Poll: Reign

[Kris]

As I pointed out here, I think the following is a useful stat.
At yesterday's rate, Team Anandtech will hold first place for 79.672706 days.
The addition of one athlon at 1 ghz will increase that time by 4.1047198 hours.
The addition of one P3 at 1ghz will increase that time by 3.3803525 hours.

Is this interesting enough to continue posting it?
Hey..I could make a poll on that.
Let me try.

 

[Engineer]

Kris,

That is a wonderful stat system.....Makes me more motivated than ever...

Let's all take this thing and hold on to it ....#1

 

[TwoFace]

Kris:

You da man!

I love these kinda tidbits... hmmm do you have stats for different processor speeds as well? I'm gonna assimilate the GF when I visit her next week, and I'd like to know how many hours/minutes that would add to our reign...

btw: it's a duron 650 maybe I'll even OC it some just for TA...

With love and respect your fellow TA member

Two-Face

 

[JimMc]

TwoFace--here's a page listing relative CPU speeds CPU Speed By all means O'C, there is a linear increase in speed .

 

[Kris]

Two-face, Jim hit it on the head. Speed is linear within a given architecture.
So, a Duron at 650 adds 2.67 hours. (4.1*.650)
and a Duron at 900 adds 3.69 hours. (4.1*.900)

So which ya gonna go with??

Another reference point, a celeron at 550 would add 1.9 hours. (3.4*.550)

 

[dkappos]

Speed is linear but the time Team Anandtech will hold first place is not.. so if 1GHz increases the time by 4.1047198 hours that doesn't mean that 500MHz will increase the time by 4.1047198/2 or that 2GHz will increase the time by 4.1047198*2

 

[Kris]

dkappos
Please help me fix it. I am neither a mathematician nor a physicist.

Here is how I figured it:

{(Tt-Dt)/[Dr-(Tr+x)]}-{(St-Tt)/[(Tr+x)-Sr]}=d

Tt=TeamAnandtech total blocks
Dt=DPC total blocks
St=Slashdot total blocks
Tr=Teamanandtech blocks/day
Dr=DPC blocks/day
Sr=Slashdot blocks/day
x=assimilated processor blocks/day
d=days in first with processor blocks (x) added to rate

d-days in first w/o x * 24=hours in first as a linear function of x

If x is .5x or 2x it is still linear as long as you account for the impact on both ends, overtake slash and overtaken by dpc.
I could very easily have made an error, and would NOT resent any assistance.

This is all bs anyway, because we will assimilate until Tr>Dr, and d is infinite!

(I am serious, if anyone can tell me what to change...an incorrect stat serves absolutely no purpose! This is simply the best I can do, it has been almost 15 years since I had any math course.)

 

[dkappos]

I am not a mathematician but from your second post I understood that even if we assimilate all the computers of the world they will still reach as (even if this is after a few thousand years) which is obviously wrong.
I am not saying that {(Tt-Dt)/[Dr-(Tr+x)]}-{(St-Tt)/[(Tr+x)-Sr]}=d is wrong (I can't check it after a few beers).

 

[Kris]

dkappos
In no post did I intend to imply that DPC will reach us no matter what we assimilate. I thought it obvious that when our rate exceeds theirs then our reign becomes infinite. In fact, that's why I said



<< This is all bs anyway, because we will assimilate until Tr>Dr, and d is infinite! >>



Which means, for any who don't want to look it up, when our daily rate exceeds their daily rate, the time we hold first place is infinite.

Anyway. :|
I intended not to update until this was resolved.
But I don't see a problem with the math, only the fact that acceleration is not included. Accuracy is not the point, rather, the point is to provide an easier to handle, real-world estimate of the benefits of your efforts to assimilate. So...


Despite a boost from slash, a dip from dpc yesterday caused projected 'reign' to artificially climb to 108 days. Artificially because our own rate decreased.

1000 AMD mhz --->6.37 hours
1000 Intel mhz__>5.25 hours

Go get some!

 

[BurntKooshie]

I would like to point out http://teamanandtech.dhs.org/speed.asp. Its more accurate than D.nets, because we have many, many users values for each CPU, and we can verify them better than D.net

 

[Kris]

B|K
I used the page at Mika's for this stuff, I assume you are referring to JimMC's link?