Hi - If you are running it to see just today's results ( n=1 ), I think your
Filter will only pass those tickers that had a cross today and so both
columns will probably show today's close... You might try something like
this ...
MACross = Cross(C, MA(C,10));
CloseLastCross = ValueWhen( MACros
Very good question. That was a head scratcher.
So the thing is, AmiBroker does a lot more work in a optimization
pass then execute AFL code. In fact, the AFL code may take very
little of the total run time.
As an example, using a database with good amount of data, write a afl
file that doe
Clearly in the future we will have more cores to work with.
The intel design is a very different approach to the problem. It
will be interesting to see what kind of performance they get out of
it when they have real silicon to demo.
The thing that impressed me the most about the cuda archite
I spent over an hour trying to figure this out and failed. It seems to
me the tools available in AFL should provide a solution but I could
not figure it out.
All I want is to create a column in my exploration that would compare
today's close to the close on the date the signal was created.
For inst
thanks for your insight.
I hope you dont mind sharing a little bit more detail
You said "
Get get the best performance, my AFL code makes one pass over the
> data, calling a Dll. The Dll takes all of the data needed by the
> calculation and loads a copy to the video card. This upload is
slow,
Maybe it all becomes easier with this here?
http://www.intel.com/pressroom/archive/releases/20080804fact.htm?iid=pr1_releasepri_20080804fact
Sorry,
I should have been more clear. The backtest I was referring to is
coded in AFL, not the built-in backtest. It is just part of the
indicator execution going on all the time in my system, so my comment
still holds, except the increased compute power is a beneficial factor.
BR,
Dennis
Wrong!
One line is parsed/executed ONCE, REGARDLESS of number of BARS under test.
So your 5 bar backtest will STILL consume 0.001 sec (0.1 microsec) for
your comment.
Note also that this measurement was done long time ago and on your modern
computers
timings should be ~5 times less.
B
I am running the following code (which I did not write)to create a
correlation matrix with an exploration. It works great.
What I want AB to do next though is to sift out all the symbols that
that meet a certian criteria and put them in a new watchlist.
ie. Sift or pool all symbols that have a
Tuzo, Joe,
So if I understand this correctly, in my code I use many loops for
complex functions and comment many lines because of the complexity.
If each line I comment takes 100 ns and I am running a 50,000 bar
backtest, I could end up with 800 ms of time wasted in // comments!
Since the
--- In amibroker@yahoogroups.com, "J. Biran" <[EMAIL PROTECTED]> wrote:
>
> It was first mentioned in release notes for AB 4.88. also
> quoting from another message:
> > -Original Message-
> > From: [EMAIL PROTECTED] On Behalf Of Tomasz
> Janeczko
> > Sent: Friday, October 20, 2006 12:32 P
The following code appears for RWIHI in AB
RWIHi( minperiods, maxperiods);
I have only found a single period used in any of Michael Poulos calculations,
and he talks about comparing a long term RWIHI(38) [for example]with a short
term RWILO(7) , but each is calculated using just one period and t
It was first mentioned in release notes for AB 4.88. also
quoting from another message:
> -Original Message-
> From: [EMAIL PROTECTED] On Behalf Of Tomasz
Janeczko
> Sent: Friday, October 20, 2006 12:32 PM
> To: [EMAIL PROTECTED]
> Subject: Re: [AmiBroker-at] Re: Conditional #include
>
>
I, too, would like to have nested /* */ commenting.
As for //, they are great when commenting out only one or two lines.
However, given a large block of code, or a wordy multi-line comment, /*
*/ is what I want to use.
As for compatibility with other programming languages, none of the
langu
roc()
- Original Message -
From: Mr. Valley
To: amibroker@yahoogroups.com
Sent: Tuesday, August 05, 2008 10:18 AM
Subject: [amibroker] % Year over Year
How does one code a Year over year chart?
For example, on a relative percentage chart, the years 2005,2006,2007,2008
c
--- In amibroker@yahoogroups.com, "J. Biran" <[EMAIL PROTECTED]> wrote:
>
> Well, as you can tell, I am for solution 1. If afl can keep
> track of parentheses I see no reason why it cannot do the
> same for /* */.
You're right -- it can be done but I can think of a couple of reasons
why it shouldn
Wow, if I ever had to worry about all that spirituality, psychology,
philosophy, Indian spiritual philosophy,
God, yoga, meditation, prayer, singing/chanting, dancing, spiritual
practices,
temporary intimacy with the Gods, psyhic powers, Western spirituality,
and spiritual teachings,
I don't
Interesting. Looks like they are using DSP chips and a custom board
for a similar application. 32MFLOPS is not very much. The modern
graphics cards deliver GFLOPS. The high end tops a teraflop.
As far as execution environments, personally, I use Tradestation for
execution and AmiBroker fo
dloyer,
Nice comments.
Is the card in this link potentially something that could be used in the
same manner as the video cards?
http://www.modulusfe.com/nnpc12/default.asp
I am currently looking at AB and another platform for auto trading program
development and trying to see what is the "best"
Amazing... thanks so much for bringing this to our attention.
- Original Message -
From: "dloyer123" <[EMAIL PROTECTED]>
To:
Sent: Tuesday, August 05, 2008 10:06 AM
Subject: [amibroker] Re: Freakishly fast backtest using 64 cores
> The same hardware that lets a game render life like 3
How does one code a Year over year chart?
For example, on a relative percentage chart, the years 2005,2006,2007,2008
could all be different colors...
Would it be a foreign symbol for the date range of the one year?
Thanks
Keep in mind that this is hand optimized code. This explores the
edge of what is currently possible. The logic of the trading system
is coded directly on the graphics cores.
A Dll that allowed the trading system logic to be defined in AFL
would be much more useful. .
--- In amibroker@yaho
The same hardware that lets a game render life like 3d graphics.
They perform single percision floating point math at rates that are
in the relm of the super computers of a few years ago. Lots of
floating point operations per second and lots of bandwidth to
memory. The graphics cores have ad
This uses the mid range video card that happened to come with my
system, a 9800GT. The newer 260 and 280 cards are 3 to 4 times
faster. The 260 can be found at best buy for $300. Some laptops
have compatible cards as well.
The video card has its own memory, mine has 512MB, some have as much
Re: [amibroker] Freakishly fast backtest using 64 coresYes, I'm puzzled too.
Looping should depend upon the speed of the CPU (or so I thought). Why would
putting the data in a different memory store increase looping speed?
I noticed that AB runs its second backtest much faster than the firs
Thank you very much,
you give me the missing informations. I just wondered in the past why
AB used margin even if I had enough cash. That's clear now ...
WIth regards, Guido
--- In amibroker@yahoogroups.com, "Mike" <[EMAIL PROTECTED]> wrote:
>
> Tip,
>
> Sorry, I used AmiBroker terms in the c
Thank you very much,
you give me the missing informations. I just wondered in the past why
AB used margin even if I had enough cash. That's clear now ...
WIth regards, Guido
--- In amibroker@yahoogroups.com, "Mike" <[EMAIL PROTECTED]> wrote:
>
> Tip,
>
> Sorry, I used AmiBroker terms in the c
27 matches
Mail list logo
