Intel vs. M1 – The empire strikes back

Intel throws down the gauntlet.

• Intel has put up a comparison of its latest chip against the M1 that it says tells a very different story but the tests are clearly skewed in Intel’s favour and don’t match real-world testing.
• However, this is a good sign that Intel intends to slug it out with Apple and remains committed to regaining its manufacturing crown rather than slowly slipping into irrelevance and oblivion.
• Intel’s tests its 11^th generation Core i7-1185G7 which is on Intel’s 10nm process rather than the benchmark which Apple picked which was on 14nm (see here).
  • First: productivity: Intel demonstrated significantly better performance on photo enhancement and various Microsoft Office tasks but Intel could easily be accused of picking and choosing the tests.
  • This is especially the case with Microsoft Office which has been native on Intel for years and is a relative newcomer to Arm.
  • Second, AI-related: tasks such as noise reduction on pictures and so on.
  • Here, Intel crushes the M1 and in this specific area, this has also been observed by independent observers.
  • Third: Games: which could easily be described as an unfair comparison as everyone knows Macs are not designed for games.
  • Here, Intel beats the M1 easily on many games and the ones that the M1 won’t run, Intel cruelly scores the M1 at zero.
  • Apple will take no prisoners in this battle and I see no reason why Intel should either as this is an existential fight for Intel.
  • Fourth, battery life: which is where Intel’s test is highly suspect and does not stack up with the real world experience of almost every reviewer.
  • Here Intel puts up an Acer Swift 5 against the MacBook Air and finds a dead heat of watching Netflix of 10hrs and 12 minutes for the MacBook and 10hrs and 6 minutes for the Acer Swift 5.
  • There is a distinct possibility that Intel has chosen this benchmark very carefully and this will need to be replicated by third parties before I give it much credibility.
  • This is because all of the reviewers have been Apple to broadly match Apple’s battery life claims for M1 powered devices with real daily usage and testing.
  • Hence, I suspect that this test is an outlier but I am open to being convinced otherwise if this can be replicated elsewhere by third parties on different devices.
• Given, how different these results are compared to the results that Apple ran, I can only conclude that the difference between Intel’s 14nm and its 10nm is significant.
• The problem is that there is only one production device that uses this 10nm chipset in production at the moment which is the MSA Prestige 14 Evo which will complicate reliable 3^rd party verification of these tests for a while.
• The main take-home from these tests is that the 10nm process may be making a substantial difference to Intel’s performance which bodes well for when Intel finally makes it onto its equivalent of TSMC’s 5nm.
• However, these tests will not be convincing until they have been replicated in the real world by 3^rd party independent testers.
• The good news for Intel here is that a lot of the M1’s leap in performance and battery life may be down to TSMC’s 5nm process meaning that everyone else will be able to catch-up.
• Intel needs to do more than that though as it will have to regain its performance edge in order to offset the power efficiency advantage that Arm-based processors always have.
• The way ahead for Intel is going to be difficult and there is the possibility that the recent rally driven by a short-term outsourcing theme completely unwinds.
• This will create a very large opportunity for those that believe that Intel will be successful in its turnaround.
• I am not yet convinced that Intel will succeed but I am encouraged by the fact that if it goes down, it will go down fighting.