Jukka Larja (roskakori2006@gmail.com) on 10/1/09 wrote:
---------------------------
>AM (myname4rwt@jee.male) on 9/30/09 wrote:
>---------------------------
>
>> Your example has very little to do with
>> the subject of the discussion you joined.
>
>Yes, and I'm pretty sure I stated that before. I certainly stated it in my last
>message. Do you even read these messages through before you start answering? Or
>do you just prefer to tackle them from beginning to the end in such a small pieces
>that there's bound to be misunderstandings.
>
>The whole three statements thing came from the fact that I thought interpreted
>what you were writing (not to Jouni, but to me) to mean that there is no cost to
>add OoOE to a core[1].

I seriously doubt I said that. Are you sure? I recall I mentioned on several occasions that I don't dispute power and die overheads (have you seen those?).

If this is the whole reason why you keep posting to "correct my misunderstanding", I'd like to see the post+sentence you're talking about.

For me, it was pretty clear, and still is, that there is
>a cost, and OoOE has to bring some benefits to make sense. Whether that cost is
>meaningfull compared to cost X or whether the benefits on workload Y are ten times
>greater than the cost is another discussion.
>
>However, as long as there is a cost, it reflects to clockspeed, if that is the
>only variable you can change. In practise it may not matter, because there are other
>variables and even if it did, absolute clockspeed isn't important anyway.
>
>It doesn't really matter to me whether the cost is in power or somewhere else.
>I thought power was a clear example and chose it.

You chose something that was never a point of disagreement in the prior discussion. And yet you find that someone, and not you!, doesn't get any of it? Superb.

>[1] Not taking into account the cost of more competent engineers or longer design
>cycle. I think those aren't the problem here.
>
>> If you dispute anything of what I said,
>> please point it out and as clearly and
>> concisely as possible so we can
>> concentrate on that w/o wasting time.
>
>I think you stated (or implied or whatever) that implementing OoO has no cost.
>Dont's ask me to show where, that didn't work then and probably won't work now.

No, I don't believe I stated OoO is free. On the contrary, you should be able to find me saying that OoOE needs to be justified, or doesn't contribute to higher performance on certain class of workloads regardless of how well-designed and advanced it is.

>Most likely I just misunderstood you. If you want to further the discussion instead
>of finding someone to blame, just answer as well as possible.

I hope the above answer helps (except links to the posts), so if this is something you were missing or misunderstanding at the outset, what's the meaningful point of your "trying to point out the misunderstanding" posts, let alone any further discussion in the same key?

>> Why change the subject?
>
>Not really sure at this point what subject you are referring to. I think that you
>have been changing the subject the whole time since I got into this discussion,
>or forced me to change it, but I gather you don't agree on that. Which to me means
>that we have a communication problem and for you apparently that I'm some sort of a troll.
>
>> It's not quite clear what you're trying to
>> say here. Do you disagree with anything
>> what I said?
>
>If I interpret what you have said in a suitable context then no, I don't disagree.
>But there's nothing special about that. Language is inaccurate tool. I think I have
>stated before that your facts aren't the problem, just the meaning you assign to
>them (or more accurately, the meaning you assign to facts of other people).

What are those?

>> Yes, so what? Please, state clearly what
>> you're trying to support/dispute using
>> energy argument, because from what you
>> say, the connection with your accusation
>> of me "not getting any of the
>> discussion" is far from obvious.
>
>There is no direct connection. Its about theoretical cost of OoO which came up later.

OK, this was addressed above.

>>> By the way, this isn't (as far as I can see)
>>> not what Jouni was after and not the reason
>>> I got into this discussion. I brought these
>>> up because it seemed you can't see OoOE to
>>> have any negative impact to clock speed, no
>>> matter how theoretical or how unrealistic
>>> (or rather irrelevant) in practise.
>
>I would have appreciated a lot if you would have taken this into account before writing your answer thus far.
>
>> Once again: no, OoOE is not *bound* to have
>> any impact on clock rate. Your line of
>> thinking -- if I understand correctly --
>> that *anything* that increases power
>> consumption must have a negative impact on
>> clock rate,
>
>I think the bound and must may be a problem here. I'm just trying to say that if
>there is a cost and you choose not to pay it in anywhere else, then you have to
>pay it in clockspeed.

No, no, this is a very... well, primitive look. As primitive as Jouni arguing that since 603ev clocked higher than original 603, then it must have had a longer pipeline.

I'm not trying to argue that clockspeed is the only, or best
>or even sensible place to pay the cost.
>
>I think Jouni's argument is better one since he at least is talking about propagation
>distances and added stages. I'm a lot farther from practical issues.

Thank you for the better-argument part, for I believe the matter of propagation delays across larger cores was in fact accentuated by me (along with large-cache example which makes it simple to understand).

>> is ridiculously far from theory, from practice,
>
>Well, with "bound" and "must" it surely is, but I never tried to argue that.
>
>> and from what was being discussed in the thread
>> you jumped in in the first place for quite
>> clear reasons I gave you clues to. So, do you
>> understand why or do you want a more detailed
>> explanation?
>
>Well, I didn't even mean to get to technical argument and I certainly don't see
>why I would even try to support Jouni's point if you keep interpreting it on your terms.

I'm talking about your own argument.

>> And why do you think clock rate is the
>> only free variable to adjust, to begin
>> with?
>
>Because I chose it to be. This was my example and my three statements remember?
>I didn't expect to get them clear enough on the first time I wrote them. I thought
>we could discuss them to clear out what I was meaning and see whether we actually
>have an argument over facts or just a misunderstanding of the issue.

OK, since I answered your question re. OoOE costs above, I gather it's up to you now to conclude whether misunderstanding took place or not.

>> I refer to the discussion that was taking
>> place. If you're really disputing something
>> from that, you shouldn't impose your own
>> context in the first place.
>
>(I'm not quite sure how well I understand the finer meaning of word impose, but
>anyway...) It didn't seem like you were considering other possible context yourself.
>Neither did it seem like you and Jouni were discussing in the same context or that
>you were going to stop anytime soon. I also think that you weren't addressing the
>really relevant aspects of what Jouni was saying, but mostly just getting stuck with less relevant mistakes.
>
>But this is my just how I read the discussion. I don't think there's any absolute truth to be found.

I think that if you're really trying to be helpful, you should instead refer to posts and point out his "really relevant" aspects etc. That would be more useful.

>> Then I have a question: does your "everething
>> else being equal" include die space or not,
>> and why?
>
>As we are talking about _adding_ OoO to a chip, no, it doesn't. Unless there is
>something that you clearly remove from an in-order design when adding OoO, then we need to talk about it.
>
>(By the way, I have a horrible feeling the some day someone is going to take these
>words out of context and I'm going to look like a total idiot :) .)

You don't need to wait until someday, mate. I'm asking you today: what's the logic behind fixing the power envelope ("not a nanowatt more", IIRC), but allowing the die space?

>> Excuse me, but why does the "wrong" or
>> "different" context apply to one and not
>> the other side? Who are you to decide,
>> let me ask, even if that were true?
>
>Well, different obviously applys to both sides (that's by definition). I used the
>word wrong for the same reason I chose to reply to you. I think Jouni tried to make
>his context known and you were ignoring it, so that made you a logical candidate
>for my answer. Wrong is probably a bad word because it has emotional baggage, but
>it's not like I was going to step in if I thought you were right, right?

Then I'd like to see what exactly you mean as requested above.

>> Let's see. You jump in a thread w/o caring to
>> read it through, as it turns out,
>
>I've said that I don't care to read the thread again. I have read pretty much every
>message posted to this forum since I started visiting it (I have skipped parts of
>political discussions and some messages which I couldn't comprehend in grammatical sense).
>
>Reading the whole thread, or even just the subthread from Jouni's first post, again
>would take me hours. It isn't something I'm going to do just to please you.

Not to please me. If you read the thread indeed, how can it be unknown to you that I never disputed power and die costs etc.?

>> esp. considering you're changing the
>> original context?
>
>From my point of view I was forced to seek a more simple example because you didn't
>seem to understand me. After that point we were far from what you were talking with
>Jouni about. I don't think that change of subject was in any ways deceitful or me
>trying to prove Jouni's arguments with some obscure examples.

Deceitful or not, it's of no help to your accusations of your claims of "not getting the discussion".

And it was brought up when I requested some clarity on the obvious contradiction in your own text (smth. about the waste in OoOE). Remember?

>> Well, are you sure you're the one who
>> understands properly everything?
>
>No. I have stated that. If I was, we could have perhaps not a meaningful discussion
>but at least a discussion on equal footing. As long as you think you're absolutely
>right and I'm willing to admit that I'm may not be, we're just stuck dancing aroung each other.

My questions to Jouni (and yourself) when I couldn't really decypher what was said, and complete lack of such questions on his part (or yours, for that matter, when you jumped in claiming someone doesn't understand instead of clarifying) should give a way more meaningful indication as to who tends to think he's right, then whatever one can say in his defence.

>> If that's really the case, then why do
>> you start with insults rather than
>> simple pointing out of
>> misunderstandings?
>
>I didn't mean to insult. I meant to make a strong, perhaps somewhat exaggerated
>statement and forgot all my good advice about communication especially between people
>of different languages and cultures. I didn't mean to be totally polite either,
>but what I wrote most likely was stronger than I had in mind.
>
>Apologies for that.

IIRC, your statement was something about theoretical standpoint. As such, it's an empty point as long as there's no theoretical substantiation -- and I haven't seen one. Perhaps I shouldn't have even expected to see anything from a person who seriously suggests that higher clockrate in 603ev must be coming from a longer pipeline or that someone else's TTA/systolic array-type arch won't work.

>>> By the way, you are talking about
>>> Hans de Vries, right?
>>
>> Yes. Don't recall any other Hans
>> posting here.
>
>Neither do I, but I'm not quite arrogant enough to think my memory's perfect. And
>it was easier to ask than to go searching not to find :) .
>
>> This is not a matter of thinking/context,
>> just a clear example of how he developed
>> his argument.
>
>I think Jouni was mostly reacting to what you said, whether it was particularly
>relevant to his argument or not. I don't think we can get anywhere analyzing what
>he said. He and I share cultural and lingual background, so I don't think I'm going
>to take your interpretations over mine just because.

No need to analyze or interpret, just take it for what it is.

>(For example, you didn't seem to understand Jouni's reference to debating culture.
>Now I don't know whether he was right to make that reference or not, but I do understand
>why he made it. I'm also not in any way surprised if you don't. It's one of the
>issues that often causes problems in communication between Finns and Europians and North Americans.)

What ref. is it?

>> It didnt't lead anywhere because with that
>> you're just changing the problem. Why
>> would you do it?
>
>You assume I changed the problem when I posted a different example, but fail to
>see that the misunderstanding had already happened at that point. I didn't change
>the problem, I reformulated it in simpler form. If that simpler form shows that
>we agree, then there's no need to discuss the original, more complicated problem.
>We can simply conclude that I misunderstood you.
>
>And just to be clear, the problem discussed here still isn't the one you and Jouni were discussing.

So, are you just exercising in debating?

>> The way it goes from your side, there's
>> very little point really. If you set out
>> to dispute something of what was said,
>> you should begin with pointing it out.
>
>I'm not sure whether dispute is a right word, but what I wanted to dispute was
>your interpretation of the point Jouni was trying to make.

Then -- in case you're really trying to help communication -- I look forward to seeing above-requested examples.

>> The "three basic truths" hardly count as
>> theory, this is just a specific context
>> (same power budget) you have chosen to
>> show how someone else was "wrong" or
>> "misunderstanding".
>
>Well, I was trying to show that I had misunderstood. Of course, before that you
>probably misunderstood something I wrote, but that may simply be because I had formulated
>it badly, assumed a context that wasn't as clear as I thought and so on.
>
>The thing is, I expected "the three basic truths" discussion to quickly get to
>a point where I write that I just misunderstood what you meant, no need to argue this anymore.

Well, if you feel you misunderstood me, what's the point in continuing this discussion?

In case you want to see my posts where I said this or that, but don't want to spend your time searching them, I can spend *my* time if it will be enough. I just suspect that even this will not be enough and will likely result in yet another post from you how I seem to not respect context, not listen to what others say etc.

>> Rudeness is one thing, another thing is you
>> say that someone doesn't get something
>> when apparently you didn't even care to
>> read the discussion in full to understand
>> where it comes from.
>
>I don't know where you got the idea I haven't read the whole discussion, but it is wrong.
>
>> It's simply surprizing and very
>> frustrating to see such cheap shots
>> from someone with self-proclaimed
>> academic background and admiration
>> of theory.
>
>I'm writing with my real name shown. If you want to confirm my academic background, you can do it.
>
>And since you like to flaunt my admiration of theory around I should probably add
>that I admire practise as well. (Actually, there are very few things I don't admire.)
>You can add that to your oh-I-just-mentioned-it-because-you-brought-it-up list of characterizations of me.

Done.

>> That's exactly why he got from me a fair and
>> well-deserved share of examples. Starting to
>> get it now?
>
>Well, no. I still just see the same argument that you know what he meant, and my
>interpretation (and as far as I can see, Jouni's own) are just changing the argument.
>Like this was some sort of a contest where your job is to show that we are trying to cheat.

I gave you already a short list of his fantastic arguments, if there's anything I forgot to mention, it's in the thread.

>> Actually, I never tried to switch the
>> context.
>
>I can believe that. Can you believe that neither was I? The context got screwed
>up. It happens all the time. Trying to assign a blame is pointless.

I think you admitted that the context you introduced differed significantly. No?

>>>> Yes, and one doesn't have to look far: in
>>>> this very thread you can find numerous
>>>> examples where communication doesn't seem
>>>> to work or words like "first" are
>>>> mistakenly interpreted as "last". :)
>>>
>>> Should I interpret this to mean that you
>>> don't think understanding can be an issue
>>> here :) .
>>
>> I suggest you should interpret it as the
>> smile suggests -- I started putting them
>> more often to help you get the meaning
>> right. To put it even more clearly for
>> you, attempts at defending one's argument
>> with made-up facts (Jouni) or theory
>> (you) is hardly a matter of
>> misunderstanding. Word "trolling" comes
>> to mind.
>
>Making something up is different from making a mistake (for example, remembering
>incorrectly). My point here was that what Jouni said about Prescott wasn't that
>relevant to his point (thus the mistake isn't that relevant), and you only think
>it is because you two were argueing in different context.

Prescott is just a particular (but telling) example. The context -- no, it's you who thinks of different context etc.

>> No, what you're actually doing is trying
>> to switch the context of original
>> discussion.
>
>You don't need to repeat yourself. It just wears me down trying to decide where
>to write my answers and trying to keep up whether I have addressed all relevant issues.

I don't try to repeat myself, it comes naturally.

>> No, I'm talking about the discussion you
>> jumped in w/o caring to read it. It was
>> me asking Jouni to clarify what he said.
>
>I think Jouni was doing pretty good job in general. He just shouldn't have addressed the details that weren't relevant.
>
>But what do I know. Perhaps I interpreted Jouni all wrong.

Or just defended your friend not caring much about the subject and essence of the discussion?

>>>>> What I have done is spent hours talking to
>>>>> students and working out wordings to make
>>>>> exercises (both in Finnish and English) as
>>>>> unambiguous as ever. Only to see someone
>>>>> completely ignore the context,
>>>>> misunderstand the point and come yelling
>>>>> to us that we are idiots.
>>>>
>>>> Well, it appears it only shows that careful
>>>> working out the wordings was a waste. :)
>>>> Someone ignored the context anyway, while
>>>> others probably wouldn't need the extra
>>>> work you put in to understand things
>>>> correctly (or just ask if in doubt).
>>>
>>> Funny, but your making assumptions out of
>>> context again :) (context I didn't give,
>>> so no wonder. People usually have very
>>> weird ideas about teaching. Often very
>>> idealistic and simplyfied, ignoring the
>>> practise. Many ideas that are
>>> theoretically speaking sound, simply don't
>>> work /selfirony).
>>
>> I didn't assume anything for what was said,
>> certainly not #students as you suggest
>> below. Just made a direct observation that
>> there were people for whom extra work was
>> unneeded and could have well been
>> unnecessary for others.
>
>Yeah, but that observation makes no sense for me in my context whatsoever. Given
>large enough body of students, it is statisticly speaking certain that someone won't
>understand no matter how hard you try to explain, and that someone will understand
>(or at least guess correctly), no matter how much you muddy the assignment.
>
>You are stating the obvious, after you have made an incorrect statement about waste.
>I took that to mean you used that latter statement to justify the first. That you
>thought the hours used to clear up the assignment were wasted. That only makes sense
>if you think number of students was small and thus the one complainer was significant
>part of them and thus somehow showed that we had been wasting time improving the excercise.
>
>Sure, there was a smiley of yours, but so there was a smiley of mine. Just to balance
>things out. If you were just stretching the truth for humorous tone, just read my answer in a same way.
>
>> Yet another example of how you prefer to
>> blame others for "out-of-context
>> assumptions" when it's not even true.
>
>I did say that you couldn't have known the context, because I didn't state it,
>I had a smiley there and even a rant marked with /selfirony. I don't see what more
>I could do to state that I don't blame you.
>
>Simply put, you were totally wrong and I assumed it was because I didn't specify
>very important part of the context, the number of students that is. Of course, there are other possible reasons.

Like the whole example was meant to yield a response and then show how easily I can misunderstand what was "really" meant? Ah, geddit now.

>-JLarja