Solved

How come P4 is faster than i5 2500K !!!

Hello,
i'm optimizing a program that analyzes the market (2 yrs of data split over seconds)

Core to Core performance

i have p4 (with HT) -3.2 Ghz- computer with 512 mb ram .. it's solves 1 calculation every 15 min

the irony thing , my new one i5-2500k -3.3 Ghz- with 8 Gb RAM ... solves only 1 calculation every 42 min !!!!!!!

what is wrong ?!
12 answers Last reply Best Answer
More about faster 2500k
  1. you could try ocing ur ram in bios to its stock settings??

    "dont quote me on that im no expert "
  2. I'd guess this is a programming issue not a hardware one. There is no reason the P4 would be faster.
  3. Best answer
    My guess is that the developer of this program put it together with specific optimizations for the P4 (family 7 CPU). When Intel released the C2D, they went back to family 6 (and stayed there) which may mean that newer instructions sets introduced on the P4 are not being used. This way of detecting processor features is completely wrong and Intel makes this blatantly obvious if the programmer ever glanced at an x86 optimization guide from Intel.
  4. Are you running the program in the same OS and bit version (32/64)?

    I think wolf has it right. Programming/compatability issues with your software.
  5. If you're doing some kind of assembly programming and using some obscure SSE instruction then yeah I can see that. What compiler are you using? If you're using an older version of Visual Studio do you have the latest patches?
  6. aicom said:
    My guess is that the developer of this program put it together with specific optimizations for the P4 (family 7 CPU). When Intel released the C2D, they went back to family 6 (and stayed there) which may mean that newer instructions sets introduced on the P4 are not being used. This way of detecting processor features is completely wrong and Intel makes this blatantly obvious if the programmer ever glanced at an x86 optimization guide from Intel.



    +1 :bounce:
  7. aicom said:
    My guess is that the developer of this program put it together with specific optimizations for the P4 (family 7 CPU). When Intel released the C2D, they went back to family 6 (and stayed there) which may mean that newer instructions sets introduced on the P4 are not being used. This way of detecting processor features is completely wrong and Intel makes this blatantly obvious if the programmer ever glanced at an x86 optimization guide from Intel.


    Good to know. The things you learn in these forums is amazing.
  8. Without knowing your algorithm, language, and compiler there isn't much anyone can tell you. You may want to go to a programming forum though as this certainly sounds like an issue with your code.
  9. Try running benchmarks for each. This'll give you a good indication of which is actually faster.

    Download Geekbench - IMO it's the best basic benchmark tool out there and it's free!
  10. kaimbanks said:
    Try running benchmarks for each. This'll give you a good indication of which is actually faster.

    Download Geekbench - IMO it's the best basic benchmark tool out there and it's free!


    Never heard of this before, pretty neat! I got a 8552 :)
  11. Best answer selected by DrGM1.
  12. DrGM1 said:
    Hello,
    i'm optimizing a program that analyzes the market (2 yrs of data split over seconds)

    Core to Core performance

    i have p4 (with HT) -3.2 Ghz- computer with 512 mb ram .. it's solves 1 calculation every 15 min

    the irony thing , my new one i5-2500k -3.3 Ghz- with 8 Gb RAM ... solves only 1 calculation every 42 min !!!!!!!

    what is wrong ?!



    thanks guys ... i'll recheck the code
    plus i think the difference here is mainly either because of System 64bit or the HDD is 5200 RPM which is SLOW !
    thanks again :)
Ask a new question

Read More

Core RAM Intel i5 Overclocking