Discussion Cinebench 2026

[Hulk]

Created a little table with GPU scores. Trying to find scores at stock speeds. If you have one of the missing cells let me know!

Also, I used AI to help me fill in some of the specs so if you see something that is just wrong, that is why.

 

[DaaQ]

I can't extract your cpu average frequency from what you posted. You would need to watch in HWinfo during the 3 tests, or better/easier lock down max frequency in Power Plan.

It was around 3.5ish in general MHZ.

 

[Hulk]

It was around 3.5ish in general MHZ.

Are you sure? Seems like it should be more around 5.5?

 

[DaaQ]

Are you sure? Seems like it should be more around 5.5?

AMD said 5.2ghz and hwinfo during second test is ss and avg maxed about 5.126ghz.

EDIT: SC was about 5.2 on AMD graph and 17% ut. HWinfo about 4.4 bouncing between 2.8, 3.5, 4.8Ghz 105TDP

This was back to 105tdp disabled. Below was 5 mins in.

 

[Joe NYC]

Created a little table with GPU scores. Trying to find scores at stock speeds. If you have one of the missing cells let me know!

Also, I used AI to help me fill in some of the specs so if you see something that is just wrong, that is why.

View attachment 136191

5060 Ti looks like an PPA efficiency champ.

 

[SPBHM]

I had a glance at task manager during the SC run, task manager reported 4.17GHz and it maintained that clock for the duration. I will check in depth later using HW info

for MT it is probably running much lower, this is 6 core coffee lake at 4ghz locked on MT, ST 4.2

 

[johnsonwax]

IMO kinda pointless to talk about Zen7 and M8 efficiency this much when we haven’t even gotten their preceding architectures yet.

Surely we can get a Zen8 Speculation Thread going from this.

 

[OEta97]

Mysterious CPU

 

[Nothingness]

Assuming that's not a fake @Det0x will have to push a bit his mysterious CPU

 

[Det0x]

Can see what i can do when i get home from work, but i can tell you guys its not Zen6. 😅
More like a gimmick of the benchmark itself when you start to disable cores...

1087 screenshot above is most likely a Raptor Lake since SMT is enabled

 

[eek2121]

Surely we can get a Zen8 Speculation Thread going from this.

Just wait till I tell you about Zen 9! 🤣

 

[MS_AT]

More like a gimmick of the benchmark itself when you start to disable cores...

It's using 2 physical cores thinking it's running on single SMT enabled core?

 

[desrever]

Just wait till I tell you about Zen 9! 🤣

Zen 9 factorial is Zen 362,880 🤪

 

[CouncilorIrissa]

It's using 2 physical cores thinking it's running on single SMT enabled core?

That's what I'm thinking.
Assuming somewhat linear scaling from 1 -> 2 threads, the 1087 score would amount to 543, which is somewhat typical for RPL in 1t

 

[ArrogantHair]

285K P56 E50
- have iGPU enabled, so not ideal scores

 

[mikegg]

"M4 Pro is roughly 52% faster in Cinebench ST and 3.6x more efficient than Strix Halo."

Which is clearly a flawed assumption. Neither of the two (performance and efficiency) hold true when you do a "sanity check" by checking MT numbers.

Why should MT numbers even factor into this? MT speed and efficiency is mostly influenced by how many cores there are and then run them at a lower clock. I clearly stated ST.

So, you started with flawed assumption of +360% efficiency, doubled it and ended up twice as flawed +720% efficiency that AMD would have to make up.

Why is the flawed? You claimed Zen 7 should beat Apple in speed and efficiency. I showed what Zen 7 needed to do to beat M8, right?

Nothing is flawed. We are speculating with logic here.

Apple is leading, but by 1/10th of what you thought.

Source?

 

[Tup3x]

X2 Elite Extreme.

Non-Elite 18 core version is apparently 6xxx.

 

[poke01]

X2 Elite Extreme.
View attachment 136348
Non-Elite 18 core version is apparently 6xxx.

faster than m4 max and Strix halo

 

[511]

X2 Elite Extreme.
View attachment 136348
Non-Elite 18 core version is apparently 6xxx.

Similar to 275HX but I am pretty sure SDX2EE has lower power draw

 

[mvprod123]

X2 Elite Extreme.
View attachment 136348
Non-Elite 18 core version is apparently 6xxx.

QRD always shows higher scores.

 

[adroc_thurston]

QRD always shows higher scores.

Eh anything with good cooling will get there.

 

[mikegg]

And you made this projection, in the same post based on this assumption:

"M4 Pro is roughly 52% faster in Cinebench ST and 3.6x more efficient than Strix Halo."

It's not an assumption. It's a fact based on Notebookcheck's data.

Which is clearly a flawed assumption. Neither of the two (performance and efficiency) hold true when you do a "sanity check" by checking MT numbers.

MT numbers are a result of having more cores and running them at lower clock rate. AMD can get higher MT efficiency than Apple today, as long as they put more cores and run them at lower clocks. This is comparatively very easy to achieve than ST speed and efficiency.

So, you started with flawed assumption of +360% efficiency, doubled it and ended up twice as flawed +720% efficiency that AMD would have to make up.

Again, I didn't make an assumption. It's a fact based on Notebookcheck's Cinebench 2024 testing.

Yes, I projected 720% efficiency because that's how much Zen 7 would have to beat Strix Halo. Remember that Zen 7 is going up against M8, not M4.

 

[511]

It's not an assumption. It's a fact based on Notebookcheck's data

You can lower the voltage and reach a significant less figures

 

[mikegg]

You can lower the voltage and reach a significant less figures

Sure, lower the voltage and efficiency should improve. But M4 is already 50% faster in ST at default voltages. How much slower can you tolerate Strix Halo ST?

If you lower M4 voltage, I'm sure it can get more efficient as well.

Anyways, we all know chip efficiency is on a curve. This is not debated. What we're talking about is the performance and power usage at default settings that AMD and Apple want you to run the chips at.

 

[511]

Sure, lower the voltage and efficiency should improve. But M4 is already 50% faster in ST at default voltages. How much slower can you tolerate Strix Halo ST?

Can't say without looking at V/F curve also Node matters N4P vs N3E not an iso node comparison we would have better chance next year when everyone has designs on N2 .

If you lower M4 voltage, I'm sure it can get more efficient as well.

For sure but how much V/F curve is a bi***.

Anyways, we all know chip efficiency is on a curve. This is not debated. What we're talking about is the performance and power usage at default settings that AMD and Apple want you to run the chips at.

Intel/AMD in the end are the mercy of what OEMs do so there is lots of delta vs Apple you need a reference design for that comparison.