16

For an Intel based system keeping the number of cores constant how much difference would there be between two dual core processors and one quad core processor? We run Oracle OLTP databases on two dual core processors. With fewer dual core processors available we may need to switch to quad core, but we are limited to four cores due to licensing issues, so will have to switch to one processor. If we make the switch should I expect to see a change in real world processing speed?

Improve this question
3
  • Interesting to know, I would've thought they'd changed that model to per socket by know.
    – Oskar Duveborn
    May 6, 2009 at 19:35
  • Depending on the platform the licensing is .5*NumberOfCores which generally allows Oracle to charge more than per socket, so I don't expect it to change.
    – Leigh Riffel
    May 6, 2009 at 21:21
  • 2
    In regards to the application performance of Oracle on 2x dual-core vs 1x quad-core, Jeff is likely correct that there wouldn't be a significant performance difference. However, Chopper3 is also correct in the general sense that the on-die transports and shared cache on the quad core provides a higher theoretical performance than the 2x dual-core configuration. Depending on the specific use case, the outcome could be different (say, heavy matrix computation). If Jeff's answer is the accepted one, I suggest that the title of the question be updated to be Oracle specific.
    – Burly
    May 8, 2009 at 17:16

3 Answers 3

Reset to default
13

Should be very little difference in performance between two dual core CPUs and a single quad core CPU, if you are using current Intel Xeons.

However, you might expect to see slightly better performance from the quad-core CPU if it is a "true" Quad core design, ala the new AMD Opterons or the new Xeon W5xxx (aka the Core i7 / Nehalem) series.

http://techreport.com/articles.x/16656/12

The Nehalem Xeons' truly astounding leap in performance over prior generations, in a range of applications, speaks for itself. The largest gains came in our scientific computing/HPC tests, where the Xeon W5580 proved to be between 50% and 100% faster than the Harpertown Xeon X5492. We saw a massive performance increase in SPECjbb 2005, as well, along with more modest but still substantial improvements everywhere else.

Improve this answer
2
  • 1
    That's more because of the better memory interconnects on the Nehalem EP series than better core separation within the chip, though.
    – Richard Gadsden
    May 5, 2009 at 12:09
  • 3
    I couldn't disagree more sorry, off-die latency (even over a 6.8GT QPI link) is 4x slower than on-die - and if it's sharing L1/L2 cache this could be several times faster again. Also 55xx-series Xeons make this even more accute.
    – Chopper3
    May 6, 2009 at 21:51
5

Check the max memory you can put in each system - it's possible that you may be able to stuff more RAM into a 2xdualcore box than you can into a 1xquadcore box. And more RAM will likely help.

Improve this answer
0
5

Sorry for the delay spotting this question. There is a huge difference in performance between a dual-processor, dual-core setup and a single-processor, quad-core setup - the latter will be much quicker whichever make and model you go for, the main reason is because everything happens 'on-die' (although there's an argument to be had if you really want to get into it), so latency and overall bandwidth is much better. This is especially true on a true quad-core chip such as the Xeon 55xx-series.

Improve this answer
1
  • But the dual-processor setup may have more CPU cache in total, that sometimes may be inportant.
    – Ian Ringrose
    May 27, 2010 at 9:19

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .