News AMD Phoenix 2 Review Evaluates Zen 4, Zen 4c Performance

[Admin]

Hardware enthusiast David Huang analyzes the Ryzen Z1 (Phoenix 2) silicon to measure Zen 4 and Zen 4c performance.

AMD Phoenix 2 Review Evaluates Zen 4, Zen 4c Performance : Read more

 

[GustavoVanni]

Wait, wait, wait!
You're saying that ROG Ally uses AMD's first hybrid chip with Zen4 and Zen4c and nobody mencioned this in any review in the last few months?
Or is this a different AMD Ryzen Z1?

 

[bit_user]

Wow, the 7840U performed only 18.2% better @ 30 W, in spite of having 33.3% more cores (not to mention all big cores)!! It sure would be nice if they had the same memory speeds, so we could know if that's helping the Z1.

Ryzen 7 7840U's iGPU has 12 RDNA 3 compute units with a max 2.7 GHz clock, while the Ryzen Z1 only has four RDNA 3 compute units at 2.8 GHz.

IMO, this makes it pretty nuts that it's being used in gaming handhelds. They really should've gone with 10 or 12 CU.

 

[greenreaper]

Fun fact: the RDNA 2 cores in the 7000-series iGPU can run at 2.7-2.8Ghz too, at about 1.3 V (+9 GPU offset). ASUS's convenient tweaker setting for this doesn't seem to work, though; I had to set the one in AMD Overclocking instead (unlike the regular cores).

 

[NeoMorpheus]

If i recall correctly, zen4c also support avx-512 unlike intels e cores.

Heck, i think that even some p-cores don’t support it either.

 

[bit_user]

Heck, i think that even some p-cores don’t support it either.

On the P-core -only Xeon processors, they all support AVX-512.

 

[The Hardcard]

Wait, wait, wait!
You're saying that ROG Ally uses AMD's first hybrid chip with Zen4 and Zen4c and nobody mencioned this in any review in the last few months?
Or is this a different AMD Ryzen Z1?

The prrevious reviews were with the Z1 Extreme.

 

[Gillerer]

Wow, the 7840U performed only 18.2% better @ 30 W, in spite of having 33.3% more cores (not to mention all big cores)!! It sure would be nice if they had the same memory speeds, so we could know if that's helping the Z1.


IMO, this makes it pretty nuts that it's being used in gaming handhelds. They really should've gone with 10 or 12 CU.

As implied in the article, mobile chips are mostly limited by power so the performance delta at the same power isn't that surprising. If both were allowed more power, the 7840U would run away with it.

AMD already has more CUs on the bigger brother Phoenix (1) - which is the 7840U. While enthusiasts would always want a cheap gaming machine, this smaller and cheaper variant is primarily intended for cheaper thin laptops; general use, video playback etc. Tripling the CUs would not benefit anyone but gamers, but would make it less inexpensive, which sort of defeats the purpose of having two different Phoenix dies to begin with.

 

[SirNathan]

Wow, the 7840U performed only 18.2% better @ 30 W, in spite of having 33.3% more cores (not to mention all big cores)!! It sure would be nice if they had the same memory speeds, so we could know if that's helping the Z1.


IMO, this makes it pretty nuts that it's being used in gaming handhelds. They really should've gone with 10 or 12 CU.

The regular Z1 has 4 CUs, but the Z1 Extreme in the ROG Ally and similar handhelds has 12 CUs. Maybe you already knew that, but it definitely confused me the first couple times I read it, and there are probably others who are confused too...

 

[TerryLaze]

Wow, the 7840U performed only 18.2% better @ 30 W, in spite of having 33.3% more cores (not to mention all big cores)!! It sure would be nice if they had the same memory speeds, so we could know if that's helping the Z1.


IMO, this makes it pretty nuts that it's being used in gaming handhelds. They really should've gone with 10 or 12 CU.

Huang highlighted that the HP Elitebook 835 G10 has a default SPL (sustained power limit) of 25W, which isn't modifiable by the user.

Difference between 25w and 30w is 18.2% ,so much wow, many improve...



The real improvement is this, the c cores can run at lower volts/clocks which means they can run at lower power, they are still way above the 800Mhz that intel can achieve but at least they are doing something.

Zen 4c needs a higher core voltage to reach the same clock speeds as Zen 4. The VID (voltage identification definition) charts revealed that Zen 4 hits the Vmin (the minimal voltage that a processor requires for a workload at a particular frequency) at 2.3 GHz. In contrast, Zen 4c arrives at the Vmin below 1.5 GHz. The V/F (voltage-to-frequency) curve for both cores overlaps at 1.5 GHz. Zen 4c's power efficiency resides in between 1.5 GHz and 2 GHz.