i5 vs i7 rmbp13/2015

Discussion in 'MacBook Pro' started by terencedriver, May 5, 2015.

  1. terencedriver, May 5, 2015
    Last edited: May 5, 2015

    terencedriver macrumors member

    Joined:
    Mar 16, 2015
    #1
    Why is there so much advice against an i7 configuration vs an i5 for the retina macbook pro 13 2015?

    Surely the extra level3 cache is beneficisl to developers and design work?

    Is the argument purely value for money ie 5% geekbench improvement for £200 ?
    Does battery life suffer much under low intensity usage (web email etc)?

    UPDATE- to be clear .. My query is about which scenarios benefit from an i7. I'm interested in rationale and evidence vs 'you won't notice it'

    Update 2: working with several (3-5) virtual machines is a scenario. Clearly ram helps but what is the effect of i7 over i5?
     
  2. yjchua95 macrumors 604

    Joined:
    Apr 23, 2011
    Location:
    GVA, KUL, MEL (current), ZQN
    #2
    The extra L3 cache doesn't mean anything.

    In fact, the performance delta between the 2.9GHz i5 and 3.1GHz i7 is <5%.

    Battery life doesn't suffer since they both have the same TDP.

    Ask yourself this - is it worth spending $200 for an upgrade just for the sake of 5% extra performance (and also the capability to brag to others that you can have an i7)?
     
  3. Samuelsan2001 macrumors 603

    Joined:
    Oct 24, 2013
    #3
    In order

    1. Because it is not very good value for money.

    2 Yes if you are maxing out both cores for longs periods of time, but at that point a 15 inch with quad core will be a much better upgrade.

    3. Yes this is the argument it really makes very little difference for the cash.

    4. Battery life will be just about the same.
     
  4. Barna Biro macrumors 6502a

    Barna Biro

    Joined:
    Sep 25, 2011
    Location:
    Luzern, Switzerland
    #4
    Realistically speaking, for 99% of the users, it would make no difference. Personally, where I live, the other stuff I want for the notebook already pushes it into the ~2800$ mark, so that additional ~200$ for the CPU is a no brainer... it makes me sleep better at night too :D

    You'll most likely be more than happy with the i5 and you'll have 200$ to spend on something else.
     
  5. terencedriver thread starter macrumors member

    Joined:
    Mar 16, 2015
    #5
    Really? I'm genuinely interested in this. Surely it had benefit but in which scenarios? Does it mitigate the infamous Yosemite UI lag? Does it speed compiling code?

    As a scientist I can't believe it had no effect at all.

    My OP was about which scenarios benefit from the i7 .. Sorry, should have been clearer.
     
  6. yjchua95 macrumors 604

    Joined:
    Apr 23, 2011
    Location:
    GVA, KUL, MEL (current), ZQN
    #6
    I'm a software engineer. There is no scenario that will benefit from the dual core i7, because it's identical to the dual core i5 aside from the slightly higher clock speed and slightly larger cache. The UI lag is due to poorly written GPU drivers, not because of the processor. Even I experience frame drops sometimes with my 15" (late-2013 2.6/16/1TB/750M).

    My personal 13" sometimes experiences lag on the internal retina display, but the moment I put it into clamshell mode and stick it into an external 1080p display, all lag disappears. For the record, it's a late-2013 i7/16/512 13".

    If it was a quad core i7, then hyper threaded tasks like video editing and rendering, media production and audio production would greatly benefit from it. But since you're comparing a dual core i7 and a dual core i5, nothing would benefit from it.

    Compiling code is a very trivial task for the most part. Even a tiny Intel Atom can do that.

    In where I work (VMware), a lot of my colleagues compile code on just sub-2GHz i5s (like the MacBook Air processors).
     
  7. terencedriver thread starter macrumors member

    Joined:
    Mar 16, 2015
    #7
    I'm reading in CPU tech sites that for i7 vs i5:

    * l3 cache improves power consumption and overall speed

    * extra i7 instructions for some code eg multimedia (codec?) and vector operations

    * hyper threading allows code to use unused capacity effectively improving throughout by approx 30%

    Are these not benefits? Feel free to tell me why these are wrong.
     
  8. yjchua95, May 5, 2015
    Last edited: May 5, 2015

    yjchua95 macrumors 604

    Joined:
    Apr 23, 2011
    Location:
    GVA, KUL, MEL (current), ZQN
    #8
    Hyper threading differences is only valid when core count is the same but thread count is different. In the case of the 13", both i5s and i7s have 2 cores and 4 threads, so both have hyper threading. So the stuff you posted are all invalid.

    In desktops, the quad core i5 has 4 cores and 4 threads, while the quad core i7 has 4 cores and 8 threads. So in this case, the i5 lacks hyper threading and the i7 has hyper threading. Hyper threading only exists when the thread count is twice the core count.

    All processors in an SKU share the same instructions set, be it i5 or i7.

    For instance, the i5 and i7 in the 13" are all Broadwell-U, so they have the same instructions set as all processors in the Broadwell SKU (be it Broadwell-Y, Broadwell-H, Broadwell-K....etc.)

    Difference between non-hyperthreaded i5 and hyper threaded i7 (both quad core): http://browser.primatelabs.com/geekbench3/compare/591026?baseline=611544

    The i7 here is a full 30% faster in multicore (i.e. hyper threaded) benchmarks.

    Difference between hyper threaded i5 and i7 (both dual core):
    http://browser.primatelabs.com/geekbench3/compare/2173902?baseline=2170758

    The i7 here is only 2.3% faster than the i5 in multicore (i.e. hyper threaded) benchmarks. So is this 2.3% worth an extra $200? Let's put it in this context - encoding a video. If it takes half an hour to encode a video on the i5, it'll only be 36 seconds faster on the i7. At this stage, a difference of 36 seconds don't matter. Still worth the $200?
     
  9. T5BRICK macrumors 604

    T5BRICK

    Joined:
    Aug 3, 2006
    Location:
    Oregon
    #9
    L2 cache is much more important than L3. 1MB additional L3 cache isn't going to cause a huge improvement in performance or noticeably reduced power consumption.

    As someone else already pointed out, this only applies if the i7 has a different instruction set than the i5 you're comparing it to. In the 13" rMBP, the i5 and i7 processors are basically identical except for the difference in cache, CPU speed.

    Again, that would make a difference if the i5 didn't also have hyperthreading.

    No, those aren't benefits because both the i5 and i7 have most of the same features. If you're interested in buying the i7 anyway, go ahead. From a performance per dollar perspective, it costs more than 10% extra for less than 5% increase in performance.

    For the record, I recently purchased a 13" rMBP i7/16GB/512GB. It's a great computer, but the only reason I went with the i7 is because it was the only configuration available at the Apple Store with 16GB of RAM.
     
  10. snaky69 macrumors 603

    Joined:
    Mar 14, 2008
    #10
    To keep things simple: unless you are pegging your CPU at 100%, 100% of the time, any difference you would notice would be entirely in your head.
     
  11. terencedriver thread starter macrumors member

    Joined:
    Mar 16, 2015
    #11
    Yjchua .. Wow!! That comparison of i5 and i7 rmbp13 is pretty damning.

    Really seems like £200 makes little difference.
     
  12. CaptainChunk macrumors 68020

    CaptainChunk

    Joined:
    Apr 16, 2008
    Location:
    Phoenix, AZ
    #12
    Yep. I ended up getting a custom 2.7GHz i5/16GB/256GB model because I found a lot more value in spending the extra money for 16GB of RAM than a marginally faster dual-core CPU.
     
  13. PaulWog Suspended

    Joined:
    Jun 28, 2011
    #13
    The speed bump is minimal. If someone needs a potential 30% speed bump that usually is in the 0-5% range (and 30% is more or less a theoretical scenario), then this isn't the laptop for them anyways. Other slightly larger laptops exist with much more powerful processors.

    What further puts such an upgrade into an additional category of niche-buyers is this: Not only do you need to be someone who wants an ultra-portable laptop that needs a very minimal speed boost, but you want an ultra-portable laptop that you'll be running basically at 100% CPU capacity (case and point: an ultraportable laptop that runs out of batteries in 2 hours flat is an ultraportable chunk of metal).

    The upgrade isn't a waste of you know you need the extra power. As in, a developer who already knows the nature of his work (the number of cores used). If you think you might eventually use it one day, it's a waste.
     
  14. terencedriver thread starter macrumors member

    Joined:
    Mar 16, 2015
    #14
    Is the benefit of 2.7 i5 and 2.9 i5 as minimal as the above discussed 2.9i5 vs 3.1i7 ?
     
  15. Samuelsan2001 macrumors 603

    Joined:
    Oct 24, 2013
    #15
    Yes

    And to be honest the Broadwell is not much better than the equivalent Haswell either so if you don't care about force trackpad and pointlessly fast SSD then you can save a load of cash buying a 2014 refurb.

    http://store.apple.com/us/browse/home/specialdeals/mac/macbook_pro/13
     
  16. 8tm macrumors member

    Joined:
    Mar 25, 2015
    #16
    I had the chance to own & compare both i5 & i7 for about 14 days. 2,9Ghz & 3,1Ghz, both 16/512GB.

    Did a lot of benchmarking, launching huge Logic audio sessions and my experience is, you just can't tell the difference between the two computers.

    Except for one thing : Heat !

    The 3,1GHz is muchhh more prompt to trigger the fans & i've seen huge temps differences compared to the 2,9 for the same tasks.

    Maybe my 3,1Ghz is a faulty unit, bad thermal pasting or whatever, but i ended up keeping it because of a much better screen than the 2,9.

    Just my experience, but could be something really. I'll advise you to go i5.
     
  17. T5BRICK macrumors 604

    T5BRICK

    Joined:
    Aug 3, 2006
    Location:
    Oregon
    #17
    I've noticed that my 3.1GHz i7 under full load runs quite a bit cooler than my old mid-2009 13" MBP with a 2.53GHz Core2 Duo, but that's not really a fair comparison.
     
  18. FtrV8 macrumors regular

    FtrV8

    Joined:
    Sep 16, 2009
    #18
    Great info in this thread. I am in the market for a new 13 inch and was set on the i7 config, but will save my $200!

    I do run a Windows VM via Parallels, that was my only reason to get the i7. RAM will be maxed for sure.

    Thanks for the great info all.
     
  19. terencedriver thread starter macrumors member

    Joined:
    Mar 16, 2015
    #19
    I read a few online reviews suggesting the fan is prone to spinning up on the i7 more than the i5
     
  20. yjchua95 macrumors 604

    Joined:
    Apr 23, 2011
    Location:
    GVA, KUL, MEL (current), ZQN
    #20
    I disagree, all i5s and i7s used in the 13" all have the same TDP, so the prejudice that the i7 generates more heat is inaccurate.
     
  21. terencedriver, May 7, 2015
    Last edited: May 7, 2015

    terencedriver thread starter macrumors member

    Joined:
    Mar 16, 2015
    #21
    But to keep within the TDP envelope it'll need more cooling (fans)... As it runs hotter because its base frequency is higher.

    Am I wrong?
     
  22. Samuelsan2001 macrumors 603

    Joined:
    Oct 24, 2013
    #22
    Certainly possible

    and while it may be hotter it will be for a shorter time as it'll get the job done quicker.
     
  23. newellj macrumors 601

    Joined:
    Oct 15, 2014
    Location:
    Boston, MA, US
    #23
    Design differences...six years of cruft inside...as you say, really not comparable. :)
     
  24. 8tm macrumors member

    Joined:
    Mar 25, 2015
    #24
    From my experience and a few (pro) online reviews, the i7 generate noticeably more heat than the i5. For a worthless performance boost. Don't worth it if you ask me.

    ----------

    The job is done faster, but only while exporting/rendering/compiling stuff. And you gain about 5/6 seconds.. It dont compensate for the extra heat.
     
  25. yjchua95 macrumors 604

    Joined:
    Apr 23, 2011
    Location:
    GVA, KUL, MEL (current), ZQN
    #25


    From my experience, it doesn't.

    My late-2013 13" rMBP (i7/16/512) is still running really cool (45ºC when doing light tasks like web browsing and using Word, and barely goes above 60ºC when doing some medium tasks).
     

Share This Page