Early M2 benchmarks show clear CPU, GPU performance gains over M1

Posted:
in macOS edited June 15
New benchmark results for what appears to be a 13-inch MacBook Pro with an M2 chip have surfaced, showing clear performance advantage over the previous-generation M1.

M2 benchmark
M2 benchmark


The Geekbench 5 results, which were published on June 15, cover a Mac with a model number of 14,7 running macOS 12.4. The Mac itself is equipped with 16GB of memory.

In single-core Geekbench 5 testing, the M2 chip received a score of 1919. The same chip got a multi-core Geekbench score of 8928. For comparison, a late 2020 MacBook Pro with an M1 chip received a single-core score of 1749 and a multi-core score of 7719.

In other words, the Apple M2 chip has a demonstrable boost in performance.

Like the M1, the M2 is an eight-core Apple Silicon system-on-chip (SoC) based on a 5-nanometer fabrication process. Apple itself says that the M2 chip's CPU is 18% faster than the M1.

Per the Geekbench 5 comparison, the M2 chip is 11.6% faster than the M1 in single-core scoring and 19.5% faster in multi-core scoring. Apple also says that the M2 provides 25% better performance at comparative power levels than the M1.

Additionally, early GPU benchmarks tested via Geekbench Metal indicate that the M2 chip is significantly faster than the M1 when it comes to graphics processing. The 10-core GPU received a score of 30627, up significantly from the M1's score of 21800.

Read on AppleInsider
«1

Comments

  • Reply 1 of 25
    danoxdanox Posts: 1,057member
    Nice now put it in a larger enclosure and allow it to breathe a just little.
    edited June 15 TheObannonFile
  • Reply 2 of 25
    thttht Posts: 4,440member
    Good result!
    killroywatto_cobra
  • Reply 3 of 25
    killroykillroy Posts: 214member
    Man I miss Robert Arthur 'Rob ART' Morgan . We would've got some no nonsense testing for the real world.
    blastdoormcdavernb2michelb76watto_cobra
  • Reply 4 of 25
    rob53rob53 Posts: 2,988member
    danox said:
    Nice now put it in a larger enclosure and allow it to breathe a just little.
    With these benchmarks it doesn't need a larger enclosure with multiple fans allowing it to be overclocked. I actually don't believe Apple's SoCs would benefit that much from overclocking. They've been designed for much lower amperage. Other ARM chips are designed for overclocking, which ends up messing everything up and requiring way too much amperage, like two 120V/20A outlets. 
    Beatswatto_cobra
  • Reply 5 of 25
    sdw2001sdw2001 Posts: 17,873member
    rob53 said:
    danox said:
    Nice now put it in a larger enclosure and allow it to breathe a just little.
    With these benchmarks it doesn't need a larger enclosure with multiple fans allowing it to be overclocked. I actually don't believe Apple's SoCs would benefit that much from overclocking. They've been designed for much lower amperage. Other ARM chips are designed for overclocking, which ends up messing everything up and requiring way too much amperage, like two 120V/20A outlets. 
    What system requires that? I’ve had two 1500 W electric heaters run on outlet like that.  
    watto_cobra
  • Reply 6 of 25
    sflocalsflocal Posts: 5,976member
    danox said:
    Nice now put it in a larger enclosure and allow it to breathe a just little.
    What the heck are you talking about?
    killroybageljoeywatto_cobra
  • Reply 7 of 25
    blastdoorblastdoor Posts: 2,696member
    killroy said:
    Man I miss Robert Arthur 'Rob ART' Morgan . We would've got some no nonsense testing for the real world.
    Agreed — he was a bright light in the Mac community. He is missed.
    killroymichelb76watto_cobra
  • Reply 8 of 25
    In single-core Geekbench 5 testing, the M2 chip received a score of 1919. The same chip got a multi-core Geekbench score of 8928. For comparison, a late 2020 MacBook Pro with an M1 chip received a single-core score of 1749 and a multi-core score of 7719.
    [...]
    Per the Geekbench 5 comparison, the M2 chip is 11.6% faster than the M1 in single-core scoring and 19.5% faster in multi-core scoring.
    Math fail. 1919/1749=1.097 ; 8928/7719=1.1566.

    So by the numbers provided in the article, geekbench (obviously not definitive, but a useful benchmark) says the M2 in the 13" Pro chassis is a hair under 10% faster in single-core and 16% faster in multicore.
    edited June 15 Alex1Nmuthuk_vanalingamwatto_cobra
  • Reply 9 of 25
    jdwjdw Posts: 1,068member
    While I appreciate this article very much, it surprisingly overlooks the obvious fact that the Air lacks an internal fan and therefore can and does limit performance during sustained operations.  Please consider adding some details about throttling in future articles on the Air which pertain to performance.  Anything less could be construed as being misleading.  Knowing how the lack of an internal fan affects performance is a key consideration that needs to be presented to the reader.  Thanks!
    Alex1Nwatto_cobra
  • Reply 10 of 25
    entropysentropys Posts: 3,507member
    But hard to do that prerelease, Jdw.

    and for MBA buyers not a common issue I expect.
    Alex1Nwatto_cobra
  • Reply 11 of 25
    jdw said:
    While I appreciate this article very much, it surprisingly overlooks the obvious fact that the Air lacks an internal fan and therefore can and does limit performance during sustained operations.  Please consider adding some details about throttling in future articles on the Air which pertain to performance.  Anything less could be construed as being misleading.  Knowing how the lack of an internal fan affects performance is a key consideration that needs to be presented to the reader.  Thanks!
    Actually the M1 air didn’t throttle very much, even under very heavy load. Check out MaxTech benchmarks/tests on YouTube 
    dewmedavwatto_cobra
  • Reply 12 of 25
    danox said:
    Nice now put it in a larger enclosure and allow it to breathe a just little.
    Yes. 15” “MacBook” with single fan would be wonderful. (Replaces 13.3” MBP)
    watto_cobra
  • Reply 13 of 25
    geekmeegeekmee Posts: 573member
    The M series of chips, I think proves that it’s no longer a muscle car war.
    watto_cobra
  • Reply 14 of 25
    DAalsethDAalseth Posts: 2,115member
    sdw2001 said:
    rob53 said:
    danox said:
    Nice now put it in a larger enclosure and allow it to breathe a just little.
    With these benchmarks it doesn't need a larger enclosure with multiple fans allowing it to be overclocked. I actually don't believe Apple's SoCs would benefit that much from overclocking. They've been designed for much lower amperage. Other ARM chips are designed for overclocking, which ends up messing everything up and requiring way too much amperage, like two 120V/20A outlets. 
    What system requires that? I’ve had two 1500 W electric heaters run on outlet like that.  
    Or as they are more commonly known, Pentium IIIs
    LOL
    muthuk_vanalingamdavwatto_cobra
  • Reply 15 of 25
    For comparison, a late 2020 MacBook Prowith an M1 chip received a single-core score of 1749 and a multi-core score of 7719.

    Why do people always cherry-pick single results?

    Geekbench shows the result for a 2020 MacBook Pro 13" M1 to be 1707 single core, not 1749 like you stated.

    Geekbench averages out numerous tests and discards scores that are unusually low/high before posting a result for their benchmark charts. This is why it often takes days/weeks (or months for less popular devices) before a device appears in the Geekbench Charts. For example, the i7-1260P/1280P aren't in the chart yet because not enough test results have been submitted,

    watto_cobra
  • Reply 16 of 25
    CheeseFreezeCheeseFreeze Posts: 1,015member
    What’s equally interesting for video editors; there’s a media engine in the M2, so you don’t necessarily need a Pro for that. 
    That’s really useful (and powerful enough) for most editors on the go. 
    foregoneconclusionwatto_cobra
  • Reply 17 of 25
    mcdavemcdave Posts: 1,919member
    The Mac14,7 device name is odd for a MacBook Air, Pro or Mac mini. Are we looking at a new desktop?
    thtwatto_cobra
  • Reply 18 of 25
    entropysentropys Posts: 3,507member
    I bet you are fun at parties Eric.

    TBH I round off anyway and would not chuck a wobbly over 40 points.
    sorry. 42 points.
    edited June 16 davwatto_cobra
  • Reply 19 of 25
    michelb76michelb76 Posts: 401member
    Let the years of selling small incremental upgrades continu!
    edited June 16
  • Reply 20 of 25
    JinTechJinTech Posts: 892member
    What’s equally interesting for video editors; there’s a media engine in the M2, so you don’t necessarily need a Pro for that. 
    That’s really useful (and powerful enough) for most editors on the go. 
    For the professionals, the M2 Pro and M2 Pro Max will surely be an absolute beast of a laptop with likely double that of what we have available today. I can't wait to see the performance, and I will keep telling myself that the M1 Pro Max is still just as good  :D
    watto_cobra
Sign In or Register to comment.