Find the answer to your Linux question:
Results 1 to 8 of 8
Like Tree1Likes
  • 1 Post By cynwulf
I miss the days when you could look at the name of a chip and tell how it compared to another chip. You knew that P500 was about 20% to ...
Enjoy an ad free experience by logging in. Not a member yet? Register.
  1. #1
    Linux Engineer MASONTX's Avatar
    Join Date
    May 2010
    Location
    Mason Texas
    Posts
    934

    A rant about cpu naming


    I miss the days when you could look at the name of a chip and tell how it compared to another chip. You knew that P500 was about 20% to 25% more powerful than a P400, or that if your current computer had a P400 cpu and you were looking at a new machine, that an 800 mhz cpu would be about twice as fast. Now, with the numbers having nothing to do with the speed or performance, it is a real PITA to tell what kind of a performance boost you will get by replacing one cpu with another. For example, how does an i3-2370 compare to an i3-370, both listed as a 2.4 Ghz 2 core processor? Now throw in i5-2430 and i5-450, also listed as 2.4 Ghz, not to mention low voltage and ultra low voltage chips that are advertised as high performance with less power. Why can't they come up with an objective performance test and a reference board for a cpu that allows you to compare one chip to another. Granted, some tests test raw number crunching, some graphics capability, or photo rendering, but a composite number, like with fertilizer, could solve this. A cpu labeled 100-200-100 versus one labeled 200-200-100 would tell you that the first chip was half as fast on test A as the second chip, but the same on test B, and C. All the marketing mumbo jumbo and naming scams tell you nothing.
    Registered Linux user #526930

  2. #2
    Administrator jayd512's Avatar
    Join Date
    Feb 2008
    Location
    Kentucky
    Posts
    5,023
    I agree with you all the way.
    Two words for you though...
    1. Marketing
    2. Marketing

    The rest is true benchmark testing that most consumers aren't going to bother to perform.
    Jay

    New users, read this first.
    New Member FAQ
    Registered Linux User #463940
    I do not respond to private messages asking for Linux help. Please keep it on the public boards.

  3. #3
    Linux Newbie SL6-A1000's Avatar
    Join Date
    May 2011
    Location
    Australia
    Posts
    120
    Granted, some tests test raw number crunching, some graphics capability, or photo rendering, but a composite number, like with fertilizer, could solve this.
    Sorry to trash a thread with nonsense, but at first i thought this said you could test these things with fertilizer... I must admit i had to read it twice before it made sense.

    Anyway being slightly more serious, It is confusing i agree, but i think like Jayd said its marketing but also it has a large part to with the fact that they are multicore processors and most general consumers aren't going to be interested in the mathematical equation or true performance details of a CPU that as multiple factors (i.e. cores, TDP, etc...) that determine its true potential unlike single-cored CPU's, as most probably won't understand what it means.

    Most people only get that its a 2.4Ghz processor with X cores. So why present the technical information when they can present the lamen information that is only partially accurate but easier to understand.

  4. #4
    Linux Engineer MASONTX's Avatar
    Join Date
    May 2010
    Location
    Mason Texas
    Posts
    934
    Whether most people understand it or not, it would be nice to have the information. Most people don't read the nutrition and ingredients list on food either, but for those that care the information gives them the ability to make informed choices. Too bad cpu manufacturers don't do the same.
    Registered Linux user #526930

  5. #5
    Linux Guru Rubberman's Avatar
    Join Date
    Apr 2009
    Location
    I can be found either 40 miles west of Chicago, in Chicago, or in a galaxy far, far away.
    Posts
    11,452
    Some systems at 2.4GHz will outperform others at 3.0GHz on single-threaded tests, if the 2.4GHz system has a faster memory bus and most operations are not in cache. There are a HUGE number of factors that go into system performance, only a few of which are RAW CPU issues... So, as the sage once said, "test, test, test" to see if a particular system will be better for your needs than another! My current system is an 8 core 3GHz E5450 Penryn system (4 years old), and ONLY new systems with faster buses will beat it, if they are also 3GHZ or faster! Newer != Better all of the time!
    Sometimes, real fast is almost as good as real time.
    Just remember, Semper Gumbi - always be flexible!

  6. #6
    Linux Newbie SL6-A1000's Avatar
    Join Date
    May 2011
    Location
    Australia
    Posts
    120
    Quote Originally Posted by MASONTX View Post
    Whether most people understand it or not, it would be nice to have the information. Most people don't read the nutrition and ingredients list on food either, but for those that care the information gives them the ability to make informed choices. Too bad cpu manufacturers don't do the same.
    I totally agree with you! But i don't think it will happen, unless they decide to change how they market it them.

    Quote Originally Posted by Rubberman View Post
    Some systems at 2.4GHz will outperform others at 3.0GHz on single-threaded tests, if the 2.4GHz system has a faster memory bus and most operations are not in cache. There are a HUGE number of factors that go into system performance, only a few of which are RAW CPU issues... So, as the sage once said, "test, test, test" to see if a particular system will be better for your needs than another! My current system is an 8 core 3GHz E5450 Penryn system (4 years old), and ONLY new systems with faster buses will beat it, if they are also 3GHZ or faster! Newer != Better all of the time!
    Yeah that is true most of the time. With that said though, from some reading about the new AMD Opteron 6200's interlagos chipset, there performance compared to the 6100's mangy-cours chipset is minimal if non-existent; if comparing a 12-core 6100 to a 12-core 6200. The only real jump in performance is going from a 12-core to a 16-core.

    Just looking at this shows it:
    SPECfp®_rate2006 Performance and Processor Price

    Not going to start a intel or AMD flame war, because i am using an AMD CPU currently. But I think intel actually does a better job of improving the performance of their CPU from one generation to the next. The new E7 processors a quite frankly insane, a 10-core cpu with 10 threads at 2.8GHz (Turbo to 3.4GHz). The only other architecture i know that can currently do better is a SPARC CPU.

  7. #7
    Linux Engineer MASONTX's Avatar
    Join Date
    May 2010
    Location
    Mason Texas
    Posts
    934
    I know they wont change, unfortunately, but I get frustrated. When I go to a site I'll see a dozen computers with different cpu's, and no practical way to quickly tell how they compare to one another. Benchmarking is great, if you have the time, the equipment, and can compare the computers side by side, but on line you can't do that. Even google lets you down when you try and get real comparisons of different cpu's. The one's you see on a site are only partly represented on line. The latest haven't been tested yet, or older cpu's are no longer listed, or different testers used different benchmark tests.
    Registered Linux user #526930

  8. #8
    Guest
    Join Date
    Feb 2005
    Posts
    312
    Quote Originally Posted by MASONTX View Post
    I know they wont change, unfortunately, but I get frustrated. When I go to a site I'll see a dozen computers with different cpu's, and no practical way to quickly tell how they compare to one another.
    That's the whole intent.

    //edit:

    Ok to expand a little...

    I started building my own PCs back in the early 90s. Back then it was ok to have a CPU called, e.g. a 386 or a 486, the next model was to be the 586 and so on... back then not much in the way of marketing went into hardware as it was assumed that one knew what one was doing...

    In the days of yore a motherboard with black ISA slots, white PCI slots a white or beige CPU socket and a mess of jumper configuration (or if you were lucky dip switches and a manual with instructions on how to set them!) was acceptable... as were those two AT power connectors which blew the board up if you plugged them in the wrong way around...

    Your parallel port was black, your PS2 ports (if you had them) were black, your serial ports were also black. Your memory slots were either white, black or brown/beige...

    If for example you got stuck and were trying to plug your CPU into the parallel port... a quick glance at the manual (with the page the right way up) would quickly reveal your error. If you didn't read the manual and plugged your new CPU in without setting the correct jumpers, crossed your fingers and powered up, you'd get a nice pop and the smell of burning silicon. The idea there being that next time you'd read the manual...

    Back in those days if Jimmy had a 133MHz CPU and Johnny had a 200MHz CPU it was pretty obvious as to who had the better CPU... it was even clearly printed on the CPU somewhere.

    Then one day the god of PC hardware lay down with the goddess of Lego(r) and from his loins was sprung the case modding, cooling modding, cold cathode sporting hardware fanboi...

    As a result we had colour coded motherboard connectors... we had fancy indecipherable product names for CPUs... we had video cards, which spend their lifetime buried inside a computer case, designed to look like something out of a Star Trek film... We also had the reappearance of what I like to call "bastardised" hardware...

    This is the common scenario where fanboi A goes to cooloverclockingandmoddingfanboiforums.org and tells everyone there about his l33t new graphics hardware. He is quickly shot down by fanbois B, C and D who inform him that the GTXZ 9000 SE-2+ Special Edition is actually a standard GTXZ 8999 SE which is in turn based on the GTXZ 7000 from five years ago..... but that's not all. The board actually ships with crippled RAM... and it doesn't end there - 34 1/2 of the 36 pipelines have been turned off... if only he'd bought the GTXZ 9001 SE-2+ Special Edition - which is a real 9xxx series... oh dear...

    In other words - confusing naming, to confuse the already confused consumer.

    No fanbois were harmed during the making of this post.
    Last edited by cynwulf; 02-14-2012 at 03:37 PM.
    PikeMinnow likes this.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •