Support NeoGAF

[Thick Thighs Save Lives]

David Huang has shared a review of the first processor over on Chinese platform Zhuanlan/Zhihu shortly after the first die shot featuring AMD Phoenix2 (aka Small Phoenix) silicon emerged.

As anticipated, the Ryzen Z1 APU and Ryzen 5 7540U APU are set to feature the 6-core Phoenix2 CPU, incorporating a hybrid layout of Zen4 and Zen4c (dense) cores. These processors are specifically tailored for low-power applications, such as lightweight laptops and, in the case of the Ryzen Z1, handheld gaming consoles.

The review primarily focuses on the Z1 APU, which employs a customized power profile distinct from its laptop counterpart. Nevertheless, both variants share the same silicon, configured in an unusual configuration, with the objective of delivering a cost-effective and efficient solution based on AMD’s latest architecture.

The Zen4c (dense) cores operate at a different frequency compared to the ‘classic’ cores, with Zen4c running at 3.5 GHz and Zen4 at 4.9 GHz. However, when both types of cores are configured to run at the same 3.2 GHz frequency, their performance becomes nearly identical. This is due to the fact that both core types utilize the same ISA and frontend. An interesting aspect is that four out of six cores are designed to run at 3.5 GHz. The Phoenix2 requires less power to attain its maximum boost clock of 4.9 GHz, with power consumption measured at a mere 12.7W.

David’s review includes initial synthetic benchmarks and 3D gaming tests. Given that the Phoenix2 has a distinct core layout and reduced GPU capabilities, featuring only one-third of the Phoenix1 RDNA3 CU count (4 vs. 12), the performance drop is noticeable. However, the gaming experience can vary depending on factors like power limits, clock configurations, system memory, and cooling efficiency.

Comparing the Ryzen 7 7840U (aka Ryzen Z1 Extreme) with a 25W power limit, one can anticipate an 80% improvement in performance compared to the Ryzen Z1 running at 15W.

In various 3DMark graphics tests, a substantial performance gap is evident between the Z1 processor and the fully specced 7840U. The 25W 7840U generally outperforms the 15W Z1 by about 80%, with even the smallest gap in Fire Strike reaching approximately 70%. Despite being configured for 30W, the Phoenix2, with three times fewer cores than Phoenix1, can sustain a GPU clock of 2.8 GHz with just 22W of power, which drops to 1.8-2.0 GHz at 15W.

In synthetic tests, the difference between the 12CU RDNA3 iGPU in the Z1 and the 4CU RDNA3 iGPU in the 7840U narrows down to 28.6% when comparing the 7840U at 25W and the Ryzen Z1 at 30W. Notably, the Z1 running at 30W performs 34% better than at 15W. In more graphics-intensive titles like Metro Exodus or Cyberpunk 2077, the 7840U is 24.6% to 57% faster, but the performance gap between the Z1 at 30W and 15W narrows to 15% – 19%.

In conclusion, the AMD Phoenix, also known as Ryzen Z1 or Ryzen 5 7540U, is clearly optimized for MMO titles and games that can run effectively at lower TDPs. For handheld gamers seeking modern titles with higher fidelity, resolution, and support for high-resolution textures, which may require sticking to a higher TDP profile, the Ryzen Z1 Extreme might be a better choice. However, for gamers primarily focused on popular titles that are well-optimized for lower-spec hardware, the Ryzen Z1 may offer a more suitable performance-to-power balance.

Regarding pricing, the ASUS ROG Ally with the Z1 non-Extreme chip has not been released yet, but the company has indicated it may be priced $100 lower than the Extreme variant. Whether this price difference justifies the spec downgrade remains to be seen.

AMD Ryzen Z1 non-Extreme featuring Phoenix2 silicon has been tested - VideoCardz.com

AMD Ryzen Z1/Phoenix2 APU seems to perform well at 15W David Huang has shared a review of the first processor over on Chinese platform Zhuanlan/Zhihu shortly after the first die shot featuring AMD Phoenix2 (aka Small Phoenix) silicon emerged. AMD Phoenix2 APU, Source: Golden Pig Upgrade As...

videocardz.com

 

[Setsuna Mudou]

Pretty sure I have a handheld with this apu for a couple months now.
The ROG Ally also has the Z1 Extreme, so this feels like it was written last year. What am I missing?

 

[Thick Thighs Save Lives]

Pretty sure I have a handheld with this apu for a couple months now.
The ROG Ally also has the Z1 Extreme, so this feels like it was written last year. What am I missing?

Z1 is the cut-down sibling of the Z1 extreme which has yet to launch, but was announced a few months back by AMD. The biggest pro is that it runs much better at lower wattage (e.g. 15W) than the Z1 extreme when taking into consideration its specs. It's also cheaper, which means we'll see less expensive handhelds on the markets hopefully soon.

Fake edit: They're also using these new Zen4c cores which are much better suited to tablets and handhelds.

AMD Zen 4c Not an E-core, 35% Smaller than Zen 4, but with Identical IPC

AMD on Tuesday (June 13) launched the EPYC 9004 "Bergamo" 128-core/256-thread high density compute server processor, and with it, debuted the new "Zen 4c" CPU microarchitecture. A lot had been made out about Zen 4c in the run up to yesterday's launch, such as rumors that it is a Zen 4 "lite"...

www.techpowerup.com

 

[Fafalada]

The biggest pro is that it runs much better at lower wattage (e.g. 15W) than the Z1 extreme when taking into consideration its specs.

I mean it'd help if they showed some Z1E benchmarks at 15W for comparison - but instead they compare 15/30W to 25W Z1E - making this really hard to say anything about.

 

[Thick Thighs Save Lives]

I mean it'd help if they showed some Z1E benchmarks at 15W for comparison - but instead they compare 15/30W to 25W Z1E - making this really hard to say anything about.

Doesn't really matter since it doesn't really compete with the Z1E in the first place. The increased efficiency it's where it's at; look at the graphs barely any loss at 15W vs 30W.

 

[Setsuna Mudou]

Z1 is the cut-down sibling of the Z1 extreme which has yet to launch, but was announced a few months back by AMD. The biggest pro is that it runs much better at lower wattage (e.g. 15W) than the Z1 extreme when taking into consideration its specs. It's also cheaper, which means we'll see less expensive handhelds on the markets hopefully soon.

Fake edit: They're also using these new Zen4c cores which are much better suited to tablets and handhelds.

mhmmm... wasn't the ROG Ally sold in 2 SKUs, the Z1 and Z1 Extreme? I thought it was already out for some time, no?

Seems a bit limiting though, especially paying $800+ for such a machine in 2023. At that level might as well keep the steam deck which is already quite efficient at such low wattage.

 

[Thick Thighs Save Lives]

mhmmm... wasn't the ROG Ally sold in 2 SKUs, the Z1 and Z1 Extreme? I thought it was already out for some time, no?

Seems a bit limiting though, especially paying $800+ for such a machine in 2023. At that level might as well keep the steam deck which is already quite efficient at such low wattage.

Don't think the ROG Ally SKU with the base Z1 is in people' hands yet, but maybe I'm mistaking.

 

[SABRE220]

Yeah Sony really has no excuses for not using zen4 in the ps5 pro seeing the designs are mature enough to push out handheld apu chips. I cant help but feel underwhelmed by the actual performance put out by the apu though compared to the hardware specs listed. I cant help but feel this will get embarrassed by the switch 2 chipset based on reports so far.

 

[Fafalada]

Doesn't really matter since it doesn't really compete with the Z1E in the first place. The increased efficiency it's where it's at; look at the graphs barely any loss at 15W vs 30W.

Reading that the other way - there are no gains past 15W (because pumping more power just does nothing for perf). Efficiency increase would be interesting against Z1E because at 15W - that chipset is power limited, and it only really starts to perform to its potential at 22W+.
Ie. the delta at 15W might actually be quite narrow if the efficiency argument is true.

 

[Thick Thighs Save Lives]

Reading that the other way - there are no gains past 15W (because pumping more power just does nothing for perf). Efficiency increase would be interesting against Z1E because at 15W - that chipset is power limited, and it only really starts to perform to its potential at 22W+.

Yeah I guess that ideally AMD should put out something exactly in-between the Z1 (4CUs) and Z1 Extreme (12CUs) as perf/watt is the most crucial aspect in a handheld imo.

 

[Kataploom]

Sony and MS might start planning a handheld lower power version of their consoles for next gen, Switch is gonna end up eating their spaces if rumors are true since most people don't care about resolution (1440p vs 4K? That's a GAF thing... and not even that much of a thing)

 

[Reallink]

The CPU cores are wildly overpowered in these chips given the context of portable "gaming" devices, they steal way too much power and die space from GPU portion. The Z1 Extreme in 30W mode with a proper fan curve outperforms the Desktop 5800X/5800X3D and is right on the heals of the Desktop 7700's in Cinebench multicore. Meanwhile the GPU barely beats a 1050Ti and is signficantly behind a 1060. Comically lopsided for a handheld gaming device. The Z1 extreme's CPU is significantly faster than the PS5 or Series X's for example, which is stupid as fuck for a portable device whose GPU can't even manage 30fps in most contemporary titles at Low settings.

 

[FearlessBelgian]

You're the same boy or girl from the Switch 2 leak topic, you bring back some ridiculous numbers from a chip probably more powerful than the one the Switch 2 will have and you still think the Switch 2 will be able to offer BOTW in 4K 60 fps... LOL

 

[Tsaki]

"when both types of cores are configured to run at the same 3.2 GHz frequency, their performance becomes nearly identical"
This confused me. Doesn't Zen4c have half the L2 cache?

 

[shamoomoo]

Yeah Sony really has no excuses for not using zen4 in the ps5 pro seeing the designs are mature enough to push out handheld apu chips. I cant help but feel underwhelmed by the actual performance put out by the apu though compared to the hardware specs listed. I cant help but feel this will get embarrassed by the switch 2 chipset based on reports so far.

The pro,if it's real is just a stopgap so there isn't a need for Zen 4 if an enhanced Zen 2 can give a performance uplift to make 60+ fps easier.

 

[Tsaki]

Sony and MS might start planning a handheld lower power version of their consoles for next gen, Switch is gonna end up eating their spaces if rumors are true since most people don't care about resolution (1440p vs 4K? That's a GAF thing... and not even that much of a thing)

A chip that is made to run at handheld wattage in a console that will be sold at profit by Nintendo from day 1 will not come close to PS5/Series X in performance.

 

[Crayon]

The CPU cores are wildly overpowered in these chips given the context of portable "gaming" devices, they steal way too much power and die space from GPU portion. The Z1 Extreme in 30W mode with a proper fan curve outperforms the Desktop 5800X/5800X3D and is right on the heals of the Desktop 7700's in Cinebench multicore. Meanwhile the GPU barely beats a 1050Ti and is signficantly behind a 1060. Comically lopsided for a handheld gaming device. The Z1 extreme's CPU is significantly faster than the PS5 or Series X's for example, which is stupid as fuck for a portable device whose GPU can't even manage 30fps in most contemporary titles at Low settings.

This is my primary issue with these at the moment. The balance of cpu to gpu is way off if you are primarily using it for games.

 

[Ozriel]

Asus needs to shelve any ideas they might have about releasing the z1 model of the Ally. Even at $100 cheaper than the regular z1 extreme, it's a terrible deal

 

[Klosshufvud]

This article is just horribly written.

The CPU cores are wildly overpowered in these chips given the context of portable "gaming" devices, they steal way too much power and die space from GPU portion. The Z1 Extreme in 30W mode with a proper fan curve outperforms the Desktop 5800X/5800X3D and is right on the heals of the Desktop 7700's in Cinebench multicore. Meanwhile the GPU barely beats a 1050Ti and is signficantly behind a 1060. Comically lopsided for a handheld gaming device. The Z1 extreme's CPU is significantly faster than the PS5 or Series X's for example, which is stupid as fuck for a portable device whose GPU can't even manage 30fps in most contemporary titles at Low settings.

Well that has a fairly simple explanation. That GPU is so much more power consuming than CPU to run. Steam Deck for instance has the GPU take up 10W and CPU 4W at a 15W power envelope. Scaling back on CPU cores/clocks don't really result in massive GPU gains.

Also while more GPU power is great, I will say that a weaker CPU can also be a pretty bothersome bottleneck. The 7840U destroys Steam Deck at emulation for instance. SD would probably see some decent perf gains if it had Zen4 cores.

 

[BlackTron]

You're the same boy or girl from the Switch 2 leak topic, you bring back some ridiculous numbers from a chip probably more powerful than the one the Switch 2 will have and you still think the Switch 2 will be able to offer BOTW in 4K 60 fps... LOL

Bookmarked for later see you at the Switch 2 reveal