> >
> That's because barriers are generally slow and annoying to use and weakly ordered
> architectures need more of them.

And *that's* because you did not see the cost earlier in the code flow, because the ordering was relaxed. Intel's store buffer filled up and stalled on a remotely held cacheline 200 instructions ago, so you did not see the cost of the fence.

> Strongly ordered architectures need less barriers
> and because its effectively a common case, handle it much better.
>
>
> > And for the case of low level synchronization details, programmers be damned (to some extent). The average
> > app programmer will never even grasp x86 ordering rules anyway,
> > so it's a lost cause. Have them use language/library
> > features, and everyone is happy. The language and library programmers who are incapable of understanding
> > power rules, probably aren't the correct people to be writing the x86 code either.
> >
> Except even the "good" programmers aren't that "good" at dealing with memory order models and sync.

Right. So they should use libraries and languages. Seriously. Everyone goes on about making the programmer's life easier. The way to make it easier is to have them use those things instead of being cowboys and doing retarded shit. Actually that applies even to good programmers who do understand such things, because bugs are always inevitable, they should not venture outside of synchronization libraries until they need to. There is actually very little need to, until already much more fundamental problems are solved, such as how to split and recombine work between CPUs. And even then, the availability of language and library primitives to do the job have significantly increased such that, really, almost nobody should ever need or want to access shared memory without going through such features. Just don't do it.

The same goes for x86 too, by the way. Programmers should use synchronization libraries *everywhere*, including ones who know memory consistency quite well, because they *will* get subtle bugs.

> This has been true for almost forever and will most likely remain true for almost ever. It is
> a rather complex area with very very few people who are truly proficient at it. Even on the hardware
> side where the percentage of people who understand it all is much higher, we tend to rely quite
> heavily on formal models and formal proofs for this stuff cause it is so complex. The state space
> for this in a modern coherent system is absolutely massive. Adding a layer of software on top
> basically makes it impossible, so everyone has to err on the side of caution.

Which means going through synchronization APIs.

< Previous Post in ThreadNext Post in Thread >
TopicPosted ByDate
An ARM-based Mac article by GasseeYuhong Bao2014/08/03 01:24 PM
  An ARM-based Mac article by GasseeJoe User2014/08/03 03:51 PM
  An ARM-based Mac article by GasseeExophase2014/08/03 09:06 PM
    An ARM-based Mac article by GasseeMichael S2014/08/03 11:44 PM
      An ARM-based Mac article by GasseeExophase2014/08/04 07:45 AM
        DiscountsRob Thorpe2014/08/04 12:57 PM
          DiscountsAaron Spink2014/08/05 01:34 AM
        An ARM-based Mac article by GasseeAaron Spink2014/08/05 01:32 AM
          Current trajectory favors arm-book in futureanon2014/08/06 01:48 PM
            Current trajectory favors arm-book in futureUngo2014/08/06 04:34 PM
            Current trajectory favors arm-book in futureDoug S2014/08/06 08:21 PM
            The return to irrelevance.Alberto2014/08/07 02:04 AM
    An ARM-based Mac article by GasseeJacob Marley2014/08/04 03:07 AM
  An ARM-based Mac article by Gasseeanon2014/08/03 09:29 PM
    An ARM-based Mac article by GasseeExophase2014/08/03 10:05 PM
      An ARM-based Mac article by GasseeNiels Jørgen Kruse2014/08/03 10:32 PM
    An ARM-based Mac article by Gasseeex-apple2014/08/04 05:20 PM
      An ARM-based Mac article by GasseePatrick Chase2014/08/04 06:51 PM
        Mac Intel urban legendsWes Felter2014/08/05 07:19 PM
          Mac Intel urban legendsUngo2014/08/06 02:09 PM
  An ARM-based Mac article by Gasseejuanrga2014/08/04 01:05 AM
    Why either or, why not both for the Pro line?Doug S2014/08/04 07:01 AM
      Why either or, why not both for the Pro line?Exophase2014/08/04 02:20 PM
        Why either or, why not both for the Pro line?Patrick Chase2014/08/04 06:46 PM
        Why either or, why not both for the Pro line?Doug S2014/08/04 07:14 PM
          Why either or, why not both for the Pro line?Aaron Spink2014/08/05 01:11 AM
            Why either or, why not both for the Pro line?Patrick Chase2014/08/05 07:40 AM
      Why either or, why not both for the Pro line?vvid2014/08/05 08:51 AM
        Why either or, why not both for the Pro line?Doug S2014/08/05 01:11 PM
          Why either or, why not both for the Pro line?Doug S2014/08/05 01:15 PM
          Why either or, why not both for the Pro line?vvid2014/08/05 04:42 PM
            Why either or, why not both for the Pro line?Ungo2014/08/06 05:04 PM
              Why either or, why not both for the Pro line?vvid2014/08/08 03:35 AM
                Why either or, why not both for the Pro line?Ungo2014/08/08 01:58 PM
      Why either or, why not both for the Pro line?juanrga2014/08/05 09:27 AM
        Why either or, why not both for the Pro line?Michael S2014/08/08 03:58 AM
          Why either or, why not both for the Pro line?juanrga2014/08/08 10:00 AM
            Why either or, why not both for the Pro line?Klimax2014/08/08 12:17 PM
            Why either or, why not both for the Pro line?Ungo2014/08/08 03:48 PM
              Why either or, why not both for the Pro line?juanrga2014/08/08 05:40 PM
                Why either or, why not both for the Pro line?David Kanter2014/08/08 10:51 PM
                  Why either or, why not both for the Pro line?juanrga2014/08/09 04:57 AM
                    Why either or, why not both for the Pro line?Aaron Spink2014/08/09 06:43 AM
                      Why either or, why not both for the Pro line?juanrga2014/08/09 07:06 AM
                      Why either or, why not both for the Pro line?Jukka Larja2014/08/17 03:40 AM
                  Why either or, why not both for the Pro line?dmcq2014/08/09 08:50 AM
                  Intel binning advantagePaul A. Clayton2014/08/09 09:58 AM
                  Why either or, why not both for the Pro line?Ronald Maas2014/08/10 06:12 PM
                    Where is the profit?Paul A. Clayton2014/08/10 08:23 PM
                Why either or, why not both for the Pro line?Ungo2014/08/11 02:02 PM
        Why either or, why not both for the Pro line?David Kanter2014/08/08 10:41 PM
          Why either or, why not both for the Pro line?juanrga2014/08/09 05:01 AM
            Why either or, why not both for the Pro line?dmcq2014/08/09 09:10 AM
              Why either or, why not both for the Pro line?Exophase2014/08/10 09:20 AM
  Shouldn't Gassee get his basic facts right?Mark Roulo2014/08/04 07:43 AM
    Shouldn't Gassee get his basic facts right?Anon2014/08/04 12:17 PM
      Shouldn't Gassee get his basic facts right?Kevin G2014/08/05 10:11 AM
        Shouldn't Gassee get his basic facts right?Ricardo B2014/08/05 01:30 PM
    Shouldn't Gassee get his basic facts right?Maxwell2014/08/04 06:55 PM
      Shouldn't Gassee get his basic facts right?gallier22014/08/05 12:28 AM
        Shouldn't Gassee get his basic facts right?Patrick Chase2014/08/05 07:48 AM
        Shouldn't Gassee get his basic facts right?David Hess2014/08/05 03:59 PM
      Shouldn't Gassee get his basic facts right?David Hess2014/08/05 03:56 PM
        Shouldn't Gassee get his basic facts right?rwessel2014/08/05 04:40 PM
          Shouldn't Gassee get his basic facts right?David Hess2014/08/06 10:59 AM
  A strong answer to this crap article on ExtremetechAlberto2014/08/06 02:51 AM
    A strong answer to this crap article on Extremetechdmcq2014/08/06 04:01 AM
    A strong answer to this crap article on Extremetechjuanrga2014/08/06 10:55 AM
      A strong answer to this crap article on Extremetechanon2014/08/06 06:54 PM
        A strong answer to this crap article on Extremetechanon2014/08/06 06:57 PM
        A strong answer to this crap article on ExtremetechPatrick Chase2014/08/07 08:39 AM
        A strong answer to this crap article on Extremetechjuanrga2014/08/08 09:49 AM
          A strong answer to this crap article on ExtremetechMegol2014/08/08 10:23 AM
            A strong answer to this crap article on Extremetechjuanrga2014/08/08 11:09 AM
              A strong answer to this crap article on ExtremetechKlimax2014/08/08 12:12 PM
                A strong answer to this crap article on Extremetechjuanrga2014/08/08 03:36 PM
                  A strong answer to this crap article on ExtremetechKlimax2014/08/09 12:31 AM
                    Actually, he did.iz2014/08/09 01:05 AM
                      Actually, he did.Klimax2014/08/09 01:36 AM
                        Actually, he did.Klimax2014/08/09 01:51 AM
                          Actually, he did.juanrga2014/08/09 04:26 AM
                            Actually, he did.Klimax2014/08/10 10:01 AM
                              Actually, he did.Wilco2014/08/10 11:29 AM
                                Actually, he did.Klimax2014/08/10 01:04 PM
                                  Actually, he did.Wilco2014/08/11 01:52 AM
                                    Actually, he did.Klimax2014/08/11 10:08 AM
              A strong answer to this crap article on Extremetechjoel2014/08/08 01:26 PM
                A strong answer to this crap article on ExtremetechMichael S2014/08/09 10:01 AM
                  A strong answer to this crap article on ExtremetechKlimax2014/08/09 01:10 PM
                    x87 taxMichael S2014/08/09 01:37 PM
                      x87 taxKlimax2014/08/10 07:48 AM
                        Visual C and x87Eugene Nalimov2014/08/11 03:07 PM
                          Visual C and x87rwessel2014/08/11 05:32 PM
                            Visual C and x87Eugene Nalimov2014/08/12 08:29 AM
                  A strong answer to this crap article on Extremetechjoel2014/08/11 03:21 PM
            A strong answer to this crap article on Extremetechanon2014/08/08 11:29 PM
              A strong answer to this crap article on ExtremetechMegol2014/08/09 03:22 AM
                A strong answer to this crap article on Extremetechanon2014/08/09 04:03 AM
                  A strong answer to this crap article on ExtremetechDavid Kanter2014/08/09 08:55 AM
                    Complexity = f (area, X, Y) not f(area)Paul A. Clayton2014/08/09 12:47 PM
                    x86 theoretical limit on maximum performance.Jouni Osmala2014/08/10 03:27 AM
                      x86 theoretical limit on maximum performance.dmcq2014/08/10 04:09 AM
                        x86 theoretical limit on maximum performance.Jouni Osmala2014/08/10 04:53 AM
                      x86 theoretical limit on maximum performance.juanrga2014/08/10 05:24 AM
                        x86 theoretical limit on maximum performance.Aaron Spink2014/08/10 06:24 AM
                          x86 theoretical limit on maximum performance.juanrga2014/08/11 06:05 PM
                            x86 theoretical limit on maximum performance.Aaron Spink2014/08/11 07:28 PM
                        Stop trollingDavid Kanter2014/08/10 08:12 PM
                          Stop trollingExophase2014/08/10 08:34 PM
                            Thunder chips dissipate 20 to 95 watts for SoCsjuanrga2014/08/11 05:38 PM
                              Thunder chips dissipate 20 to 95 watts for SoCsKlimax2014/08/11 09:03 PM
                          Is comparing to highest bin fair?Paul A. Clayton2014/08/11 12:37 AM
                            Is comparing to highest bin fair?David Kanter2014/08/13 11:18 AM
                              What kind of comparison was intended?Paul A. Clayton2014/08/13 03:18 PM
                          Stop ignoring datajuanrga2014/08/11 06:00 PM
                            Stop ignoring dataKlimax2014/08/11 09:06 PM
                              Seriously guys stop this anti-ARM charadejuanrga2014/08/12 01:21 AM
                                Seriously guys stop this anti-ARM charadeAaron Spink2014/08/12 01:24 AM
                                  Seriously guys stop this anti-ARM charadejuanrga2014/08/12 02:00 AM
                                    Seriously guys stop this anti-ARM charadeAaron Spink2014/08/12 02:08 AM
                                      Seriously guys stop this anti-ARM charadejuanrga2014/08/12 02:25 AM
                                Seriously guys stop this anti-ARM charadeKlimax2014/08/12 05:29 AM
                                  Seriously guys stop this anti-ARM charadejuanrga2014/08/12 10:19 AM
                                    Seriously guys stop this anti-ARM charadeKlimax2014/08/12 01:08 PM
                                Seriously guys stop this pro-ARM troll called JuanrgaJ Random Hacker2014/08/19 04:47 PM
                                  Seriously guys stop this pro-ARM troll called Juanrgaanon2014/08/20 02:46 PM
                            Design targets aren't dataDavid Kanter2014/08/13 12:07 PM
                              I agreejuanrga2014/08/13 02:16 PM
                                I agreeDavid Kanter2014/08/14 10:37 PM
                                  Not again, thanksjuanrga2014/08/15 02:22 AM
                                  I agreeRonald Maas2014/08/15 06:40 AM
                                    I agreeMichael S2014/08/15 06:55 AM
                                      I agreeDavid Kanter2014/08/15 08:43 AM
                                      I agreejuanrga2014/08/15 10:28 AM
                                    I agreeDavid Kanter2014/08/15 08:41 AM
                                      I agreejuanrga2014/08/15 10:39 AM
                                        I agreeKlimax2014/08/15 12:57 PM
                                        I agreeDavid Kanter2014/08/15 02:01 PM
                                          on ISAs and a plea for pieceMaynard Handley2014/08/15 03:41 PM
                                            on ISAs and a plea for piecejuanrga2014/08/16 02:15 AM
                                              on ISAs and a plea for pieceKlimax2014/08/16 02:56 AM
                                                on ISAs and a plea for piecejuanrga2014/08/16 03:39 AM
                                                  on ISAs and a plea for pieceKlimax2014/08/16 06:52 AM
                                                    on ISAs and a plea for piecejuanrga2014/08/16 10:21 AM
                                            on ISAs and a plea for pieceRicardo B2014/08/16 07:51 AM
                                              on ISAs and a plea for pieceMaynard Handley2014/08/16 08:45 AM
                                                on ISAs and a plea for pieceRicardo B2014/08/16 09:14 AM
                                              on ISAs and a plea for piecejuanrga2014/08/16 09:54 AM
                                                on ISAs and a plea for pieceRicardo B2014/08/16 10:32 AM
                                              on ISAs and a plea for pieceSymmetry2014/08/17 05:32 AM
                                                on ISAs and a plea for pieceRicardo B2014/08/17 06:07 AM
                                          K12Alexko2014/08/15 04:43 PM
                                            K12David Kanter2014/08/15 05:54 PM
                                              K12Alexko2014/08/16 02:14 AM
                                              K12juanrga2014/08/16 03:34 AM
                                                Sorry cannot share any detailsDavid Kanter2014/08/16 10:31 AM
                                                  K12, some guesses/leaksjuanrga2014/08/18 01:48 PM
                                          "Intel no longer has an obvious lead in process technology" [electronicsweekly]juanrga2014/08/16 01:56 AM
                                            "Intel no longer has an obvious lead in process technology" [electronicsweekly]Kira2014/08/16 08:07 AM
                                              "Intel no longer has an obvious lead in process technology" [electronicsweekly]Maynard Handley2014/08/16 08:56 AM
                                                "Intel no longer has an obvious lead in process technology" [electronicsweekly]Doug S2014/08/16 09:28 AM
                                                  "Intel no longer has an obvious lead in process technology" [electronicsweekly]juanrga2014/08/16 11:01 AM
                                                    "Intel no longer has an obvious lead in process technology" [electronicsweekly]tarlinian2014/08/16 12:56 PM
                                                      "Intel no longer has an obvious lead in process technology" [electronicsweekly]juanrga2014/08/17 01:29 PM
                                                        "Intel no longer has an obvious lead in process technology" [electronicsweekly]tarlinian2014/08/17 02:08 PM
                                                          Metal pitchesDavid Kanter2014/08/17 02:20 PM
                                                          "Intel no longer has an obvious lead in process technology" [electronicsweekly]juanrga2014/08/17 03:14 PM
                                              "Intel no longer has an obvious lead in process technology" [electronicsweekly]juanrga2014/08/16 10:49 AM
                                            Intel manufacturing advantage is still hereDavid Kanter2014/08/16 10:27 AM
                                              Intel manufacturing advantage is still herejuanrga2014/08/16 11:45 AM
                                                Intel manufacturing advantage is still hereDavid Kanter2014/08/16 01:16 PM
                                                  Intel manufacturing advantage is still herejuanrga2014/08/17 01:52 PM
                                                    History is very usefulDavid Kanter2014/08/17 02:18 PM
                                                      History is very useful, but can not be used to make infallible predictionsjuanrga2014/08/17 03:33 PM
                                                        History is very useful, but can not be used to make infallible predictionsDavid Kanter2014/08/17 05:28 PM
                                                          History is very useful, but can not be used to make infallible predictionsPurana Archer2014/08/18 02:32 AM
                                                            History is very useful, but can not be used to make infallible predictionsGabriele Svelto2014/08/18 07:17 AM
                                                              History is very useful, but can not be used to make infallible predictionsMaynard Handley2014/08/18 12:36 PM
                                                                History is very useful, but can not be used to make infallible predictionsGabriele Svelto2014/08/19 12:53 AM
                                                                History is very useful, but can not be used to make infallible predictionstarlinian2014/08/19 12:14 PM
                                                        History is very useful, but can not be used to make infallible predictionsAaron Spink2014/08/17 09:10 PM
                                                        Intel has process advantage.Jouni Osmala2014/08/18 12:36 AM
                                                          Intel has process advantage.juanrga2014/08/19 02:00 AM
                                                            Intel has process advantage.Jouni Osmala2014/08/20 03:33 AM
                                                              Intel has process advantage.dmcq2014/08/20 07:17 AM
                                                                Intel has process advantage.Michael S2014/08/20 07:58 AM
                                                                  Intel has process advantage.dmcq2014/08/20 04:11 PM
                                                                Intel has process advantage.Jouni Osmala2014/08/20 10:49 PM
                                                                  Intel has process advantage.dmcq2014/08/21 12:51 AM
                                                                    Intel has process advantage.juanrga2014/08/21 02:05 AM
                                                                  Intel process advantage is shrinkingjuanrga2014/08/21 01:15 AM
                                                                    Could you please stopDaniel B2014/08/21 03:41 AM
                                                                      Could you please stopanon2014/08/21 08:20 AM
                                                                    Intel process advantage is shrinkingjuanrga2014/08/25 11:29 AM
                                                                      Intel process advantage is shrinkingKlimax2014/08/25 01:14 PM
                                                                        Intel process advantage is shrinkingjuanrga2014/08/25 02:11 PM
                                                                          Intel process advantage is shrinkingKlimax2014/08/25 09:12 PM
                                                                            Intel process advantage is shrinkingjuanrga2014/08/26 01:16 AM
                                                                              Intel process advantage is shrinkingKlimax2014/08/26 02:38 AM
                                                                                Intel process advantage is shrinkingjuanrga2014/08/26 02:52 AM
                                                                                  Intel process advantage is shrinkingKlimax2014/08/26 02:55 AM
                                                                                    Intel process advantage is shrinkingKlimax2014/08/26 03:07 AM
                                                                                      Intel process advantage is shrinkingjuanrga2014/08/26 03:40 AM
                                                                                        Intel process advantage is shrinkingAaron Spink2014/08/26 03:50 AM
                                                                                  Intel process advantage is shrinkingDavid Kanter2014/08/26 05:25 PM
                                                                                    Intel process advantage is shrinkingjuanrga2014/08/27 03:10 AM
                                                                                      Intel process advantage is shrinkingjuanrga2014/08/27 03:15 AM
                                                                            Intel process advantage is shrinkingMaynard Handley2014/08/26 01:49 PM
                                                                              Intel process advantage is shrinkingjuanrga2014/08/26 03:44 PM
                                                                              Intel process advantage is shrinkingKlimax2014/08/26 09:37 PM
                                                                      Intel process advantage is shrinkingDoug S2014/08/25 09:24 PM
                                                                        Intel process advantage is shrinkingdmcq2014/08/25 11:17 PM
                                                                        Intel process advantage is shrinkingjuanrga2014/08/26 01:22 AM
                                                                      Are you really that daft?Daniel B2014/08/26 04:00 AM
                                                                        Are you really that daft?Ronald Maas2014/08/26 03:37 PM
                                                                          Conservative, or just realistic?Daniel B2014/08/27 12:27 AM
                                                                      Some new data to scrap overMaynard Handley2014/08/27 12:34 PM
                                                                        Some new data to scrap overAaron Spink2014/08/27 01:21 PM
                                                                          Some new data to scrap overWilco2014/08/27 02:38 PM
                                                                            Some new data to scrap overAaron Spink2014/08/27 10:22 PM
                                                                            Some new data to scrap overgallier22014/08/28 01:26 AM
                                                                          Some new data to scrap overMaynard Handley2014/08/27 02:56 PM
                                                                            Some new data to scrap overAaron Spink2014/08/27 10:14 PM
                                                                            Some new data to scrap overgallier22014/08/28 01:42 AM
                                                                              Some new data to scrap overWilco2014/08/28 04:11 AM
                                                                        Some new data to scrap overjuanrga2014/08/27 03:35 PM
                                                                          Some new data to scrap overAaron Spink2014/08/27 11:04 PM
                                                                            Some new data to scrap overjuanrga2014/08/28 05:57 AM
                                                                        Some new data to scrap overRicardo B2014/08/29 08:14 AM
                                                Will there ever be a desktop Broadwell?Doug S2014/08/16 07:06 PM
                                                  Will there ever be a desktop Broadwell?Ricardo B2014/08/17 06:01 AM
                                                  Will there ever be a desktop Broadwell?juanrga2014/08/17 01:43 PM
                                            "Intel no longer has an obvious lead in process technology" [electronicsweekly]Aaron Spink2014/08/16 08:10 PM
                                              "Intel no longer has an obvious lead in process technology" [electronicsweekly]dmcq2014/08/17 01:58 AM
                                              "Intel no longer has an obvious lead in process technology" [electronicsweekly]Maynard Handley2014/08/17 09:06 AM
                                                "Intel no longer has an obvious lead in process technology" [electronicsweekly]tarlinian2014/08/17 09:40 AM
                                                  "Intel no longer has an obvious lead in process technology" [electronicsweekly]Maynard Handley2014/08/17 11:37 AM
                                                    "Intel no longer has an obvious lead in process technology" [electronicsweekly]tarlinian2014/08/17 12:12 PM
                                                    "Intel no longer has an obvious lead in process technology" [electronicsweekly]Alberto2014/08/17 12:12 PM
                                                    "Intel no longer has an obvious lead in process technology" [electronicsweekly]Michael S2014/08/17 12:15 PM
                                                      "Intel no longer has an obvious lead in process technology" [electronicsweekly]Maynard Handley2014/08/17 12:45 PM
                                                        "Intel no longer has an obvious lead in process technology" [electronicsweekly]Michael S2014/08/17 01:13 PM
                                                        "Intel no longer has an obvious lead in process technology" [electronicsweekly]David Kanter2014/08/17 02:29 PM
                                                          "Intel no longer has an obvious lead in process technology" [electronicsweekly]Maynard Handley2014/08/17 03:10 PM
                                                            "Intel no longer has an obvious lead in process technology" [electronicsweekly]tarlinian2014/08/17 04:08 PM
                                                            "Intel no longer has an obvious lead in process technology" [electronicsweekly]Klimax2014/08/17 09:42 PM
                                                              "Intel no longer has an obvious lead in process technology" [electronicsweekly]Maynard Handley2014/08/18 12:21 PM
                                                            "Intel no longer has an obvious lead in process technology" [electronicsweekly]Alberto2014/08/18 03:19 AM
                                                            Die Size matters a LOT!Jouni Osmala2014/08/18 12:21 PM
                                        I agreeDavid Hess2014/08/17 02:24 PM
                                      I agreeRonald Maas2014/08/15 10:14 PM
                                        I agreedmcq2014/08/16 02:53 AM
                                          Beyond web serversjuanrga2014/08/16 03:55 AM
                                            Beyond web serversdmcq2014/08/16 04:40 AM
                                              Beyond web serversjuanrga2014/08/16 10:43 AM
                                                Beyond web serversAaron Spink2014/08/16 09:07 PM
                                                  Beyond web serversjuanrga2014/08/17 01:35 PM
                                                    Beyond web serversAaron Spink2014/08/17 03:24 PM
                                                      Beyond web serversdmcq2014/08/18 04:50 AM
                                                        Beyond web serversAaron Spink2014/08/18 02:58 PM
                                            Beyond web serversMichael S2014/08/16 11:43 AM
                                              Beyond web serversjuanrga2014/08/16 11:55 AM
                                                Beyond web serversMichael S2014/08/16 12:04 PM
                                                  Beyond web serversjuanrga2014/08/17 07:19 AM
                                                    Beyond web serversrwessel2014/08/17 08:34 AM
                                                    Beyond web serversDavid Kanter2014/08/17 09:23 AM
                                                      Beyond web serversMichael S2014/08/17 10:04 AM
                                                        Beyond web serversRicardo B2014/08/17 10:37 AM
                                                          Beyond web serversDavid Kanter2014/08/17 11:31 AM
                                                      Beyond web serversjuanrga2014/08/17 03:59 PM
                                                        Beyond web serversDavid Kanter2014/08/17 05:42 PM
                                                          Beyond web serversjuanrga2014/08/18 01:39 PM
                                              Beyond web serversGabriele Svelto2014/08/16 10:33 PM
                                                Beyond web serversRonald Maas2014/08/17 08:06 AM
                                                  Amdahl's law (NT)Michael S2014/08/17 10:08 AM
                                                Beyond web serversjuanrga2014/08/17 01:41 PM
                                      Xeon-class performancePaul A. Clayton2014/08/16 09:31 AM
                                        Xeon-class performancejuanrga2014/08/16 10:53 AM
                                          Xeon-class performanceKlimax2014/08/17 01:29 PM
                              Design targets aren't datadistantobserver2014/08/16 11:22 AM
                      x86 theoretical limit on maximum performance.Kevin G2014/08/10 09:26 PM
                    A strong answer to this crap article on Extremetechanon2014/08/10 04:25 PM
                      A strong answer to this crap article on ExtremetechDavid Kanter2014/08/10 08:15 PM
                        A strong answer to this crap article on ExtremetechExophase2014/08/10 09:03 PM
                          A strong answer to this crap article on ExtremetechAaron Spink2014/08/10 11:33 PM
                            A strong answer to this crap article on Extremetechnone2014/08/11 12:35 AM
                              A strong answer to this crap article on Extremetechnope2014/08/11 01:05 AM
                              A strong answer to this crap article on ExtremetechAaron Spink2014/08/11 03:02 AM
                            A strong answer to this crap article on ExtremetechRicardo B2014/08/11 03:04 AM
                  A strong answer to this crap article on ExtremetechMegol2014/08/11 07:10 AM
                    A strong answer to this crap article on Extremetechanon2014/08/11 07:44 PM
              A strong answer to this crap article on ExtremetechGabriele Svelto2014/08/09 11:37 PM
                A strong answer to this crap article on Extremetechjuanrga2014/08/10 05:07 AM
                A strong answer to this crap article on Extremetechanon2014/08/11 04:13 AM
                  A strong answer to this crap article on ExtremetechGabriele Svelto2014/08/11 11:08 AM
                    A strong answer to this crap article on Extremetechanon2014/08/11 08:32 PM
              A strong answer to this crap article on ExtremetechMichael S2014/08/10 02:11 AM
                A strong answer to this crap article on Extremetechanon2014/08/10 04:27 PM
                  A strong answer to this crap article on ExtremetechMichael S2014/08/10 10:30 PM
                    A strong answer to this crap article on Extremetechanon2014/08/11 02:56 AM
                      A strong answer to this crap article on ExtremetechMichael S2014/08/11 04:23 AM
                        A strong answer to this crap article on Extremetechanon2014/08/11 05:29 AM
                      A strong answer to this crap article on ExtremetechMegol2014/08/11 07:30 AM
                        A strong answer to this crap article on Extremetechanon2014/08/11 07:46 PM
                          A strong answer to this crap article on ExtremetechMegol2014/08/12 04:48 AM
                            A strong answer to this crap article on ExtremetechMegol2014/08/13 08:25 AM
                              A strong answer to this crap article on Extremetechanon2014/08/13 03:55 PM
          A strong answer to this crap article on ExtremetechDavid Kanter2014/08/08 10:36 PM
            A strong answer to this crap article on Extremetechanon2014/08/08 11:12 PM
              A strong answer to this crap article on Extremetechdmcq2014/08/09 02:00 AM
              A strong answer to this crap article on ExtremetechAaron Spink2014/08/09 02:54 AM
                A strong answer to this crap article on Extremetechanon2014/08/10 08:50 PM
                  A strong answer to this crap article on ExtremetechExophase2014/08/10 09:35 PM
                    A strong answer to this crap article on Extremetechanon2014/08/11 03:20 AM
                  A strong answer to this crap article on ExtremetechAaron Spink2014/08/10 11:48 PM
                    A strong answer to this crap article on Extremetechanon2014/08/11 03:10 AM
                      A strong answer to this crap article on ExtremetechMichael S2014/08/11 04:43 AM
                        A strong answer to this crap article on Extremetechanon2014/08/11 05:21 AM
            A strong answer to this crap article on Extremetechjuanrga2014/08/09 05:38 AM
              ISA doesn't matterDavid Kanter2014/08/09 09:50 AM
                ISA doesn't matterBrett2014/08/09 10:51 AM
                  ISA doesn't matterLinus Torvalds2014/08/09 12:36 PM
                    ISA doesn't matterBrett2014/08/09 09:37 PM
                      ISA doesn't matterMichael S2014/08/10 01:21 AM
                        ISA doesn't matterdmcq2014/08/10 03:06 AM
                        ISA doesn't matterBrett2014/08/10 01:21 PM
                          ISA doesn't matteranon2014/08/10 08:03 PM
                            ISA doesn't matterBrett2014/08/10 10:45 PM
                              ISA doesn't matterAaron Spink2014/08/10 11:57 PM
                              ISA doesn't matteranon2014/08/11 03:18 AM
                                ISA doesn't matterBrett2014/08/16 02:38 PM
                              ISA doesn't matter -- PARMAook2014/08/26 10:50 AM
                          IBM fab business is losing money, the POWER line is notMark Roulo2014/08/11 06:44 AM
                    ISA doesn't matterTREZA2014/08/10 03:27 AM
                    ISA doesn't matterjuanrga2014/08/10 04:34 AM
                      ISA doesn't matterMichael S2014/08/10 05:01 AM
                        ISA does matterjuanrga2014/08/11 06:28 PM
                          ISA does matterMichael S2014/08/12 12:20 AM
                            ISA does matterjuanrga2014/08/12 01:38 AM
                              ISA does matterAaron Spink2014/08/12 02:06 AM
                                ISA does matterjuanrga2014/08/12 02:27 AM
                                  ISA does matterUngo2014/08/12 02:28 PM
                                    ISA does matterjuanrga2014/08/12 02:43 PM
                                      Banning of Juanrgagallier22014/08/13 07:56 AM
                                        bad idea (NT)Michael S2014/08/13 08:02 AM
                      ISA doesn't matterAaron Spink2014/08/10 06:48 AM
                      Best *Student* Paper Paul A. Clayton2014/08/10 11:31 AM
                ISA doesn't matterjuanrga2014/08/10 03:32 AM
                  ISA doesn't matterWilco2014/08/10 04:48 AM
                    ISA doesn't matterLinus Torvalds2014/08/10 10:38 AM
                      ISA doesn't matterUngo2014/08/11 02:59 PM
                ISA doesn't matterEtienne2014/08/11 06:17 AM
                  ISA doesn't matterGabriele Svelto2014/08/12 10:31 PM
                    ISA doesn't matterEtienne2014/08/13 02:09 AM
          A strong answer to this crap article on ExtremetechAaron Spink2014/08/09 02:44 AM
            A strong answer to this crap article on Extremetechjuanrga2014/08/09 05:43 AM
              A strong answer to this crap article on Extremetechanon2014/08/09 09:17 AM
                A strong answer to this crap article on ExtremetechMichael S2014/08/09 11:38 AM
                  A strong answer to this crap article on ExtremetechAlberto2014/08/09 12:41 PM
                  A strong answer to this crap article on ExtremetechWilco2014/08/09 12:49 PM
                    A strong answer to this crap article on ExtremetechAlberto2014/08/09 01:13 PM
                      A strong answer to this crap article on ExtremetechWilco2014/08/09 01:30 PM
                        A strong answer to this crap article on ExtremetechAlberto2014/08/09 02:15 PM
                        A strong answer to this crap article on Extremetechcarop2014/08/10 01:01 AM
                          A strong answer to this crap article on ExtremetechWilco2014/08/10 03:50 AM
                            A strong answer to this crap article on ExtremetechMichael S2014/08/10 05:23 AM
                              A strong answer to this crap article on ExtremetechWilco2014/08/10 05:38 AM
                                Thank you (NT)Michael S2014/08/10 06:16 AM
                                Nope, Samsung ARM cpu was canned.Alberto2014/08/11 01:57 AM
                                  Nope, Samsung ARM cpu was canned.Wilco2014/08/11 02:14 AM
                                  Nope, Samsung ARM cpu was canned.David Hess2014/08/11 10:56 AM
                                    Nope, Samsung ARM cpu was canned.juanrga2014/08/11 06:32 PM
                              ISA does matterjuanrga2014/08/11 06:34 PM
                          A strong answer to this crap article on Extremetechjuanrga2014/08/10 05:19 AM
                      A strong answer to this crap article on ExtremetechWilco2014/08/09 02:10 PM
                        A strong answer to this crap article on ExtremetechMichael S2014/08/10 02:01 AM
                        A strong answer to this crap article on ExtremetechAlberto2014/08/10 05:17 AM
                A strong answer to this crap article on Extremetechjuanrga2014/08/10 02:40 AM
                  A strong answer to this crap article on ExtremetechMichael S2014/08/10 03:27 AM
                    A strong answer to this crap article on ExtremetechWilco2014/08/10 04:00 AM
                      A strong answer to this crap article on ExtremetechMichael S2014/08/10 04:50 AM
                        A strong answer to this crap article on ExtremetechWilco2014/08/10 05:51 AM
                      A strong answer to this crap article on ExtremetechAaron Spink2014/08/10 07:04 AM
                        A strong answer to this crap article on ExtremetechWilco2014/08/10 07:30 AM
                          A strong answer to this crap article on ExtremetechAaron Spink2014/08/10 07:46 AM
                            A strong answer to this crap article on ExtremetechWilco2014/08/11 04:37 PM
                              A strong answer to this crap article on ExtremetechMichael S2014/08/12 12:58 AM
                                A strong answer to this crap article on ExtremetechAaron Spink2014/08/12 01:11 AM
                                  A strong answer to this crap article on ExtremetechMichael S2014/08/12 01:22 AM
                                    A strong answer to this crap article on ExtremetechAaron Spink2014/08/12 01:40 AM
                                Applied Micro systems not benchmarkedDavid Kanter2014/08/13 10:29 AM
                                  Applied Micro systems not benchmarkedjuanrga2014/08/13 10:51 AM
                          A strong answer to this crap article on ExtremetechMichael S2014/08/10 07:48 AM
                            A strong answer to this crap article on ExtremetechAaron Spink2014/08/10 08:19 AM
                    A strong answer to this crap article on Extremetechjuanrga2014/08/10 05:00 AM
                      A strong answer to this crap article on ExtremetechMichael S2014/08/10 05:18 AM
                        A strong answer to this crap article on ExtremetechJouni Osmala2014/08/10 01:12 PM
                          A strong answer to this crap article on Extremetechdmcq2014/08/10 01:43 PM
          A strong answer to this crap article on ExtremetechRonald Maas2014/08/10 09:56 PM
            A strong answer to this crap article on ExtremetechpTmd2014/08/10 11:57 PM
            A strong answer to this crap article on ExtremetechAaron Spink2014/08/11 12:02 AM
              ISA importance and bike sheddingPaul A. Clayton2014/08/11 01:39 AM
              A strong answer to this crap article on ExtremetechJouni Osmala2014/08/11 02:43 AM
                A strong answer to this crap article on ExtremetechMichael S2014/08/11 04:56 AM
                  A strong answer to this crap article on ExtremetechJouni Osmala2014/08/11 11:14 AM
            ISA does matterjuanrga2014/08/11 06:48 PM
              ISA does matterUngo2014/08/12 03:37 PM
                ISA does matterjuanrga2014/08/13 04:18 AM
            Look at the system level, ISA doesn't matterDavid Kanter2014/08/13 10:47 AM
              POWER guys commented on this years ago ...Mark Roulo2014/08/14 07:30 AM
                POWER guys commented on this years ago ...juanrga2014/08/16 03:29 AM
                  POWER guys commented on this years ago ...Maynard Handley2014/08/16 08:35 AM
                    POWER guys commented on this years ago ...juanrga2014/08/16 10:35 AM
                    POWER guys commented on this years ago ...Michael S2014/08/16 11:55 AM
                      POWER guys commented on this years ago ...Maynard Handley2014/08/16 01:24 PM
                        POWER guys commented on this years ago ...Michael S2014/08/16 01:54 PM
                          POWER guys commented on this years ago ...Maynard Handley2014/08/16 02:56 PM
                        POWER guys commented on this years ago ...Michael S2014/08/16 02:02 PM
                        POWER guys commented on this years ago ...David Kanter2014/08/16 03:09 PM
                        POWER guys commented on this years ago ...Ricardo B2014/08/16 04:43 PM
                          POWER guys commented on this years ago ...Maynard Handley2014/08/16 05:35 PM
                            POWER guys commented on this years ago ...Ricardo B2014/08/16 06:17 PM
                              OS/2Ricardo B2014/08/16 06:20 PM
                              POWER guys commented on this years ago ...rwessel2014/08/16 08:36 PM
                              POWER guys commented on this years ago ...Maynard Handley2014/08/17 11:24 AM
                                POWER guys commented on this years ago ...Ricardo B2014/08/17 12:41 PM
                                POWER guys commented on this years ago ...Gabriele Svelto2014/08/18 01:20 AM
                          POWER guys commented on this years ago ...rwessel2014/08/16 08:43 PM
                          POWER guys commented on this years ago ...Michael S2014/08/16 10:30 PM
                            POWER guys commented on this years ago ...Ricardo B2014/08/17 05:28 AM
                        POWER guys commented on this years ago ...Aaron Spink2014/08/16 09:25 PM
                          POWER guys commented on this years ago ...anon2014/08/16 11:10 PM
                            POWER guys commented on this years ago ...Aaron Spink2014/08/16 11:42 PM
                              POWER guys commented on this years ago ...anon2014/08/17 02:40 AM
                                POWER guys commented on this years ago ...Ricardo B2014/08/17 05:43 AM
                                  POWER guys commented on this years ago ...anon2014/08/17 06:53 AM
                                POWER guys commented on this years ago ...anon2014/08/17 06:42 AM
                              POWER guys commented on this years ago ...Ronald Maas2014/08/17 08:57 AM
                                POWER guys commented on this years ago ...rwessel2014/08/17 09:40 AM
                                POWER guys commented on this years ago ...Ricardo B2014/08/17 11:06 AM
                                  POWER guys commented on this years ago ...Aaron Spink2014/08/17 12:56 PM
                                  POWER guys commented on this years ago ...Michael S2014/08/17 01:22 PM
                                    POWER guys commented on this years ago ...Ricardo B2014/08/17 02:14 PM
                                      cheap membarsMichael S2014/08/17 04:56 PM
                                        cheap membarsnksingh2014/08/21 10:34 AM
                                          cheap membarsMichael S2014/08/21 11:31 AM
                                            cheap membarsnksingh2014/08/21 01:54 PM
                                              cheap membarsMichael S2014/08/21 05:40 PM
                                            cheap membarsanon2014/08/21 10:17 PM
                                              cheap membarsMichael S2014/08/22 02:16 AM
                                                cheap membarsanon2014/08/22 06:33 AM
                                                  cheap membarsMichael S2014/08/22 07:18 AM
                                                    cheap membarsanon2014/08/22 04:50 PM
                                                      cheap membarsPatrick Chase2014/08/24 10:06 AM
                                                        cheap membarsPatrick Chase2014/08/24 10:11 AM
                                                          cheap membarsanon2014/08/24 06:21 PM
                                                            cheap membarsPatrick Chase2014/08/25 10:02 AM
                                                              cheap membarsdmcq2014/08/30 01:15 AM
                                      POWER guys commented on this years ago ...anon2014/08/17 09:55 PM
                              POWER guys commented on this years ago ...Howard Chu2014/08/28 08:17 PM
                                POWER guys commented on this years ago ...Aaron Spink2014/08/29 06:35 AM
                                  POWER guys commented on this years ago ...Michael S2014/08/29 07:34 AM
                                  POWER guys commented on this years ago ...anon2014/08/29 02:49 PM
                                    POWER guys commented on this years ago ...foobar2014/08/30 08:26 PM
                                      POWER guys commented on this years ago ...dmcq2014/08/31 03:54 AM
                                        POWER guys commented on this years ago ...Wilco2014/08/31 11:52 AM
                          POWER guys commented on this years ago ...Maynard Handley2014/08/17 08:52 AM
                            POWER guys commented on this years ago ...Maynard Handley2014/08/17 09:31 AM
                              POWER guys commented on this years ago ...Gabriele Svelto2014/08/18 01:11 AM
                            POWER guys commented on this years ago ...nksingh2014/08/25 02:55 AM
                            POWER guys commented on this years ago ...Patrick Chase2014/08/25 10:15 AM
                              POWER guys commented on this years ago ...Patrick Chase2014/08/25 10:23 AM
                              POWER guys commented on this years ago ...nksingh2014/08/25 08:26 PM
                                POWER guys commented on this years ago ...Patrick Chase2014/08/25 10:32 PM
                                  POWER guys commented on this years ago ...Aaron Spink2014/08/26 03:41 AM
                                    POWER guys commented on this years ago ...anon2014/08/26 05:06 AM
                                      POWER guys commented on this years ago ...Aaron Spink2014/08/26 04:27 PM
                                        POWER guys commented on this years ago ...anon2014/08/26 10:02 PM
                                          POWER guys commented on this years ago ...Aaron Spink2014/08/26 10:51 PM
                                            POWER guys commented on this years ago ...anon2014/08/28 07:25 PM
                                              POWER guys commented on this years ago ...anon2014/08/28 09:45 PM
                                              POWER guys commented on this years ago ...rwessel2014/08/29 12:14 AM
                                                POWER guys commented on this years ago ...anon2014/08/29 02:49 AM
                                                  POWER guys commented on this years ago ...Michael S2014/08/29 05:43 AM
                                          POWER guys commented on this years ago ...nksingh2014/08/26 10:54 PM
                                        POWER guys commented on this years ago ...dmcq2014/08/27 01:29 AM
                                          POWER guys commented on this years ago ...dmcq2014/08/27 07:16 AM
                                    Idiom recognition for LL/SC?Paul A. Clayton2014/08/26 12:28 PM
                        POWER guys commented on this years ago ...Gabriele Svelto2014/08/16 10:22 PM
                      AltiVec was Power6 (NT)Kira2014/08/17 10:33 AM
                        AltiVec was Power6Maynard Handley2014/08/17 11:29 AM
                          AltiVec was Power6Kira2014/08/17 11:53 AM
Reply to this Topic
Name:
Email:
Topic:
Body: No Text
How do you spell avocado?