headers
Elektron | 1 Aug 2004 18:31
Picon
Favicon

Re: RC5-72 progress


On 31 Jul, 2004, at 18:54, Per Hallström wrote:

> I don't know anything about the upcoming processors from Intel or AMD, 
> but maybe they both come to the conclusion that they would benefit 
> from an architecture similar to the one present in the G5 processors. 
> There must be some reason for IBM to make them (the G5 processors) 
> that way, so maybe the others should too.

Well, there are a couple things. First, the PowerPC family has 32 
registers (though the first two/three are special-purpose), and second, 
RISC is better (at least, it can do a=b+c, instead of a=b; a+=c).

Of course, the AMD 64 is fairly competitive with the G5 (10Mkeys for 
2500 MHz, since it's probably not the 2500+), though I'm not sure if 
this is in 64-bit mode or not (I assume it isn't).

- Purr
_______________________________________________
rc5 mailing list
rc5@...
http://lists.distributed.net/mailman/listinfo/rc5
Permalink | Reply | 2 comments |
headers
Daniel Quintiliani | 1 Aug 2004 19:34

Prediction updated for month of August.


After plotting the August 1 data (140 GKeys/sec), I recreated the trendline.
It gave this formula:

y = 47.585Ln(x) + 14.932
R^2 = 0.8898

That makes 403 years 4 months from Dec 2002.

Notice how people are starting to leave RC5? Other people will join soon, with
greater processing power. This basically evens it out with slightly more each time. That's
why I chose a logarithmic formula for the trendline. You can look at the graph and see that
this is happening and that the graph is almost logarithmic.

Think I should submit an enhancement request for the stats to automatically predict
based on the graphs? The stats system would be able to get a much better estimate
than I would because the graph keeps track of every day's keyrate and I'm only looking
at the first day of the month.

-Dan
Permalink | Reply | 6 comments |
headers
Christopher Hicks | 1 Aug 2004 19:39

Re: Prediction updated for month of August.

On Sun, 1 Aug 2004, Daniel Quintiliani wrote:
> The stats system would be able to get a much better estimate than I 
> would because the graph keeps track of every day's keyrate and I'm only 
> looking at the first day of the month.

Maybe you've already considered this, but I thought the stats varied 
considerably based on the day of the week.  The first of the month this 
month is on a weekend which are historicly low keyratewise while July 1 
was in the middle of the week which are histoiricly higher.

--

-- 
</chris>
_______________________________________________
rc5 mailing list
rc5@...
http://lists.distributed.net/mailman/listinfo/rc5
Permalink | Reply | 2 comments |
headers
Fuzzy Logic | 1 Aug 2004 19:21
Picon

Re: RC5-72 progress

Be careful how you phrase that. RISC is simply "reduced instruction
set computer" and can be used to describe processors such as that
found in the S/390 (what I work on), and so not all RISC processors do
what you suggest. In fact, I would venture to say that you are
describing a capability usually associated with CISC.

Fuzzy
-- 
"Men may doubt what you say, but they will believe what you do."
*Lewis Cass {1782-1866 American Politician}

On Mon, 2 Aug 2004 00:31:55 +0800, Elektron <elektron_rc5@...> wrote:
> 
> On 31 Jul, 2004, at 18:54, Per Hallström wrote:
> 
> > I don't know anything about the upcoming processors from Intel or AMD,
> > but maybe they both come to the conclusion that they would benefit
> > from an architecture similar to the one present in the G5 processors.
> > There must be some reason for IBM to make them (the G5 processors)
> > that way, so maybe the others should too.
> 
> Well, there are a couple things. First, the PowerPC family has 32
> registers (though the first two/three are special-purpose), and second,
> RISC is better (at least, it can do a=b+c, instead of a=b; a+=c).
> 
> Of course, the AMD 64 is fairly competitive with the G5 (10Mkeys for
> 2500 MHz, since it's probably not the 2500+), though I'm not sure if
> this is in 64-bit mode or not (I assume it isn't).
> 
> - Purr
(Continue reading)

Permalink | Reply | 1 comment |
headers
Daniel Quintiliani | 2 Aug 2004 11:45

Re: Prediction updated for month of August.


> On Sun, 1 Aug 2004, Daniel Quintiliani wrote:
> > The stats system would be able to get a much better estimate than I 
> > would because the graph keeps track of every day's keyrate and I'm only 
> > looking at the first day of the month.
> 
> Maybe you've already considered this, but I thought the stats varied 
> considerably based on the day of the week.  The first of the month this 
> month is on a weekend which are historicly low keyratewise while July 1 
> was in the middle of the week which are histoiricly higher.
> 

Good idea. First I fixed a few bugs with the spreadsheet (bad estimates of the
graph, an extra month for some reason).

Next, I split it into half-months rather than months and looked at the middle of the 
month as well as the month to give a better estimate.

y = 42.817Ln(x) - 1.772
R^2 = 0.8589

That gives 461 years 1 month from Dec 2002 for a better estimate. A 58 year difference...

So, I guess we can say that RC5-72 will probably be finished in about 400-500 years.

I will test the spreadsheet for a while longer and then maybe submit an enhancement
request.

-Dan
(Continue reading)

Permalink | Reply | 1 comment |
headers
Ray Booysen | 2 Aug 2004 12:54

Re: Prediction updated for month of August.

Daniel Quintiliani wrote:

>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>  
>
>>On Sun, 1 Aug 2004, Daniel Quintiliani wrote:
>>    
>>
>>>The stats system would be able to get a much better estimate than I 
>>>would because the graph keeps track of every day's keyrate and I'm only 
>>>looking at the first day of the month.
>>>      
>>>
>>Maybe you've already considered this, but I thought the stats varied 
>>considerably based on the day of the week.  The first of the month this 
>>month is on a weekend which are historicly low keyratewise while July 1 
>>was in the middle of the week which are histoiricly higher.
>>
>>    
>>
>
>Good idea. First I fixed a few bugs with the spreadsheet (bad estimates of the
>graph, an extra month for some reason).
>
>Next, I split it into half-months rather than months and looked at the middle of the 
>month as well as the month to give a better estimate.
>
>y = 42.817Ln(x) - 1.772
(Continue reading)

Permalink | Reply | 0 comments |
headers
Elektron | 2 Aug 2004 16:24
Picon
Favicon

Re: RC5-72 progress


On 2 Aug, 2004, at 01:21, Fuzzy Logic wrote:

> Be careful how you phrase that. RISC is simply "reduced instruction
> set computer" and can be used to describe processors such as that
> found in the S/390 (what I work on), and so not all RISC processors do
> what you suggest. In fact, I would venture to say that you are
> describing a capability usually associated with CISC.

Neither x86 or M65K support separate source/destination registers. RISC 
and CISC also lose their meaning when a CISC processor has 100-some 
instructions, while the successor RISC has over 200.

The only real difference between CISC and RISC is that CISC does things 
which are relatively expensive to implement (such as memory access in 
most instructions, and native support for binary-coded-decimal, and 
instructions up to 12 bytes long or so), while RISC tends not to. RISC 
platforms also tend to prefer aligned loads and stores (whereas the x86 
has one-byte instructions), and equal-sized instructions (e.g. 32 
bits).

Having different source and destination registers isn't hard to 
implement, since you have to access registers anyway (it also reduces 
data dependency stalls, so I can write very fast CRC code [1]), nor is 
branch if decrement non zero (which seems to have the same latency as a 
branch). It just requires more bits per instruction, and you free bits 
by not needing to access memory (and often get more bits per 
instruction anyway).

The PowerPC is called RISC but can load/store multiple words, byte
(Continue reading)

Permalink | Reply | 0 comments |
headers
Ray Booysen | 2 Aug 2004 17:59

Raw Participant Logs

Hey All.

I remember being able to download raw logs of your keys done.  Can 
anyone remember the URL.  I remember it was similiar to the URL that you 
navigate to to get your graphical stats.

Regards
Ray Booysen

--

-- 
Ray Booysen
dnetc@...

_______________________________________________
rc5 mailing list
rc5@...
http://lists.distributed.net/mailman/listinfo/rc5
Permalink | Reply | 1 comment |
headers
waldo kitty | 2 Aug 2004 16:31
Favicon

Re: Prediction updated for month of August.

Daniel Quintiliani wrote:
> -----BEGIN PGP SIGNED MESSAGE-----

[trim]

> Notice how people are starting to leave RC5? Other people will join soon, with
> greater processing power. This basically evens it out with slightly more each time. That's
> why I chose a logarithmic formula for the trendline. You can look at the graph and see that
> this is happening and that the graph is almost logarithmic.

one should also note that it may not be a case of people leaving the project and others joining... in my case,
my clients have 
preference for OGR stuff and they will switch over to RC5 when they are out of OGR stuff... additionally, the
clients i'm currently 
running on my boxes seem to hold results until some trigger (that i've not yet determined) is tripped...
i've had to go around to 
them occasionally and manually flush them... yes, i know that i need to upgrade them... maybe i'll do that
with my coffee this 
morning...

FWIW: those clients are all running dnetc v2.9008-490-GTR-04050719 for Win32 (Windows 4.10).

[time passes]

hummm... maybe i won't update those clients... seems that they are the latest of the win clients already...

maybe my problem of them sitting on results is a config problem? the client is set to use internal defaults
instead of x results or 
y time... i know that one of the machines flushed some 20+ results the other day and i was normally running
them so that they
(Continue reading)

Permalink | Reply | 2 comments |
headers
Jim C. Nasby | 2 Aug 2004 20:16

Re: Raw Participant Logs

>From http://stats.distributed.net/participant/phistory.php?project_id=8&id=39622:

<!-- IMPORTANT NOTE TO SCRIPTERS!
This page, like many stats pages, has a version which is far more suitable
for machine parsing.  Please try the url:
http://stats.distributed.net/participant/phistory_raw.php?project_id=$project_id&id=$id
-->

You might also want to look at
http://cvs.distributed.net/viewcvs.cgi/stats-html2/

On Mon, Aug 02, 2004 at 04:59:26PM +0100, Ray Booysen wrote:
> Hey All.
> 
> I remember being able to download raw logs of your keys done.  Can 
> anyone remember the URL.  I remember it was similiar to the URL that you 
> navigate to to get your graphical stats.
> 
> Regards
> Ray Booysen
> 
> -- 
> Ray Booysen
> dnetc@...
> 
> 
> _______________________________________________
> rc5 mailing list
> rc5@...
> http://lists.distributed.net/mailman/listinfo/rc5
(Continue reading)

Permalink | Reply | 0 comments |

Gmane