dupa

Author Topic: GPU SQL Queries  (Read 3142 times)

kullboys

  • Newbie
  • *
  • Posts: 1
    • View Profile
GPU SQL Queries
« on: March 21, 2018, 03:12:36 am »
Ch?o mọi người
b?i h?t
« Last Edit: May 18, 2018, 12:29:59 am by kullboys »

Art

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 3848
    • View Profile
Re: GPU SQL Queries
« Reply #1 on: March 21, 2018, 09:09:12 am »
Welcome!

Given the price to size ratio of the SSD's, a lot of people are still good with the standard Hard Drives.

Sure the overall speed of SSD's is a nice feature but storage being what it is, the better bang for the buck for some
is still the old fashion electro-mechanical Hard Drive.

A lot of us have noticed that Hal seems to take a bit longer to process its thoughts as the brain file increases in size. Hopefully, an SSD might help in that case.

It would be interesting to see an experiment with HAL and a large Halbrain.db file doing a comparison. One on a Mech.Hard Drive and the other on an SSD and both being timed for responses.

If anyone has done such an experiment, please consider sharing the results.

Good topic kullboys.
In the world of AI it's the thought that counts!

- Art -

cyberjedi

  • Hero Member
  • *****
  • Posts: 810
  • The Mighty Hal Machine
    • View Profile
Re: GPU SQL Queries
« Reply #2 on: March 22, 2018, 12:43:06 am »
RE: Hal Speed...

lololol.you guys..

Ok ,Schools in session:

Create a VM, Install Hal on VM , Run Hal on VM.......
 VM=Virtual Machine
Schools Out: Old tricks are the best tricks, remember all the old games that ran on a ROM drive???

best wishes
cyberjedi
« Last Edit: March 22, 2018, 12:45:43 am by cyberjedi »

lightspeed

  • Hero Member
  • *****
  • Posts: 6761
    • View Profile
Re: GPU SQL Queries
« Reply #3 on: March 22, 2018, 12:52:22 pm »
cyberjedi , while art has mentioned about hals lagging answer time response back,  i wrote a message to dooniecram  , but he has stopped working on code writing on his super plug in . so i am going to post a copy of the same message here to you now .

 i think the key to hal also sounding more intelligent in answering the user lies within plug ins like yours  but needs to have hal acess the subject word with something on wikipedia online . then follow up with subject matter already learned within hal , if none is there , then no response is given or else a random number generic phrase like , isn't that interesting?  etc. etc.
if something was created like this hal may have to be set for two paragraphs only or whatever , in case the article is very large .  :) ;)

i think doing this type of thing not only from wikipedia but many other sites on various things will help hal greatly . their may already be a wikipedia but is on a ask only bases , would be nice if hal just spoke up on his own as if he knew about it.  :)

p.s. you will have to make sure that hal doesn't speak the [1] etc. which refers to other parts or pictures in articles. and any other unwanted symbolizes will have to be eliminated so hal doesn't say them.

IF this type of plug in can addition can be made maybe one could have a in line list of various INTERNET sites that hal will acess (without opening the page ) and read info from the subject , they could even be randomly picked if more then one site conflicts with having the same subject to read from .

as i said i believe this would make hal appear a lot smarter as he will answer and or talk about subjects .
example inspiration sites, etc. etc. , using this type on many list of online access will also keep hals brain from growing to much and will also let hal have online information .
 maybe a learn true and or false setting can be included so user has option of hal learning from the website or not learning. in another words information use on the fly only.


I think as i mentioned before cyberjedi that the key to making hal seem a lot smarter is through retrieval of information etc, of the internet (similar to the cloud , ) this way hal can respond  a lot faster ( basically like sophia a.i. is doing )  if this can be done with response list of internet sites and deduction on subject , a long list of ( and able to add more ) will result in hal having a very large area of answers to choose from , maybe after accessing a page for info. hal would also add a response   on the subject from hals own data base brain info.

I hope you can create something like this that we can all add in the s.a.i.l.l.y. plug in . this is one way hal could answer a lot faster , be smarter , without upgrading to solid state drives , especially when some people can't afford to . etc.

 

Spitfire2600

  • Sr. Member
  • ****
  • Posts: 251
    • View Profile
Re: GPU SQL Queries
« Reply #4 on: March 25, 2018, 01:29:08 am »
I have considered this previously. I believe the limitation is VBScript/SQL linking as it can only use 1 core. I could be totally wrong about that though or something may have changed since I last googled it.

-Spitfire2600