Thursday, April 25th 2024
AMD "Strix Point" Mobile Processor Confirmed 12-core/24-thread, But Misses Out on PCIe Gen 5
AMD's next-generation Ryzen 9000 "Strix Point" mobile processor, which succeeds the current Ryzen 8040 "Hawk Point" and Ryzen 7040 "Phoenix," is confirmed to feature a CPU core-configuration of 12-core/24-thread, according to a specs-leak by HKEPC citing sources among notebook OEMs. It appears like Computex 2024 will be big for AMD, with the company preparing next-gen processor announcements across the desktop and notebook lines. Both the "Strix Point" mobile processor and "Granite Ridge" desktop processor debut the company's next "Zen 5" microarchitecture.
Perhaps the biggest takeaway from "Zen 5" is that AMD has increased the number of CPU cores per CCX from 8 in "Zen 3" and "Zen 4," to 12 in "Zen 5." While this doesn't affect the core-counts of its CCD chiplets (which are still expected to be 8-core), the "Strix Point" processor appears to use one giant CCX with 12 cores. Each of the "Zen 5" cores has a 1 MB dedicated L2 cache, while the 12 cores share a 24 MB L3 cache. The 12-core/24-thread CPU, besides the generational IPC gains introduced by "Zen 5," marks a 50% increase in CPU muscle over "Hawk Point." It's not just the CPU complex, even the iGPU sees a hardware update.Apparently, AMD is increasing the workgroup processor (WGP) count of the iGPU from 6 on the current "Hawk Point" processor, to 8 on "Strix Point." This works out to 16 compute units, or 1,024 stream processors, making a 33% increase in the shader engine's performance. The new iGPU is based on the updated RDNA 3+ graphics architecture. "Strix Point" also debuts AMD's 2nd Generation Ryzen AI NPU based on the XDNA 2 architecture. This NPU offers 50 AI TOPS in performance, an over 3-fold increase from the 16 TOPS offered by the NPU on "Hawk Point."
In terms of I/O, one can expect an updated display engine with DisplayPort 2.1 UHBR10, DSC, and support for 8K @ 60 Hz with a single cable. The memory interfaces are expected to remain unchanged spare for increased reference speeds, with support for DDR5 and LPDDR5(x) memory types. One area for disappointment is the PCIe interface. We had expected "Strix Point" to feature PCIe Gen 5, but it seems like AMD had other plans. The PCIe interface of "Strix Point" will be similar to that of "Phoenix" and "Hawk Point," it will stick to PCIe Gen 4. While this might not mean much for discrete GPUs, it would mean that you can't use the latest Gen 5 NVMe SSDs at their advertised speeds.
The same specs sheet also confirms a lot of specs of the larger "Strix Halo" chiplet-based mobile flagship processor, you can read all about it in our older article.
Source:
HKEPC
Perhaps the biggest takeaway from "Zen 5" is that AMD has increased the number of CPU cores per CCX from 8 in "Zen 3" and "Zen 4," to 12 in "Zen 5." While this doesn't affect the core-counts of its CCD chiplets (which are still expected to be 8-core), the "Strix Point" processor appears to use one giant CCX with 12 cores. Each of the "Zen 5" cores has a 1 MB dedicated L2 cache, while the 12 cores share a 24 MB L3 cache. The 12-core/24-thread CPU, besides the generational IPC gains introduced by "Zen 5," marks a 50% increase in CPU muscle over "Hawk Point." It's not just the CPU complex, even the iGPU sees a hardware update.Apparently, AMD is increasing the workgroup processor (WGP) count of the iGPU from 6 on the current "Hawk Point" processor, to 8 on "Strix Point." This works out to 16 compute units, or 1,024 stream processors, making a 33% increase in the shader engine's performance. The new iGPU is based on the updated RDNA 3+ graphics architecture. "Strix Point" also debuts AMD's 2nd Generation Ryzen AI NPU based on the XDNA 2 architecture. This NPU offers 50 AI TOPS in performance, an over 3-fold increase from the 16 TOPS offered by the NPU on "Hawk Point."
In terms of I/O, one can expect an updated display engine with DisplayPort 2.1 UHBR10, DSC, and support for 8K @ 60 Hz with a single cable. The memory interfaces are expected to remain unchanged spare for increased reference speeds, with support for DDR5 and LPDDR5(x) memory types. One area for disappointment is the PCIe interface. We had expected "Strix Point" to feature PCIe Gen 5, but it seems like AMD had other plans. The PCIe interface of "Strix Point" will be similar to that of "Phoenix" and "Hawk Point," it will stick to PCIe Gen 4. While this might not mean much for discrete GPUs, it would mean that you can't use the latest Gen 5 NVMe SSDs at their advertised speeds.
The same specs sheet also confirms a lot of specs of the larger "Strix Halo" chiplet-based mobile flagship processor, you can read all about it in our older article.
42 Comments on AMD "Strix Point" Mobile Processor Confirmed 12-core/24-thread, But Misses Out on PCIe Gen 5
They're a player in the market, they can influence it. And yes, if AMD would offer performance at a significant discount compared to Nvidia, AMD would see more sales (imagine that!). Why don't they do that? Because GPUs, as we have them today, are not price gauged (at the top), they're simply expensive to build.
Interconnect speeds! 4/800Gbps interconnections.
Storage density. 48 x x4 connected SSDs into a single or double U of space
Accelerator interconnects 4 or 8 x16 Accelerators in one chassis easily with space for storage/interconnects etc as needed.
The things like CXL etc on top of all this.
Where currently in the Consumer market are people really taking advantage of that.
PCI-e 5 SSDs are massively expensive,are hot,big power draw
GPUs are still at PCI-E 4.0
How many peope have >10Gbps internet/networks at home. How many are still on 1Gb switches/routers
PCIe 5.0 will bring advantages to consumers but its the server space at the moment that can really take advantage of it with ease.
Besides, what a single customer wants to buy doesn't affect everyone's choice.
Your argument doesn't even need two brands. Lets say I want to buy a cheap HP laptop. What would happen if EVERYONE wanted to buy the same model?..
You can apply this logic to almost anything but in the end it doesn't work like that, fortunately.
I dont think you understand how prices actually work, but you have one word correct and that is compete. For prices to lower, then AMD and Nvidia need to compete instead of one being a little brother to the other. The same way AMD 'competed' against intel with the Ryzen lineup. Before Ryzen, AMD was certainly second fiddle to Intel, anyone who says otherwise has a strange bias to one corporation. Look how quickly AMD made high core count CPUs standard for computers. Those Intel 10 HEDT units were priced ridiculously because there was not competition from AMD.
For GPU prices to lower, then AMD will need to compete better with Nvidia. The big gap right now is DLSS, Frame Gen, and RTX performance. In those areas AMD is significantly behind in those areas. But that's not to say AMD makes bad GPU's, they are really good, but they cant 'compete' with Nvidia in several areas.
Lastly, GPU prices are not going to suddenly drop because AMD releases a great GPU. In fact, expect GPU prices to increase, and by a lot over the next several years. There is too much use for GPU's in AI and Nvidia and AMD will certainly cater to those markets. What will happen is the demand for GPUs will increase and with no more foundry spaces, the supply will remain the same.
AMD will be fine. They make great products, both GPU's and CPUs. While they are slightly behind Nvidia right now, it's not by a lot and AMD can make up a lot of ground in upscaling resolution with AI cores in future GPUs.
A 4090 running on PCIe 3.0 loses single digit performance even at PCIe 2.0 is less than 10% worst case
www.techpowerup.com/review/nvidia-geforce-rtx-4090-pci-express-scaling/
How many of you here are going to really see/feel the difference a 4Gb/s SSD and an 8GB/s SSD.......... now when you consider that "sacrificing" things like this may actually give you an extra hour of battery life out of your laptop for a single SSD etc its more and more sensible.
So thinking that a piece of silicon designed with power constraints in mind it makes PERFECT sense to not include PCIe 5.0 as all you are doing is commiting power draw and heat for litearlly 0 benefit in situations this silicon would be deployed/used in.
It's not like consumers can't have it, just pick a CPU that got it! :rolleyes: Like Raphael, which makes me guess that Dragon range has it as well, tho I haven't looked it up.
There's no point in having PCIE 5 in EVERY CPU/APU (yet), just because it's possible.
Also Gen 5 SSDs are useless in desktops as well. PCIe is going faster and faster to be more useful in datacenters and super computers. They are already busy with Gen6 and Gen7.
You don't even need Gen4 SSDs for that matter. At this time yes PCIe 4 is more than enough for consumers. The only reason for it is actually Direct Storage that benefits from it and for that a PCIe 4 SSD is more than sufficient.
Consumers really have 0 need for PCIe 5.0 unless you want everything to use PCIe 4x connections. the RTX 4090 barely benefits from a PCIe 3.0 x16 connection over an 8x one. SSDs only for their copying and really how often do you need to copy massive files that you can't wait a few seconds longer for. It does nothing in speeding up level loading unless it supports Direct Storage which are at this time only Forspoken, Ratchet and Clank: Rift Apart, Horizon Forbidden West, and Forza Motorsport.
Gen 5 just makes your SSD a power hungry beast. I mean 15 watt for writing is no joke and has nothing to do in a laptop. Gen 4 SSDs can work with a third of that power. a Gen 3 SSD hardly registers the 2watt mark.
With your quote standing I reckon you have a 400Gbit network running at home as well. Because well why stay at 1 gbit and your 5GHz wifi spot when there is also wifi 7. Not to mention have a 128 Core CPU because.... more cores and not to forget 2TB of RAM.
Maybe now you might understand where some of us are getting at.