• Did you order new AirTags? We've opened a dedicated AirTags forum.

intrepidcase

macrumors member
Original poster
Nov 13, 2019
40
12
So I am deliberating between the i9 2.3 and i9 2.4. In theory, the speed bump is < %5, both at nominal speed and full turbo. But early geekbench scores are showing a consistent 10% + difference, both single-core and multi-core.


averages so far:

i7 2.6: 1050/5638 single/multi

i9 2.3: 1013/6205

i9 2.4: 1186/7290

Given that i7 2.6 and i9 2.4 single-core scores are roughly the same, and i9 2.3 single-core is 20% lower, the i9 2.3 looks to be the *bad* choice of cpu's here ?!#
and so it comes down to multi-core needs, in which case the i9 2.4 is 30% higher than the i7 2.6.

[edit: first numbers were wrong in 2 cases, have fixed. that said, i still question the i9 2.3 as it looks to be a trade-off with the i7 2.6 - higher multi-core but a bit lower single-core, and the i9 2.4 is still ~15% higher than i9 2.3 in both measures]

These numbers seem surprising to me ... not sure how this lines up with historical differences on say the 2019 15" MBP's. I am unclear as to how much geekbench really hits the thermal throttles, but it seems like Apple has really fixed this issue? and that the i9 2.4 might be worth it.
 
Last edited:

intrepidcase

macrumors member
Original poster
Nov 13, 2019
40
12
Yeah, I am wondering about this but one would assume it didn't? All/most of the i9 2.3 scores have 16GB of ram, and all/most of the i9 2.4 scores have 32GB of RAM. Regardless, I am getting 32GB no matter what.
 
Comment

am2am

macrumors regular
Oct 15, 2011
221
101
Undervolting in i9 2.4.
We've seen the same performance difference in 15" 2019 models - better results thanks to volting tweaking in higher end CPU.
That's at least my theory - to confirm we need more data.
Having the same dilemma which one to pick - that is why I'm still waiting - otherwise I would have ordered already :)
 
Comment

intrepidcase

macrumors member
Original poster
Nov 13, 2019
40
12
Did the undervolted high end 15" 2019 actually perform better outside of benchmarks?
What are the real-world consequences/downsides of undervolting a cpu? why not do it to all models?

(edit: if apple has figured out to successfully undervolt some cpu's without stability issues, why not do it across the board .... why not other vendors as well? ... guess i need to research this further)
 
Last edited:
Comment

Howard2k

macrumors 68040
Mar 10, 2016
3,013
2,231
(WRT single threaded score vs multithreaded score) it depends how you use your MacBook. If you’re using one today you should keep Activity Monitor open and you can see whether your CPU bottlenecks are single threaded or multithreaded.
 
Comment

intrepidcase

macrumors member
Original poster
Nov 13, 2019
40
12
@Howard2k Thanks. I develop highly multi-threaded/resource intensive server apps so I am not questioning if I will use any available cpu ... what I am trying to understand is what the trade-offs are in terms of thermal throttling, fan noise, battery life etc. If the 2.4 only provides marginal increases in performance at the cost of much higher throttling, heat, etc then it might not be worth it. So trying to wrap my head around the early benchmarks.
 
Comment

wallysb01

macrumors 68000
Jun 30, 2011
1,576
775
Thanks for aggregating those scores. When I looked no i7s were up yet. Great to see their single core score is so high. It will be interesting to see how longer running benchmarks do, because you are definitely right that the 2.3 i9 looks like a funny spot. Noticeably worse single core performance than the cheaper i7 and not a huge upgrade price to the 2.4 i9 brings what are looking like dramatic improvements. Obviously more thorough testing is needed, but this is encouraging.
 
Last edited:
  • Like
Reactions: Macintosh IIcx
Comment

The Mercurian

macrumors 68020
Mar 17, 2012
2,077
2,318
I don't suppose anyone knows an R package that will filter and download geekbench scores? ?
 
Comment

athanis

macrumors newbie
Nov 15, 2019
4
18
You miscalculated the single-core average for the i7 and the i9 2.3:

i7 : 1050 (8402 / 8 = 1050.25)
i9 2.3: 1014 (17234 / 17 = 1013.76)

So the i9 2.3 single-core performance is just about 4 % slower, not 20 %.
 
Comment

mmjrk

macrumors newbie
Oct 21, 2019
23
16
Seems to me that the geekbench scores are actually slightly worse than 15'' 2019 models.
 
Comment

pshifrin

macrumors 6502
Mar 14, 2010
478
348
My 2019 15 inch 2.4/32/1TB gets 1173/7105 (run 1) and 1251/7113 (run 2) with every app closed. I also quit dropbox, creative cloud and most other menu bar apps.

I had to get a work replacement in July and this unit has been rock solid. Needed it same day so paid the Apple Tax of in store high end BTO otherwise would have gotten the 2.3. So of course i’d like the 16 but i like the 19 revision of the keyboard and the 16 would be essentially the same speed. That is until next spring when this one will become a hand me down ;)
 
  • Like
Reactions: The Mercurian
Comment

davecom

macrumors member
Aug 10, 2009
95
7
Yeah I originally ordered the i9 2.3 right after launch, but then I saw the first few geekench results and on a hunch cancelled and ordered the i9 2.4. I’m increasingly glad I did. It seems Apple really has done something for the thermals. That said, the difference does seem surprisingly large between both the 2.3 and the 2.4. Also surprising the 2.3 is down from the previous generation.
 
Comment

cap7ainclu7ch

macrumors 6502
Jun 30, 2010
383
331
Well you guys are a bad influence. Cancelled my 2.3 order and went with the high end 2.4 BTO option. 10% increase in performance for 150$ (edu discount) seems like a no brainer for a device I plan on using for at least 4 years in a high end capacity. Get it a week earlier as well.
 
Comment

MacBookGamer

macrumors regular
Nov 15, 2019
114
117
They aren't showing a massive improvement. The 2.3GHz is consistently scoring at 6900 multicore. The 2.4GHz even has one score in 6500.
 
Comment

wallysb01

macrumors 68000
Jun 30, 2011
1,576
775
They aren't showing a massive improvement. The 2.3GHz is consistently scoring at 6900 multicore. The 2.4GHz even has one score in 6500.

no the 2.3 isn’t showing consistent 6900. Range is more 6200-6800.
Also, it’s extremely foolish to look at the low end outliers. By this logic we should dismiss the 9880H because one computer has a 4900? Those computers are probably doing something else or who knows what. The 9980HK’s range is a bit more like 6900-7400. 7300/6700 is still about a 10% improvement.
 
  • Like
Reactions: The Mercurian
Comment

honcho

macrumors member
Apr 19, 2011
58
16
What puzzles me is that every 2.4GHz reports 32GB of RAM, while every 2.3GHz has 16GB. I could understand it if these were stock configurations, but 32GB is not standard in the 2.4GHz. Also, I would have thought that the 32GB RAM option would be popular in the 2.3GHz model. Strange.
 
Comment

MacBookGamer

macrumors regular
Nov 15, 2019
114
117
no the 2.3 isn’t showing consistent 6900. Range is more 6200-6800.
Also, it’s extremely foolish to look at the low end outliers. By this logic we should dismiss the 9880H because one computer has a 4900? Those computers are probably doing something else or who knows what. The 9980HK’s range is a bit more like 6900-7400. 7300/6700 is still about a 10% improvement.
We don't have enough data yet to identify outliers. It's possible that 6900 for the 2.3GHz is the norm and the rest are outliers. Once we get around 20 tests we should know.
 
Comment
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.