PDA

View Full Version : Opinions wanted (dead or alive)


KingArthur
Sep 29, 2001, 04:27 AM
I had an idea a few weeks ago and I bet it has been tried in the past, but I would like to see what you guys think of it. Here we go:
With the growing demand for virtual ram, why have HDs remained at one head for reading and writing data? Would not a duel-headed hard drive (each head on the opposite side of the other) astoundingly increase the efficiency of the computer? Think of it, no need for the computer to spend forever skipping from one piece of the HD to another, instead, while one head is reading, the other can be seeking for the next spot to read from or writing elsewhere. I am guessing that this has been tried in the past, but remember the strides that we are making...we may be able to do it now.

Another idea:
Why is there not a substitute to "sleep" where, instead of having to keep the ram powered so that you "wake" in your previous state, the computer saves the ENTIRE ram bit for bit to the HD and just read it bit for bit from the HD when "waking-up." Think how little time it takes to recall say 64MB of HD space to the RAM or vice verca. That would be EXTREMELY energy efficent and almost as quick as waking from sleep.

These are just some of the random ideas that I have during English IV or Calculus AP or Study Hall or my personal most hated, AMERICAN GOVERNMENT. Maybe if the classes posed a challange, I would not have so much time to think these things up. Tell me what you think, please. Who knows...These may be the kind of new things Apple needs to get back with the competition again. You can either respond here or e-mail me at KingArthur10@mac.com

mnkeybsness
Sep 29, 2001, 07:01 PM
that idea for the hardrive is good, does anyone know if that really has been tried before???

mymemory
Sep 30, 2001, 10:50 AM
I thought about 2 optical eyes reading the same disc, each one in opposites sides may be, or on the same harm but pointing to 2 different places. That would increase the transfer data. May be there is something in the technology that make the actual devices more efficient.

What is having me thinking was a dream I had last night about a dog attacking me. I had the same dream a few days before the NY attack. I was a very unusual dream. The relation between one and the other, I got it a week after the incident when I saw in the discovery channel a show where a man could predict some things in the future, and usually the dogs mean terrorist attack. I spec this to be ******** of my head but the best way to avoid a dream from coming true is talking about it. In a week I'll delete this post.

biker4mac
Oct 1, 2001, 03:49 PM
In a way, what you are talking about is RAID 0 (striping) on a single drive. The question would be whether the two heads would work either (a)in tandem (closer to true RAID 0) -- one reading (or writing) the odd bits while the other was reading (or writing) the even bits OR (b) separately -- each reading and writing information independent of the other.

The first one would gain speed advantages without a lot of fuss -- the heads could be offset so that you would have sections of odd and even segments on the disk like rings on the cross section of a tree.

The second one would be more versatile -- one could be reading while the other wrote, etc -- but would probably need a certain amount more complexity so that they did not interfere with one another.

It would seem that the idea is a logical step forward. Twice the data speed without spinning the actual disk twice as fast (my 52x CD already sounds like a jet engine!).

mnkeybsness
Oct 1, 2001, 05:28 PM
now that i think of it, it would be pretty complex to structure. i think that the heads working separately would work best. perhaps one head just for reading and one for writing would eliminate the complexity of the matter.

KingArthur
Oct 1, 2001, 11:07 PM
See, I figured that with the HD, you would probably have to buy a new PCI card for campatibility (like buying a SCSI card for internal SCSI drives). Basically, you would have a sort of second processor on the card (2nd for the HD) so that each would work independantly. It would also involve a new device dirver so that the OS would be able to tell it what to get one step ahead of time. I know that the complexity is great, but so is having a CD-RW/DVD-R. Hey...I think there is a real market for this with increasing virtual memory usage. Another problem I thought of was that the bus speed or may have to be increased to actually give the full performance increase or the bus would need to allow double the data to go through at a time (64 or 128 bits instead of 32 or 64 bit, respectively). I am no computer technition (can't even spell the word), so I do not know what exactly would be involved in such a monumentous task. This idea COULD be applied to many things such as 3.5" floppies, CD-ROMS, and basically every other storage medium. And this is but one of many ideas I have on inventions. Maybe one day my ideas will take root.

Thanks a bunch for the input and keep it comming. Also, I am supprised that no one talked about my other, less major idea of a dormant stage. One thing that I am waiting for is magnetic RAM. That would do the same thing as this idea.

A new idea I had a few days ago:
Everyone knows how annoying the time for 3.5" floppies takes to pop-up on the screen. They virtually freeze the system until they are done. Come on. 3.5" floppies have been around since the 80's. You think we would have made the mechinisms smaller by now and maybe put a sub-processor and ram on the drive so that it can gather the info on that and then give it all at once to the system. How many CPU cycles would be saved by that?! CDs have the same problems, so do most storage mediums. We have sub-processors and RAM on video cards and sound cards; why not on storage drives?

Next:
When the heck are we going to see increased speed for the PCI cards? Only 33Mhz! We've had that since the early-mid nineties! The CPU is not the only important thing here. We need everything (the bus, PCI bus, ram bus, etc) to be increased.

Also:
When are we gonig to see backside RAM that works independantly of the main bus? Like cache, we need to have the RAM start following in the footsteps. Who know? On chip RAM some day with a quad-head backside HD someday? We are the future: let's start getting those ideas out and make them work, not keep them bottled-up because we think that they might be stupid. The only stupid idea is the one which was not tested or expressed. Lets start sharing our ideas!

This is all I can remember for the moment dealing with computers. I need to remember to write these things down. I see my problem....it involves remembering! hehe.
And guess what everyone!!! One week 'til my 18th b-day!!!! (October 8)