UNSOLVED

jd10013

updated

20 years ago

J

jd10013

139 Posts

0

10755

November 6th, 2006 20:00

which is better?

I'm thnking of buying a new computer, and I was wondering about dual core processors.  I know the basic concept of how they work, but comparing them seems to be a bit tougher than with single core processors.  I have 2 question.
1.  Why are the speeds lower on the dual core's?  2.6ghz seems to be the top end, and many are 1.8ghz.
2.  which would be better.  a slower dual core with a higher fsb, or a faster dual core with a slower fsb.  to be more specific, I'm trying to compre the (and this is all rounded) 1.8 ghz dual core with the 1000 fsb to the 2.6 ghz dual core with the 533 fsb?
 
The slower processor with the faster fsb is more expensive.  is overall speed less important with the dual core's?
  • PETER345

    1 Rookie

    •

    5796 Posts

    579

    0

    Posted November 6th, 2006 21:00

    You need to consider the architecture. Considering Intel dual cores:

    Pentium D: based on Pentium 4 architecture
    Core 2 Duo: based on the Core/Pentium M architecture

    The Pentium D/4 are fairly inefficient and use a lot of power to get less processing power per GHz. The Core/C2D are MUCH better.

    I have both a 2.4G C2D and 3.0G PD. The C2D blows the PD out of the water. It (2.6G C2D) performs (running hours of regression tests) like a PD at 5.3GHz! Not only that, but it burns 100W less power than the PD at full load. So it is much faster, more efficient and cooler.

    I would not buy PD unless you just can't afford the C2D. By my comparision (C2D is equivalent to PD at 2.2x the clock), the 1.8G C2D is equivalent to PD at about 4G! You get a faster FSB to boot.

    My figures are pretty close to what I have seen in the published benchmarks.

    Of course, there are also the AMD dual cores. In general, they are better than the PD, but not as good as the C2D. Some AMD fans may or may not agree with the last statement.

    Hope this helps,

    Peter
  • jd10013

    139 Posts

    579

    0

    Posted November 6th, 2006 22:00

    ahhh, your right.  the 2.6 with the 533mhz fsb is a PD, and the 1.85 ghz with the 1066 mhz fsb is a core 2D.  Since you seem very knowldgeable about these things, how would you rate these specs:
     
     
    Intel® Core™2 Duo Processor E6300 (1.86GHz, 1066 FSB)
    Genuine Windows® XP Media Center 2005 Edition  
    Express Upgrade to Windows Vista
    1GB Dual Channel DDR2 SDRAM at 533MHz- 2DIMMs
    160GB Serial ATA Hard Drive (7200RPM) w/DataBurst Cache™
    16x DVD+/-RW  Drive
    No Monitor
    256MB nVidia Geforce 7300LE TurboCache
    Integrated 7.1 Channel Audio

     

    or more specificly, i'm currently using a dell dimension 8400 with a p4 3.2 ghz ht with 1 gig of ram, dual drives (cd-rw, dvd), sound blaster audio card and a 256mb pci-e ati radeon x700.

    would the above system be a noticable improvement over what i have?  like i said, i have a lot harder time comparing the new processors to the old ones, and haven't bought vid cards recently enough to know what out there is good, and whats bad.

    Message Edited by jd10013 on 11-06-2006 06:22 PM

  • PETER345

    1 Rookie

    •

    5796 Posts

    579

    0

    Posted November 7th, 2006 00:00

    My rule of thumb is that the C2D behaves like PD at 2.2x the clock. Your mileage may vary.

    So, the 1.86 C2D is like 4G PD. That might not be a huge improvement over what you have (roughly 25% faster).

    I don't know much about the graphics cards as I don't play games and just buy the cheapest ones. I have the 7300LE and my understanding is that it is a "real" graphics cards with GPU, but on the low end.

    If you want noticeable improvement I would go more for the E6600 (2.4G). The E6700 is just too expensive to justify the 10% improvement over the E6600 (unless money is no object). Otherwise, you might wait a bit for the prices on the high end to come down.

    I generally find that for just everyday use, 50% improvement is needed to be noticeable. However, for specific tasks (like my software testing or gaming), you might find less improvement to be noticeable.

    Peter
  • CmA

    1 Rookie

    •

    154 Posts

    579

    0

    Posted November 7th, 2006 11:00

    Does that 2.2x apply to single threaded apps or only ones that take full advantage of the two cores?
     
    Thanks,
  • PETER345

    1 Rookie

    •

    5796 Posts

    579

    0

    Posted November 7th, 2006 13:00

    I suspect it applies to both. However, the C2D has a bigger L2 cache so it should handle 2 apps at once better than the PD. Google for benchmarks. There are many out there.

    The actual advantage depends on many factors, but the benchmarks show the C2D really pretty much blows the PD away, it is just a question of "by how much?"

    I was amazed how my Core Duo (not Core 2) at 1.6G beats my PD 3G desktop. The laptop burns 30W at full bore compared to the desktop at 235W. Talk about a better design!

    Peter
  • CmA

    1 Rookie

    •

    154 Posts

    579

    0

    Posted November 7th, 2006 14:00

    Thanks Peter ... I've looked at LOTS of benchmarks and reviews, but was never 100% confident of how the single threaded apps would perform, or what their single threaded apps were actually comparing. I'd need to investigate more the benchmarks themselves, and that was overkill for me!!
     
    Regards,
  • CmA

    1 Rookie

    •

    154 Posts

    579

    0

    Posted November 7th, 2006 15:00

    Impressive ... Thanks for that ...
  • PETER345

    1 Rookie

    •

    5796 Posts

    579

    0

    Posted November 7th, 2006 15:00

    I think you can be pretty confident on this one. Everyone says C2D is better than pretty much everything else. Even the AMD guys are not saying otherwise.

    If you remember when Pentium 4 came out and it turned out the 1.5G P4 was about the same speed as the 1.1G P3. It is well known that the P4 does less work per clock than virtually every other architecture. C2D has many improvements over the P3.

    I just ran a single single-threaded CPU intensive app that took 618s on 3G PD (9100). It took 331s on 2.4G C2D (9200). That says that the C2D is 2.33 times the effectiveness of PD at the same clock. Hard to believe isn't it? Either C2D is that good or P4/PD is that bad (or both).

    Peter