E2160 /E2140 VS E4300 - Some benchmarks

Discussion in 'Intel x86 CPUs and chipsets' started by mAJORD, May 20, 2007.

  1. mAJORD

    mAJORD Member

    Joined:
    Jun 4, 2002
    Messages:
    9,402
    Location:
    Griffin , Brisbane
    Been a couple of requests around the place for benchmarks on these somewhat mysterious CPUs that have crept onto market early, and without a peep from intel, so seeing as I have an E4300 and E2160 in my hands, why not compare them on an identical platform on apps.


    [​IMG]

    .

    [​IMG]

    Return of the pentium.... I guess?

    I don't know about you guys, but this has to be the most bland product boxes i've seen in a long time.. I have to wonder, do intel even know where they want to place this range in the market? I think not. Apart from the model and batch # sticker, there is not one thing on the box to actually tell you what it is.. Is it a pentium 'D' - No. Is it a Pentium E? , well I thought it was, but apparently not. Is it a Core 2.. well, yeah but no but yeah but.. aparently it's not that either. I personaly don't know why they didn't just add it to the core 2 series, but anyway...

    Something else I found interesting is the cooler - No copper slug anymore.. Not sure if this is a trend with new C2Ds in general (???) or not.


    [​IMG]

    Existing benchmarks on the web compare game performance ( http://www.pconline.com.cn/diy/cpu/reviews/0704/993288.html ), showing a very small impact on FPS @ stock, turning into a large one at 3gig+ due to the lack of cache starving the core so to speak. So i've concentrated on some more every day things like encoding, photoshop, Word, compresison. Plus a few oddities i've been working on

    Before I go into the OCing woes I had with this CPU, here are a collection of benches @ stock.

    Conditons:

    All benchmarks performed at least twice, 3 times if there was any (small) difference. E4300 was tested first on a freshly started machine.

    E2160 Fitted straight after, bios settings checked, and all tests performed in exactly the same order to maintain exact consistency between tests.. confident these results are spot on.

    I've thrown some previous stock X2 3600+ results in aswell, some are missing unfortunatly. I may update with extra results when I hook that machine back up.

    Test machine:
    Intel
    CPUs:
    E2160 1.8gig (200x9)
    E2140 - Simulated 1.6gig (200x8 in bios)

    Mainboard: Gigabyte GA-965P-S3 F9
    Ram: ADATA DDR2-800 @ SPD (5-5-5-18) @ 2:1
    Hard drive: Seagate 80gig SATA 7200RPM
    OS: Win XP SP2, Febuary updates, DualCore hotfixed.

    AMD
    CPU:AMD X2 3600+ 1.9gig
    Mainboard: DFI LANPARTY UT NF590 SLI-M2R/G
    RAM: ADATA DDR-800 @ SPD (5-5-5-18)
    Hard drive: Seagate 80gig SATA 7200RPM
    OS: Win XP SP2, Febuary updates, DualCore hotfixed , AMD patch



    I won't go into software detials as it's more about comparisons, the main ones are:

    photoshop CS2 -Retouch artists speedtest macro.
    Word 2007
    Lame 3.97 intel compiler
    CDEX 1.7 (flac)
    Cinebench 9.5
    Windows media encoder 9

    As mentioned, i've included some new things I've found to be quite cpu intensive (over 80% on 1 core), but often ignored. Scrolling through a large PDF file full of images, and waiting for the acrobat installation to process both turn out to be very cpu intensive, but don't fit into any benchmark catagory i know of. Regardless, i've been including them in a few reuslts i've posted up in the past, and will be looking for more 'often overlooked' things if anyone has Ideas..

    **A note on some of the benchmark titles.

    I realised some don't make a lot of sense, the titles came striaght out of my spreadsheet and i overlooked the names. so a bit of an explaination:

    "Compress wordbench": RAR compression of a 200meg word file. lot's of images.
    "FLAC": compression of 200+meg WAV file in FLAC lossless format\
    "WMencoder" : windows media encoder 9, encoding a FRAPS BF2 avi file to wmv



    Results @ stock:


    [​IMG]

    Overclocking.

    Unfortunatly, I seem to have scored a rather poor chip. Most likely suffering from the 300FSB wall syndrome I hear about on 4300s

    I'm not 100% convinced that's the cause still, as unlike the 4300, I ran into stablity problems on stock voltage at a very low 2.6Ghz (288x9) , which went rapidly downhill approaching/ passing FSB 300.

    This was a bit of a shame as I was hoping to post results @ 3gig.

    Anyway, i had to settle on 2.7 @ 1.5v in the end.. on the stock cooler which is a bit risky, but it was only for benching, no time for serious heat to set in.

    comparions @ 2.7 (300x9 1:1 Ram:)

    [​IMG]



    My conclusion:


    I think the results speak for themselves prettymuch. At this point in time, comparing these reuslts with those on the net showing gaming performance, it's pretty apparent that @ stock speeds, or even mildly Overclocked, the Effects of the 1mb cache is almost non existent for all but a few things.

    So at the least, as a stock CPU the 2160 is great value, matching that of X2 4200+ which i'd put it up against myself (as it was only a tiny bit slower than an E4300 so this CPU would bring it about par). Those putting together stock systems and considering the E4300 may want to think otherwise if on a budget.

    The E2140 on the other hand is a mixed bag @ stock, the Mhz hit is quite severe in the benches I ran, and it was trading blows with the X2 3600+ in most cases. It's consistent edge in encoding and gaming would push it slightly ahead overall, but considering the 3600+ can be had for as little as $90-95, and can be partnered with very cheap, bust solid boards I still think it's better value.

    As for Overclocking.. well I'm not making any judgment until others have posted their results. This CPU was an absoulte pig over 2.6gig to be honest. Despite getting it to run at 2.7, it was far from day to day friendly, refusing to boot sometimes and resetting back to stock. Slotting the E4300 back in everything was fine, so I'm certain it's a CPU limitation in this case. Others on XS have reported 3gig plus overclocks, similar to E4300, so I may have just gotten very unlucky. However, it should be remembered results i've seen so far show the effect of the cache loss will show up at higher clocks than I was able to achieve, in things like games. Worth keeping in mind

    For those wishing to avoid this batch:

    Q648A415

    Providing MOST 2140's prove to hit 3gig, or at least closer to, (unlike this example), even as an AMD man I'd have to say it would win the OC value award. Despite the slightly higher cost, and need for a decent mainboard, one of these @ 3gig is still going to be a good 15% faster than a K8 (see earlier link) Time will tell if that's as achievable (or more) as it is with an E4300

    Cheers
     
    Last edited: May 20, 2007
  2. RodneyJM

    RodneyJM (Banned or Deleted)

    Joined:
    Mar 20, 2005
    Messages:
    1,916
    Location:
    Engaged in Silicon Wars !
    Interesting stuff mAJORD, got a E2160 coming early this week. Going to bang it on to my P5B-E and let her rip...:thumbup:
     
  3. skinz

    skinz Member

    Joined:
    Feb 26, 2002
    Messages:
    1,161
    Location:
    Sydney - 2565
    just wondering, did you try the e2160 with the e4300's heatsink? i doubt it would make much difference, but i was wondering just out of interest if it effected the overclock. what kind of case cooling did you use? or was it an out of case setup?

    I am currently wondering whether or not to grab one of these, a e4300, or a x2 4600+. They work out to be around the same price. The 4600+ is faster at stock, but the other two are better overclockers. Just wanting something that is as fast or faster than the 4600+, say 2.4ghz at stock volts and with stock cooling. I'd imagine that is feasible with either of the Intels on a 965 board? 2.7ghz aint too bad either, would that make it quicker than an e6600/6000+?

    Thanks again, i've been waiting for these tests.
     
  4. Onyx

    Onyx Member

    Joined:
    Jun 29, 2001
    Messages:
    4,253
    Location:
    NSW
    Thanks for going to the trouble of benching/comparing the two chips. What I'd like to see is both these two chips clocked at 266FSB and compared to an E6600 - to see the difference that L2 cache makes - at 1MB, 2MB and 4MB amounts with processor speeds being the same.

    My decision to buy an E4300 was based upon research that showed cache to matter very little. IF and it's a big if the E2160s were to clock as well as the Allendales, this would have been an even better value proposition.
     
  5. slobber

    slobber Member

    Joined:
    May 1, 2005
    Messages:
    5,207
    Location:
    Mandurah!
    Thanks so much for you time and effort mAJORD :thumbup: :thumbup: :thumbup:

    I'm sure many ppl will benefit from your advice

    I personally would still go for the AM2 solution in this price bracket simply due to available boards/feature set at a given price as Intel boards are way overpriced if you want one that is OC friendly though the price is going down
     
  6. Narcoticx2

    Narcoticx2 Member

    Joined:
    Mar 5, 2003
    Messages:
    4,339
    Location:
    The Redlands
    thanks, mAJORD.

    While they're cheap, i certainly don't want to be left in the cold regarding overclocking. When the slightly more expensive ones get to around 3GHz, you'd expect these to!

    Looking forward to seeing some more people's results - I think pairing this processor with a DS3 is a pretty cheap C2D setup, but it's only really worth it if it'll overclock to the 3+ mark! Otherwise, better off with a 3GHz 3600+...
     
  7. slobber

    slobber Member

    Joined:
    May 1, 2005
    Messages:
    5,207
    Location:
    Mandurah!
    Thankyou for your invaluable contribution :rolleyes:
     
  8. Knobi

    Knobi Member

    Joined:
    Mar 21, 2005
    Messages:
    2,382
    Location:
    Eeede-laide
    majord, your chip seems to be a dawg.............

    I am contemplating the same one or a brissy 3600+
     
  9. DAVID CLEM

    DAVID CLEM Member

    Joined:
    Apr 18, 2003
    Messages:
    372
    Great to hear about the E2160, been waiting for a 775 chip around this price for the kids PC.
     
  10. N1zmo

    N1zmo Member

    Joined:
    Apr 25, 2005
    Messages:
    608
    Location:
    Perth, 6065
    Will these be replacing the pentium D or celeron Ds?
     
  11. ricky60

    ricky60 Member

    Joined:
    Jun 14, 2003
    Messages:
    588
    These chips will replace the pentium D's (in some regards, already have, because supply of the PD 3.0GHz is already drying up since last thursday). The Celerons are being replaced with the Celeron 420's and the Celeron 440's, which already should be hitting the shelves. The Cele's are only apparantly using 35Watts, which is pretty cool for HTPC applications.
     
  12. SLATYE

    SLATYE SLATYE, not SLAYTE

    Joined:
    Nov 11, 2002
    Messages:
    26,823
    Location:
    Canberra
    It would seem that NetBurst is finally dead, if the Celerons are being replaced too. What'd be really nice is if they'd release a version of this new Pentium which works on the old P4/P-D boards.

    mAJORD, are the heatsinks the same weight? The new one might still have a copper slug, but with a cover over it (to prevent corrosion).
     
  13. OP
    OP
    mAJORD

    mAJORD Member

    Joined:
    Jun 4, 2002
    Messages:
    9,402
    Location:
    Griffin , Brisbane
    Nah SLATYE, it's considerably lighter.

    The Copper seems to be hollowed out towards the top aswell, whereas the alum' appears to just be solid all the way up, from what I can see
     
  14. chainbolt

    chainbolt Member

    Joined:
    Jun 27, 2001
    Messages:
    102,967
    Location:
    Tokyo
    Yep, that is really great work. A real performance comparison. :thumbup:

    The E2140/2160 and later 2180 are Allendales.
     
    Last edited: May 24, 2007
  15. slobber

    slobber Member

    Joined:
    May 1, 2005
    Messages:
    5,207
    Location:
    Mandurah!
    Yup.............with half the L2 disabled, just like they used to do with the Conroes

    Hmmm I wonder if we will see a 'native' e2*** series :)
     
  16. chainbolt

    chainbolt Member

    Joined:
    Jun 27, 2001
    Messages:
    102,967
    Location:
    Tokyo
    Native Conroe with 4 MB and native Allendale with 2 MB. To disable 50% of the L2 is very cost effective. I doubt they will come with a native E21xx.
     
  17. slobber

    slobber Member

    Joined:
    May 1, 2005
    Messages:
    5,207
    Location:
    Mandurah!
    Oh OK, I thought they might as then more usable cores per wafer but another 'line' would be needed :)
     

Share This Page