Massive AMD Zen 5 IPC gain of 25%+ reported as leak alleges increased L1 cache and revamped L3 cache for Ryzen 8000 CPUs

[Redaktion]

Paul from RedGamingTech is at it again with another round of leaks concerning AMD's Zen 5 processors that are expected to debut sometime in 2024. Per the leaker, the Ryzen 8000 CPUs could bring north of 25% IPC uplift, a larger L1, 1 MB L2, and a new "Ladder" L3 cache.

https://www.notebookcheck.net/Massive-AMD-Zen-5-IPC-gain-of-25-reported-as-leak-alleges-increased-L1-cache-and-revamped-L3-cache-for-Ryzen-8000-CPUs.713355.0.html

[NikoB]

Actually, the whole value of these rumors about Zen 5 (when the false AMD still cannot release Zen4 Phoenix laptops, although it is already the end of April 2023) is that AMD admits to its significant technological loss to Apple by more than 2.5 years !!!

256-bit memory controller (moreover, soldered, like Apple, and therefore significantly limited in volume, and we need ALREADY 64GB minimum, if the built-in video chip requires 16-24GB for share memory - otherwise what is their point in games?), in 2 times slower than Apple's 512-bit memory controller in the M2 Max, which is already in production and on sale.

At the same time, GDDR6x video memory in NVidia solutions is still 2-5 times faster, which means that there will be no performance level of GTX4070, a priori. At best, some "GTX4050Ti". Sometime in AMD's distant hazy future....

The shameful 128-bit memory controller in the Zen4 7945HX cannot even physically handle the 28 pci-e 5.0 lanes that AMD integrated there for some reason, instead of the integrated graphics of the Zen4 Phoenix level (which we have not even seen in any review), even without Taking into account the integrated video card, chips of the 7945HX level already need memory with a bandwidth of at least 400Gbyte/s, at least 200. But in reality, the slow 7945HX memory controller has a shameful bandwidth of only 60-65Gbyte/s! It's like putting tractor wheels on a Formula 1 car...

The entire x86 industry is in a dead end - the memory bandwidth ALREADY should be at least 3 times more than its best solutions in the consumer mobile segment (and this is no more than 85Gbyte/s on average and only for Intel, not for AMD, whose memory controllers are traditionally lose to Intel controllers by about x1.3-x1.5) to cater for all ports and technologies...

Apple needs to finish off the x86 camp by making its Arm platform freely available to all laptop manufacturers. Only in this way will it be able to turn the tide with x86's overwhelming market share and overwhelming software market share. Someone at Apple, as brave as Jobs once was, must finally make this decision.

[Apple loses every day]

Actually, the whole value of these rumors about Zen 5 (when the false AMD still cannot release Zen4 Phoenix laptops, although it is already the end of April 2023) is that AMD admits to its significant technological loss to Apple by more than 2.5 years !!!

256-bit memory controller (moreover, soldered, like Apple, and therefore significantly limited in volume, and we need ALREADY 64GB minimum, if the built-in video chip requires 16-24GB for share memory - otherwise what is their point in games?), in 2 times slower than Apple's 512-bit memory controller in the M2 Max, which is already in production and on sale.

At the same time, GDDR6x video memory in NVidia solutions is still 2-5 times faster, which means that there will be no performance level of GTX4070, a priori. At best, some "GTX4050Ti". Sometime in AMD's distant hazy future....

The shameful 128-bit memory controller in the Zen4 7945HX cannot even physically handle the 28 pci-e 5.0 lanes that AMD integrated there for some reason, instead of the integrated graphics of the Zen4 Phoenix level (which we have not even seen in any review), even without Taking into account the integrated video card, chips of the 7945HX level already need memory with a bandwidth of at least 400Gbyte/s, at least 200. But in reality, the slow 7945HX memory controller has a shameful bandwidth of only 60-65Gbyte/s! It's like putting tractor wheels on a Formula 1 car...

The entire x86 industry is in a dead end - the memory bandwidth ALREADY should be at least 3 times more than its best solutions in the consumer mobile segment (and this is no more than 85Gbyte/s on average and only for Intel, not for AMD, whose memory controllers are traditionally lose to Intel controllers by about x1.3-x1.5) to cater for all ports and technologies...

Apple needs to finish off the x86 camp by making its Arm platform freely available to all laptop manufacturers. Only in this way will it be able to turn the tide with x86's overwhelming market share and overwhelming software market share. Someone at Apple, as brave as Jobs once was, must finally make this decision.

You spam and repost this same inaccurate garbage on every AMD and Intel article jfc. ARM will never get the chance to match up to x86's power or compatibility. Ever. Apple is in panick mode knowing their chips are already hitting a performance wall since 7000 series chips and Raptor Lake absolutely smoke them in benchmarks. Apple ONLY has battery life gains on Intel and AMD but that is gonna end soon with 8000 series chips and Meteor Lake (which will be on Intel's newest fabrication process). Apple's walled garden approach for all their platforms will be their demise.

[Neenyah]

256-bit memory controller (moreover, soldered, like Apple, and therefore significantly limited in volume, and we need ALREADY 64GB minimum, if the built-in video chip requires 16-24GB for share memory - otherwise what is their point in games?)

64 gigs minimum, lol? Idk, lower the resolution? No one sane is going to play 4K (or higher) with iGPU. And with a dGPU eating 8-16 GB of VRAM system is still barely using 10-12 GB of RAM in the most demanding games. It's called optimizing your product (on the devs' side).

[Neenyah]

ARM will never get the chance to match up to x86's power or compatibility. Ever. Apple is in panick mode knowing their chips are already hitting a performance wall since 7000 series chips and Raptor Lake absolutely smoke them in benchmarks. Apple ONLY has battery life gains on Intel and AMD...

Very true. On low loads and with a very limited range of software/apps Apple's M1s and M2s are faster than Intel's and AMD's offerings but once you include all needed apps for normal regular workflow those little ARM chips gets destroyed by even i5s and Ryzens 5. And whatever some Apple's fan(atic) wants to say - no, performance per Watt is superior on x86.   
 
Apple's idle/sleep game is unmatched though and the same goes for that previously mentioned low loads - if you just watch movies and read/write stuff and emails then the battery is going to last a hell lot. But once you start to actually do work to, you know, earn money, Intels and AMDs are faster and more efficient, not just in benchmarks.

[Vovan.]

Actually, the whole value of these rumors about Zen 5 (when the false AMD still cannot release Zen4 Phoenix laptops, although it is already the end of April 2023) is that AMD admits to its significant technological loss to Apple by more than 2.5 years !!!

256-bit memory controller (moreover, soldered, like Apple, and therefore significantly limited in volume, and we need ALREADY 64GB minimum, if the built-in video chip requires 16-24GB for share memory - otherwise what is their point in games?), in 2 times slower than Apple's 512-bit memory controller in the M2 Max, which is already in production and on sale.

At the same time, GDDR6x video memory in NVidia solutions is still 2-5 times faster, which means that there will be no performance level of GTX4070, a priori. At best, some "GTX4050Ti". Sometime in AMD's distant hazy future....

The shameful 128-bit memory controller in the Zen4 7945HX cannot even physically handle the 28 pci-e 5.0 lanes that AMD integrated there for some reason, instead of the integrated graphics of the Zen4 Phoenix level (which we have not even seen in any review), even without Taking into account the integrated video card, chips of the 7945HX level already need memory with a bandwidth of at least 400Gbyte/s, at least 200. But in reality, the slow 7945HX memory controller has a shameful bandwidth of only 60-65Gbyte/s! It's like putting tractor wheels on a Formula 1 car...

The entire x86 industry is in a dead end - the memory bandwidth ALREADY should be at least 3 times more than its best solutions in the consumer mobile segment (and this is no more than 85Gbyte/s on average and only for Intel, not for AMD, whose memory controllers are traditionally lose to Intel controllers by about x1.3-x1.5) to cater for all ports and technologies...

Apple needs to finish off the x86 camp by making its Arm platform freely available to all laptop manufacturers. Only in this way will it be able to turn the tide with x86's overwhelming market share and overwhelming software market share. Someone at Apple, as brave as Jobs once was, must finally make this decision.

I understand your frustration but please don't give up hope completely on x86 just yet, NikoB. AMD are working on a larger die Apple M-Series competitor chip (codename: Sarlak / Strix Halo) with 256-bit LPDDR5X(-8533?) memory controller.

In terms of bandwidth, it's estimated to give around 270 GB/s which is still way below GDDR6X and the 400 GB/s in the 512-bit M2 Max. However, it'll come with 32 MB of infinity cache (along with up to a 40 CU RDNA3+ iGPU) which should give a massive boost and help in higher resolution gaming. Should be able to deliver RTX 3060-3070 Ti laptop dgpu performance but with like 5-6 hr battery life while gaming instead of 90 min.

It's expected to come H2 2024 but knowing AMD, just think more like early 2025. Yeah, another 2.5 years late but better late than never I suppose. (also keep in mind Qualcomm plans to release their Nuvia designed cpu core SoC in 2024 aswell - so they too are very late to the party)

While massive 512-bit to 1024-bit memory controller's are nice, you've to understand they increase the cost of the chip significantly. These costs just end up being passed down to consumers. I don't think under current climate this makes much financial sense (economic recession / RTX 4090 laptops already approaching $5000 in the EU - not too different in pricing from an M2 Max Macbook Pro)

The only reason why Apple is able to pull off what they do is because they've boatload of cash and due to having a very special relationship with TSMC, allowing them to negotiate better pricing on a custom node process which is exclusive to them. AMD came from a very different place, being in debt/bankrupt for almost over a decade. They literally just started making money now from the last few years of Epyc cpu server/data center sales.

Regarding Apple making it's platform available to everyone. This will never happen as they've never cared about anyone except for their own ecosystem. Also, ARM isn't free or very open - just look at the recent lawsuits. If I had to place any bets on a next gen cpu arch, it'd be RISC-V. Jim Keller, the guy who worked on the original zen core and A4/A5 is currently working at tenstorrent (a RISC-V startup). But we're still a very long way off from seeing any mainstream RISc-V devices.

[Vovan.]

As good as Apple hardware is, it's pretty much useless outside of being very optimized for their 1st party software, like final cut pro. No 3rd party dev wants to develop on metal. There's no games.

Windows on ARM emulation is still trash.

So despite taking forever to release their stuff, AMD is still pretty much the only option.

[Le quads]

Imo this is the right direction for Chips. Higher clock speeds arent as necessary unless its strictly for gaming, but if you want to boost the 90% of other tasks your computer will be doing this is the way to do it. Look at the efficiency and performance gains from the Apple M series chips that are doing way more low level cache per core. You can have the best of both world and have 3D Vcache + more low lvl cache for huge gains in nearly every sector. If you need higher clocks you can just adjust the cache, but theres a new segment for high efficiency & high performance at the same time.

[Nate]

Apple needs to finish off the x86 camp by making its Arm platform freely available to all laptop manufacturers.

Finish off x86? Apple's hardware is intended to run Apple software, making it totally useless to me. It's not even a consideration for me, and Apple has no reason to try to compete in high end desktops or gaming.

[rs]

Actually, the whole value of these rumors about Zen 5 (when the false AMD still cannot release Zen4 Phoenix laptops, although it is already the end of April 2023) is that AMD admits to its significant technological loss to Apple by more than 2.5 years !!!

Nonsense. Apple only had the advantage of the 5nm process. But that wasn't their effort but TSMC's. AMD also has to design more cost efficient because they offer products on more markets. Zen 2-4 were quite compact for certain reasons. If they wanted they also could have done a wide design as M1. But that wouldn't have made any sense. Zen 2 to 4 are impressive enough in their very own way.

The entire x86 industry is in a dead end - the memory bandwidth ALREADY should be at least 3 times more than its best solutions in the consumer mobile segment (and this is no more than 85Gbyte/s on average and only for Intel, not for AMD, whose memory controllers are traditionally lose to Intel controllers by about x1.3-x1.5) to cater for all ports and technologies...

You really have no clue. AMD pioneered the memory controller by integrating it into the processor. When Apple still used old PowerPC and Intel only had slow FSB.

Apple needs to finish off the x86 camp by making its Arm platform freely available to all laptop manufacturers.

Sure. Because that's what Apple is known for, an open ecosystem. 🤦�♂ Apple is worse than x86. And ARM isn't any better than x86. Only somewhat less closed, but still proprietary. Right now the only reasonable solution to replace x86 in the long run is RISC-V.