Things we need to take into consideration when treading new grounds


xiongxioi

Member
Joined
Feb 13, 2012
Messages
434
All x86's supporters here are basing their statement on the assumption that a x86 pandora 2 will be as easy to produce as an arm pandora 2.

Technically speaking, they aren't very different. Since the processes are similar.

However, this is starting a brand new project rather than coming up with a successor to an already well explored concept.

We need to do feasibility studies, and use a bit of project management skill to initialize the project (including scope, cost, time, quality, quantity, hr, communication management plans).

This stage will not cost much, but it will become doublely useful and save a lot of future troubles if conducted. (that's why I think engineers are more capable than any other professionals)

I know nothing about programming (besides writing macro in excel and doing a few matlab plots), however I do know a bit about project management, calculating structural member strength and reading building standards.

My proposal is to put our efforts into more productive thinkings.

The first step to initialize the concept, is: figuring out the bare minimium we will get from a x86 pandora and how much effort we need to get to that point.

State your assumptions clearly, and do a few easy calculations.

Things I want to know are: what's would be standard spec for x86 tablets in next year. This is to assume we are just a regular bunch of people trying to make a regular bunch of x86 tablets. If we are running windows on it, how good would it support regular uses, such as internet browsing, video streaming, downloading things, document processing, and so on. If we are to run linux on it, then what kind of software would be readily avalible without doing too much optimization and tweaking. If we are to normally use it, how long will the battery last, under regular manners (make a reasonable assumption, such as: 80% of the time, it will running at 500mhz, 10% of the time, it will what what, so on and so forth). Please base these on current product and make reasonable predictions rather than blindly believe what intel and amd have said. (we can compare this to past similar products and figure out a reasonable percentage of change)

Then, i would like to know, what's the killer app on that, how much effort we would need to make the most out of it, do we need a new ui that cost half a million to develop, how much extra money and effort do we need to spend on those features and so on.

After having a basic idea about what a bare minimium x86 p2 would be like, then we will work from there.

Edit; Since i'm not an expert on the subject matter, i would like to ask you people start to brainstorm most of the functionalities you would expect on a $600 x86 computer. Video playback and such are all included. Then have a look at what similar products are capable of when those were sent into market, then come up with a reasonable assumption on the bare minimium we can deliver without much extra effort and added costs.

Assessment criterias:

1. future proof. (thanks grench for pointing this out)
 
Last edited by a moderator:
Great thinking. I think the biggest thing to look at is battery life of current gen x86 tablets. I dont think any of them can last 10 hours but the ones that can last a long time (Such as the standard 7hrs) what is the appoximate milliamp hrs to actual battery life
 
Great thinking. I think the biggest thing to look at is battery life of current gen x86 tablets. I dont think any of them can last 10 hours but the ones that can last a long time (Such as the standard 7hrs) what is the appoximate milliamp hrs to actual battery life
Like many other things, it's all in how you measure it.

The Z3770 has a nifty trick where it can scale the backlight down and lighten the image on the display so that it looks just like it otherwise would - thus saving overall device power.  Since other devices don't have this feature, do you disable it for the test?

Do you test by pegging the CPU/GPU on all of the devices and see which one dies first?  Device A may have only done X work where device B did X*5 work, but device B died first - which one had the best battery life?

Streaming video run-down test?  That winds up being a measure of how well the video decode system was built or implemented in the software being used.

Do you allow one system to be tweaked and 'set up' for the best possible scenario then compare it to the other one 'out of the box'?

We got into this trouble in another thread.  I presented a claim of one brand new SoC being faster than another 5+ year old SoC.  I was told to 'prove it'.  I provided the closest match benchmark I could find that had been run in common to the two systems - and spent the next ~3 pages of that email chain reading how the benchmark, methodology and it's application were at fault.

No matter what benchmark you choose, no matter how carefully you apply it, there will be conflict.
 
We got into this trouble in another thread.  I presented a claim of one brand new SoC being faster than another 5+ year old SoC.  I was told to 'prove it'.  I provided the closest match benchmark I could find that had been run in common to the two systems - and spent the next ~3 pages of that email chain reading how the benchmark, methodology and it's application were at fault.
No, they said "prove it is as fast as you say it is". There wasn't one person in that topic who believed that the x86 SoC was not faster than the OMAP 3530. You said it was 20x faster and people were sceptical about that.

-God Ginrai
 
Oh boy!  Not again!  This gonna grow into a monster like the other one! :) :lol:   Well at least the same people are on this thread too! :) :eek:
 
Last edited by a moderator:
Like many other things, it's all in how you measure it.
We are going from the basics first. We need raw specs as if the product is produced by a Chinese slave factory.

No capital is budgeted for optimization and all that.

After the initialization stage, scopes will be defined: how good the end product is going to be, and how much time, labor, money are going to nail it.
 
OMG, another ARM vs x86 thread !
No it's only x86 here.  But I'm pretty sure the "ARM" word will get mixed in here pretty soon... and off it goes  :lol:
Next time, please edit your post to include any new information rather than double-posting. It wastes space.

EDIT: xiongxioi, you too... -_-

It has nothing to do with ARM.
Not if random posters have anything to say about it!

-God Ginrai
 
Last edited by a moderator:
We got into this trouble in another thread.  I presented a claim of one brand new SoC being faster than another 5+ year old SoC.  I was told to 'prove it'.  I provided the closest match benchmark I could find that had been run in common to the two systems - and spent the next ~3 pages of that email chain reading how the benchmark, methodology and it's application were at fault.
No, they said "prove it is as fast as you say it is". There wasn't one person in that topic who believed that the x86 SoC was not faster than the OMAP 3530. You said it was 20x faster and people were sceptical about that.

-God Ginrai
And I was wrong.  The only benchmark presented showed it to be 17.6x faster on a single thread - but it has 4 cores.  Even if it's only 10x or 6x faster per core...  Again, we need more and better information than that.  You're welcome to find your own benchmarks and present them.

Back on topic...

In evaluating whether or not to use the Z3770, there are a lot of other factors to consider even beyond CPU processing power and even perf/watt.  For example:  What are the compatible wireless radios that can talk to it?  Is there a selection within those that fits the need?  Are there any restrictions on what type of display or touch panel that it talks to?  How would the nubs integrate?

There are piles of questions that need to be asked and answered before a production project plan can be compiled.  I.e. find an answer to, "Is it really a fit?" first, then plan out the process.
 
Not if random posters have anything to say about it!

-God Ginrai
Those people who talk about ARM at this early stage will be set on fire and then feed to a bunch of hungry wolves.

After this project is done, then ideally another feasibility study will be conducted for ARM, and we will enter the project tendering stage to choose the best suited.

The problem however, is that we are not sure of what good arm socs we can secure at this stage.
 
OMG, another ARM vs x86 thread !
No it's only x86 here.  But I'm pretty sure the "ARM" word will get mixed in here pretty soon... and off it goes  :lol:
Next time, please edit your post to include any new information rather than double-posting. It wastes space.


EDIT: xiongxioi, you too... -_-

It has nothing to do with ARM.
Not if random posters have anything to say about it!

-God Ginrai
The last I knew you weren't the posting cop or a moderator.  Please stop wasting visual board space by asking others to consolidate posts.  Disk is cheap.  He was responding properly to two separate conversational threads.
 
The last I knew you weren't the posting cop or a moderator.  Please stop wasting visual board space by asking others to consolidate posts.  Disk is cheap.  He was responding properly to two separate conversational threads.
I have always taken time to request others to follow forum etiquette. The waste of space I was talking about was visual board space as well. You are not contributing anything to this topic by making this attack on me. And you complain about wasting visual board space but then don't even trim the quote to the section you are addressing.

@xiongxioi Good idea, a separate topic like this for ARM once we have a better idea of what chips we are looking at will go nicely with this one about x86.

-God Ginrai
 
I back Grench on that one. It might be that the moderators, after due discussion, will advice someone (in private, to start off with) that they ought to change their posting style. Not others, and not openly. If this bothers you, make a report.

However, I do strongly agree with Grench even on the actual question. It is better to reply to two different parts of the thread in two different posts - Munging everything together only breeds confusion.
 
Last edited by a moderator:
The last I knew you weren't the posting cop or a moderator.  Please stop wasting visual board space by asking others to consolidate posts.  Disk is cheap.  He was responding properly to two separate conversational threads.
I have always taken time to request others to follow forum etiquette. The waste of space I was talking about was visual board space as well. You are not contributing anything to this topic by making this attack on me. And you complain about wasting visual board space but then don't even trim the quote to the section you are addressing.

@xiongxioi Good idea, a separate topic like this for ARM once we have a better idea of what chips we are looking at will go nicely with this one about x86.

-God Ginrai
You're simply upset because I decline to bow to your claimed authority in the matter of etiquette.  Believe it or not, there are differing thoughts around this - and your continually harping on people isn't helping.  Frankly, it's annoying and rather offensive how you ignore the topic and spend so much time trying to belittle others.  It's schoolyard bully behavior.  It derails and detracts from any semblance of polite conversation.  Let the mods moderate please.  If you're that offended, then by all means, report it to the mods and let them deal with it.

Thank you.

Edit:  Sorry - I had this reply lingering while I was doing other things and hand't seen the moderator reply before posting it.
 
Last edited by a moderator:
I back Grench on that one. It might be that the moderators, after due discussion, will advice someone (in private, to start off with) that they ought to change their posting style. Not others, and not openly. If this bothers you, make a report.

However, I do strongly agree with Grench even on the actual question. It is better to reply to two different parts of the thread in two different posts - Munging everything together only breeds confusion.
The report function is, as I understand it, to report troublesome users to the staff. Shenmue and xiongxioi are not troublesome users, and reporting them as such would be mean to them, even if they can't see my report.
And I have to disagree with you on the separation of posts. On my 1600x900 screen, I can see about 2 posts or 1 large one. A Double post like the ones that Shenmue and xiongxioi made takes up the whole screen, whereas I could have seen both of their full posts had they not double-posted. On the vertical 1080x1900 screen I am using, I can usually see ~4-5 normal posts. However, their double posts together take up 4/5 of the screen. There are all of these functions for formatting our posts given to us by BBCode, we have any number of ways to separate different topics within our posts.

^ Look! Another way to separate things! ^

Just use:


 That wasn't so hard now, was it Grench?

-God Ginrai
 
Last edited by a moderator:
And I have to disagree with you on the separation of posts. On my 1600x900 screen, I can see about 2 posts or 1 large one. A Double post like the ones that Shenmue and xiongxioi made takes up the whole screen, whereas I could have seen both of their full posts had they not double-posted. On the vertical 1080x1900 screen I am using, I can usually see ~4-5 normal posts. However, their double posts together take up 4/5 of the screen. There are all of these functions for formatting our posts given to us by BBCode, we have any number of ways to separate different topics within our posts.
 
Just use the mobile theme. There are extensions for FireFox that allow you to switch to the Android-browser string and use the javascript-heavy mobile UI
That's like telling me to set +m on an IRC channel on freenode just because an unvoiced user is flooding the channel.

-God Ginrai
 
Last edited by a moderator:
Back
Top