This is a multi-part message in MIME format.
--------------010807000608050100070300
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
AAsk a écrit :
>> I only mention that I currently manipulate matrices equivalent to a
>> vector of nearly one billion numbers without any kind of APL anomaly
for
>> what I am looking for....
>>
>
> Using which APL? Running on what hardware?
>
Good morning "AAsk",
I use IBM APL2. In some areas I am rather near to APL2 implementation
limits, but this is largely enough for my cases. It runs on an assembled
system : now old Asus A7M266-D MotherBoard, 2xAthlon MP 2000+ with 3.5GB
RAM and Suse Linux O/S : this system has been rebooted very little in
many years, and only for maintenance (of which dust removal from fans)
It will be replaced by another system currently under assembly : Asus
KFNS4-DRE with 2 AMD Opteron Quad 2347 and 32GB RAM (and plans for 64GB
RAM), under same O/S. According to my estimates I may gain more than one
order of magnitude in overall efficiency : not bad if I am right !
I may face unexpected problems but this is another story ...
Asus KFNS4-DRE and AMD 2347 are very recent announcements in France.
It is worthwhile mentionning that France is generally 1 to 2 steps
behind UK or Germany .... my personal opinion : mind set is too
conservative !
I am still waiting for AMD to deliver for mid Jan hopefully.
As you notice, no need for very expensive rig although new system will
be nearly twice more expensive than old one, it remains very affordable.
I hope these details are satisfactory. If not, please feel free to
contact me privately.
Regards, Gilbert
--------------010807000608050100070300
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
AAsk a écrit :
<blockquote
cite="midebd95f9a-76d2-49df-ae68-abf988ed6495@[EMAIL PROTECTED]
"
type="cite">
<blockquote type="cite">
<pre wrap="">I only mention that I currently manipulate matrices
equivalent to a
vector of nearly one billion numbers without any kind of APL anomaly for
what I am looking for....
</pre>
</blockquote>
<pre wrap=""><!---->
Using which APL? Running on what hardware?
</pre>
</blockquote>
Good morning "AAsk",<br>
I use IBM APL2. In some areas I am rather near to APL2 implementation
limits, but this is largely enough for my cases. It runs on an
assembled system : now old Asus A7M266-D MotherBoard, 2xAthlon MP 2000+
with 3.5GB RAM and Suse Linux O/S : this system has been rebooted very
little in many years, and only for maintenance (of which dust removal
from fans)<br>
<br>
It will be replaced by another system currently under assembly : Asus
KFNS4-DRE with 2 AMD Opteron Quad 2347 and 32GB RAM (and plans for 64GB
RAM), under same O/S. According to my estimates I may gain more than
one order of magnitude in overall efficiency : not bad if I am right !<br>
I may face unexpected problems but this is another story ...<br>
<br>
Asus KFNS4-DRE and AMD 2347 are very recent announcements in France.<br>
It is worthwhile mentionning that France is generally 1 to 2 steps
behind UK or Germany .... my personal opinion : mind set is too
conservative !<br>
I am still waiting for AMD to deliver for mid Jan hopefully.<br>
<br>
As you notice, no need for very expensive rig although new system will
be nearly twice more expensive than old one, it remains very
affordable.<br>
I hope these details are satisfactory. If not, please feel free to
contact me privately.<br>
Regards, Gilbert
</body>
</html>
--------------010807000608050100070300--
|
