Tex!! Is this right?

Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
edited August 2003 in Hardware
Figured I'd ask for ya by name since you know more about hard drives/storage than anyone else I know...

So, I joined www.powernotebooks.com 's forums, since I own one of their notebooks... so some guy on there says the following
If you use RAID 1-5's you can up your Stripe Size to get better Perfromance, esspically if its a server. I dont set servers less than 128kb Stripe with RAID 1 and my RAID 1 512kb Stripe Beats out a single drive by a long shot, the only problem is that you losse alot of Hard Drive Space in the Process

Now, the guy in question isn't the sharpest knife in the drawer (he's also convinced that a Radeon 9800 Pro draws 60a of power... nevermind the fact that most PSes couldn't handle it if it did, but I'll deal with that post later, when I have some more time on my hands, probably after summer school is over). Anyhow, that goes against everything I've ever read... RAID 1 is slower than a single drive period, no exceptions, etc., correct? Or am I losing it?

http://www.powernotebooks.com/forum/viewtopic.php?p=1165#1165

Comments

  • citrixmetacitrixmeta Montreal, Quebec Icrontian
    edited August 2003
    well, raid 1 slow in writing, but faster in reading.
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited August 2003
    I always thought it was slower either way because at most you'd only get the equivalent of 1 drive's read/write speeds, plus you now have the overhead from the raid controller...
  • edited August 2003
    The read operation on a Raid 1 array will read data from both drives at once, thus improving output (assuming you are using a decently implemented raid contoller).
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited August 2003
    ah
  • TexTex Dallas/Ft. Worth
    edited August 2003
    Originally posted by Geeky1
    Now, the guy in question isn't the sharpest knife in the drawer [/url]

    I hate to tell ya this Buddy but....... he isnt a knife..... He's a spoon.. A PLASTIC spoon even.

    Tex
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited August 2003
    LOL :D
  • McBainMcBain San Clemente, CA New
    edited August 2003
    Originally posted by Tex


    I hate to tell ya this Buddy but....... he isnt a knife..... He's a spoon.. A PLASTIC spoon even.

    Tex

    Perhaps spork worthy?
  • ShivianShivian Australia
    edited August 2003
    Originally posted by Geeky1
    Now, the guy in question isn't the sharpest knife in the drawer (he's also convinced that a Radeon 9800 Pro draws 60a of power... nevermind the fact that most PSes couldn't handle it if it did, but I'll deal with that post later, when I have some more time on my hands, probably after summer school is over). Anyhow, that goes against everything I've ever read... RAID 1 is slower than a single drive period, no exceptions, etc., correct? Or am I losing it?
    A Radeon 9800 operates at 0.8V so the power consumption according to that guy would be about 48W which is realistic. There will be some transformer or something to convert the 5V? to 0.8V somewhere so that is not an unrealistic amp rating.
  • TexTex Dallas/Ft. Worth
    edited August 2003
    People usually greatly underestimate the wattage needed for the highpower video cards in general. And how it impacts the PSU choice.

    Tex
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited August 2003
    U sure it's 0.8v? I thought it was ~1.5...

    @ any rate, I think he was referring to the power consumption on the 5v line... but if he's referring to 60a @ 0.8v then yes, it's realistic...
  • ShivianShivian Australia
    edited August 2003
    Yeap I'm sure. Someone asked a little while ago so I looked it up.
    0.8V -> AGP 8x
    1.5V -> AGP 4x
    Not sure about the rest but one of them is 3.3V... just can't remember how it goes.

    And with the transistor counts of vid cards nowadays I would expect the power requirements of a vid card to be as much if not more than modern processors (and then you factor into account that vid cards generally lag processors in terms of process size).
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited August 2003
    Shivian, if that's what you mean...

    The core itself doesn't actually run at 0.8v. 0.8v is the signaling voltage- it's the voltage that the board and card use to communicate across the AGP bus. The core itself is not bus-powered; in fact, it never was. I looked up the AGP spec for this debate I'm having with this guy and AGP cards have always been powered by either the 3.3v or 5v or both... with the advent of these external power connectors, it's using 12v or 5v, or both. But the chip's core voltage is probably more along the lines of 1.5v, since vmodding gets 1.6+... and 1.5v @ 60a is 90w. Since something like 99% of the power that goes into a chip comes out as heat, the thing can't draw that kind of power because there's no way ATi could cool it with the cooler that's on it...
  • ShivianShivian Australia
    edited August 2003
    Nononono sorry I don't mean the core is 0.8V. Just the card is supplied 0.8V through the 8x AGP slot.
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited August 2003
    yea... thought so. wanted to double-check...
Sign In or Register to comment.