News Apple M3 CPUs Hit 4.05 GHz, Challenge Raptor Lake in Geekbench

[Admin]

Apple's M3-series CPUs managed to significantly increase frequency and single-thread performance, according to leaked Geekbench 6 scores.

Apple M3 CPUs Hit 4.05 GHz, Challenge Raptor Lake in Geekbench : Read more

 

[peachpuff]

I love playing geek bench, it's my fav!

 

[Amdlova]

Best benchmark for apple products lol

 

[TerryLaze]

Apple's M3-series CPUs managed to significantly increase frequency and single-thread performance, according to leaked Geekbench 6 scores.

Apple M3 CPUs Hit 4.05 GHz, Challenge Raptor Lake in Geekbench : Read more

Lol, you had to run a geekbench of the 14900k today at reduced clocks of just 3.2Ghz just to make a story?!

ASUS System Product Name - Geekbench

Benchmark results for an ASUS System Product Name with an Intel Core i9-14900K processor.

browser.geekbench.com

CPU Information
Name	Intel(R) Core(TM) i9-14900K
Topology	1 Processor, 24 Cores, 32 Threads
Identifier	GenuineIntel Family 6 Model 183 Stepping 1
Base Frequency	3.20 GHz
Cluster 1	8 Cores
Cluster 2	16 Cores
Maximum Frequency	3200 MHz

At 5.5Ghz max clocks the 13900k hits 4220 single core score , and the real single clock of the 14900k is 6Ghz...

ASRock Z690 AQUA OC - Geekbench

Benchmark results for an ASRock Z690 AQUA OC with an Intel Core i9-13900K processor.

browser.geekbench.com

 

[ezst036]

Any overclocking with a big water cooler?

Or even better, LN2?

 

[JamesJones44]

Lol, you had to run a geekbench of the 14900k today at reduced clocks of just 3.2Ghz just to make a story?!

ASUS System Product Name - Geekbench

Benchmark results for an ASUS System Product Name with an Intel Core i9-14900K processor.

browser.geekbench.com

At 5.5Ghz max clocks the 13900k hits 4220 single core score , and the real single clock of the 14900k is 6Ghz...

ASRock Z690 AQUA OC - Geekbench

Benchmark results for an ASRock Z690 AQUA OC with an Intel Core i9-13900K processor.

browser.geekbench.com

This is not the typical score for a 13900k. Looking at the search results I would say the average 13900k only checks in at ~3200 single core and ~24000 multi core when allowed to boost to 5.5 GHz.

Geekbench Search - Geekbench

browser.geekbench.com

Micro-Star International Co., Ltd. MS-7E03 - Geekbench

Benchmark results for a Micro-Star International Co., Ltd. MS-7E03 with an Intel Core i9-13900K processor.

browser.geekbench.com

The 14900K fairs about the same. I couldn't find any that were able to boost to 6 GHz but there are a few above 5.6 GHz

Micro-Star International Co., Ltd. MS-7D29 - Geekbench

Benchmark results for a Micro-Star International Co., Ltd. MS-7D29 with an Intel Core i9-14900K processor.

browser.geekbench.com

 

[thestryker]

The single threaded performance uplift over the M2 seems pretty linear with the clockspeed increase. Seems like whatever CPU architecture work was done is predominantly on the efficiency cores.

 

[TJ Hooker]

Lol, you had to run a geekbench of the 14900k today at reduced clocks of just 3.2Ghz just to make a story?!

ASUS System Product Name - Geekbench

Benchmark results for an ASUS System Product Name with an Intel Core i9-14900K processor.

browser.geekbench.com

At 5.5Ghz max clocks the 13900k hits 4220 single core score , and the real single clock of the 14900k is 6Ghz...

ASRock Z690 AQUA OC - Geekbench

Benchmark results for an ASRock Z690 AQUA OC with an Intel Core i9-13900K processor.

browser.geekbench.com

The Geekbench scores linked in this article are consistent with those from various 14900k reviews. The listed maximum frequency is probably just a glitch, possibly from Geekbench not handling E cores correctly. The result you posted had E cores disabled.

 

[TerryLaze]

The Geekbench scores linked in this article are consistent with those from various 14900k reviews. The listed maximum frequency is probably just a glitch, possibly from Geekbench not handling E cores correctly. The result you posted had E cores disabled.

Then what were the max clocks of the results that I linked?! Because it says 5.5Ghz and if e-cores are disabled it shouldn't have the glitch, right?!
If what I linked are propper results for 5.5Ghz then I still stand by my opinion.

 

[TJ Hooker]

Then what were the max clocks of the results that I linked?! Because it says 5.5Ghz and if e-cores are disabled it shouldn't have the glitch, right?!
If what I linked are propper results for 5.5Ghz then I still stand by my opinion.

I meant the results that had the E cores enabled probably had a glitch with the reported max frequency.

If Geekbench scores vary that much purely based on E cores on or off (not enough info to say this with certainty though) then either:

1. Real workload performance can vary significantly based on E core on or off, meaning the lower Geekbench score with E cores enabled is valid; or

2. Geekbench performance isn't indicative of performance in real workloads, meaning it's a crappy benchmark and we shouldn't worry about conflicting results because they don't mean anything anyways.

My hunch is the latter.

Edit: I guess Geekbench could have a bug that specifically impacts performance on heterogeneous x86 processors. But given that those have been mainstream since the Alder Lake release ~2 years ago, if Geekbench still hasn't fixed it that also doesn't paint the benchmark in a very good light.

 

[TerryLaze]

I meant the results that had the E cores enabled probably had a glitch with the reported max frequency.

If Geekbench scores vary that much based on E cores on or off, then either:

1. Real workload performance can vary significantly based on E core on or off, meaning the lower Geekbench score with E cores enabled is valid; or

2. Geekbench performance isn't indicative of performance in real workloads, meaning it's a crappy benchmark and we shouldn't worry about conflicting results because they don't mean anything anyways.

My hunch is the latter.

Edit: I guess Geekbench could have a bug that specifically impacts performance on heterogeneous x86 processors. But given that those have been mainstream since the Alder Lake release ~2 years ago, if Geekbench still hasn't fixed it that also doesn't paint the benchmark in a very good light.

We are talking about the single core results here....
IF geekbench uses the e-cores for this test when they are enabled then yes, it's an extremely bad test.
Otherwise the score changes based on clocks (and other factors like ram and so on)

 

[TJ Hooker]

We are talking about the single core results here....
IF geekbench uses the e-cores for this test when they are enabled then yes, it's an extremely bad test.
Otherwise the score changes based on clocks (and other factors like ram and so on)

Sure, but the clocks reported in the geekbench results obviously aren't trustworthy, at least for some processors.

E.g. there's plenty of other 14900k results with a reported max clock of 5.7 GHz, but still getting around the same score as the result posted in the article with a reported max clock of 3.2 GHz.

I can't say for sure why the result you found is so much higher, but I doubt the reported clock speed has anything to do with it.