Rembrandt can go and die.MLIDs leaked slide shows that Phoenix and Rembrandt sre still there in 2025 (so, 9040 and 9030 series), probably as a cheaper gaming option, since they have 12 CUs, and can still compete with intel
Why, it's an alright part, especially when cheap.Rembrandt can go and die.
680M/660M is still pretty good, dunno why you think it's badRembrandt can go and die.
Point accepted.Why, it's an alright part, especially when cheap.
Why, it's an alright part, especially when cheap.
Not trying to attack you, but you do realize that one can exist independently of the other while being owned by the same company, right?In this world where AMD tried to plough through the market with a subpar architecture only to pull themselves up by their bootstraps with Zen I firmly believe they would have had no chance splitting their R&D brainpower let alone funding.
If Globalfoundries can't break through 12nm on their own, but AMD can develop a good arch on their own, logic would dictate one group is more fruitful than the other. If it were easy enough to throw money at litho wouldn't all the other fabs also be on 7-5-3nm right now? Occam's razor says it all when Intel was the only mass market cutting edge 'foundry' for over a decade until 2019.
AMD likely has not released the microcode required to run the chip in third party boards. Zen 5 is still many months from release, after all.You mean the Eng CPU sample you got only runs on a ref MB that you got along with it, and that ref MB does not allow changing max TDP setting from stock? Yeah, in that case there’s no way of knowing.
Then we can only make educated guesses of the perf/watt improvement at 120W TDP for Zen5 vs Zen4, based on e.g. CPU arch and chip process tech changes from Zen4 to Zen5. Anyone that has any guesstimates?
Finally, if you only got Eng sample, there’s no way of knowing other results from final CPU version anyway either. So that should apply for your other Zen5 perf info predictions too, if you’re basing it on Eng sample testing alone.
I agree, but not in the way you may think. Lots of people are attacking what adroc is saying rather than hitting the ignore button like they should. I don’t necessarily like him, but he should be able to speak just like everyone else.This thread is in dire need of mod intervention.
I can agree about Lucienne, but Cezanne? No way. Zen 3 is a great chip.What actually has to be said is this:
Cezanne and Lucienne need to die
I mean, is mobile Phoenix2 actually better than mobile Cezanne in any considerable way? im not sure if being better in ST justifies the price, develop a new board, etc...What actually has to be said is this:
Cezanne and Lucienne need to die
Clocks, a slight efficiency advantage, and IGP are the only differences to my knowledge.I mean, is mobile Phoenix2 actually better than mobile Cezanne in any considerable way? because being just a little bit better in ST im not sure if it justifies the price, develop a new board, etc...
Lucienne is already out, it was only used in 5300U/5500U/5700U Zen2 laptop parts, and that's itWhat actually has to be said is this:
Cezanne and Lucienne need to die
Im not convinced yet that the GPU is better in any meaningful way, it is just too small. But battery life will be impacted yes.Clocks, a slight efficiency advantage, and IGP are the only differences to my knowledge.
If you had to use each of them in a blind test you would not be ably to tell the difference unless you used the IGP or were measuring battery life.
| AMD's U-series | Hawk Point | Kraken Point | Lunar Lake | Strix Point | Strix Halo (Sarlak) |
|---|---|---|---|---|---|
| Date | Q4 2023 | Q1 2025 ? | Q4 2024 | Q3 2024 | Q1 2025 ? |
| Node | N4 | N4P | N3B + ? | N4P | 2 x N4P + N3E |
| Die Size | 178 mm2 | ? | ? | 225 mm2 | ? |
| TDP | 15-30 W | 15-30 W? | 17-30 W | 15-30 W? | 120 W ? |
| CPU | Zen 4 8P | Zen 5+5c 4P+4E | LNC+SKT 4P+4E | Zen 5+5c 4P+8E | Zen 5 16P |
| Threads | 16 | 16 | 8 | 24 | 32 |
| L3 Cache | 16 MB | 16+4 = 20MB ? | 8 MB | 16+8 = 24MB | 16+16 = 32MB |
| Base Freq | 3.3 GHz | ? | ? | ? | ? |
| Turbo Freq | 5.1 GHz | ? | ? | ? | ? |
| GPU | RDNA3 12CU | RDNA3+ 8CU | ARC 8 XE | RDNA3+ 16CU | RDNA3+ 40CU |
| SP | 768 | 512 | 128 | 1024 | 2560 |
| ALU | 768 | 1024 | 1024 | 2048 | 5120 |
| GPU Speed | 2700 MHz | 3000 MHz ? | 1850 MHz | 3000 MHz ? | 2500 MHz ? |
| FP32 TFlops | 4.1 TF | 3 / 6.1 TF | 3.8 TF | 6.1 / 12.3 TF | 12.8 / 25.6 TF |
| AIE (TOPS) | 16 | 50 | 30 ? | 50 | 50 |
It's a shame that AMD doesn't have any chips in anything ultra-portable, like handhelds 😢 /sI just couldn't believe AMD don't have ultra-portable solution to upcoming LNL from Intel....
This could be phrased as "¹fewer ²dense cores" or "¹fewer ²compact cores", for example.Kraken Point essential is a cut down version of Strix Point with lesser¹ e-cores² and half of CU.
Exactly. The only difference in the AMD dense and regular cores is cache size and speed. The performance is close and the usability is identical . Example, they both have avx-512 for AMD. Intel used to have it only on the P cores, but is now totally gone or disabled.I find it irritating when folks take Intel's "P" and "E" terminology and apply it to AMD CPUs. Remember, this terminology denotes two very different lines of core microarchitectures of Intel's. AMD does not mix core architectures in their CPUs, and presumably won't do so either in Zen 5 CPUs. With Zen 4, AMD just introduced two different physical designs of one and the same microarchitecture, and the effects of this are different from what Intel is accomplishing with its P+E products and E-only products.
This could be phrased as "¹fewer ²dense cores" or "¹fewer ²compact cores", for example.
Then the question is, how will AMD compensate having only 2/3 of WGPs(CUs) compared to PHX, assuming this is the successor.View attachment 92503
Kraken Point will have 8 CU
Wrong ALU count and TFlops for Hawk Point.Here comes the competitor of Intel's Lunar Lake, AMD's Kraken Point specificly U -series of Kraken Point. I just couldn't believe AMD don't have ultra-portable solution to upcoming LNL from Intel....turn up Kraken Point U series which are supposedly launching in Q1 2025 is the one. I have listed down the leaked specs and compare to Lunar Lake, seems like Kraken Point wins in almost every parts:-
AMD's U-series Hawk Point Kraken Point Lunar Lake Strix Point Date Q4 2023 Q1 2025 ? Q4 2024 Q3 2024 Node N4 N4P N3B + ? N4P Die Size 178 mm2 ? ? 225 mm2 TDP 15-30 W 15-30 W? 17-30 W 15-30 W? CPU Zen 4 8P Zen 5+5c 4P+4E LNC+SKT 4P+4E Zen 5+5c 4P+8E Threads 16 16 8 24 L3 Cache 16 MB 16+4 = 20MB ? 8 MB 16+8 = 24MB Base Freq 3.3 GHz ? ? ? Turbo Freq 5.1 GHz ? ? ? GPU RDNA3 12CU RDNA3+ 8CU ARC 8 XE RDNA3+ 16CU SP 768 512 128 1024 ALU 768 1024 1024 2048 GPU Speed 2700 MHz 3000 MHz ? 1850 MHz 3000 MHz ? FP32 TFlops 4.1 TF 6.1 TF 3.8 TF 12.3 TF
- LNL has one advantage is ability to scale down to 9W with fanless design. However, in PC market I am not sure that make business sense. Even ZenBook Duo comes with two fans to cool down MTL-H CPU.
- Kraken Point essential is a cut down version of Strix Point with lesser e-cores and half of CU. There are going to have H and U versions. U-series with lower TDP should compete well against LNL with lower clock speed.
- With HT, Kraken Point should have double threads compared to LNL. As for GPU performance, it depends on AMD RDNA3+ DI implementation.
- It is rough comparison table, but at least we know how AMD's answer to Intel's LNL
- I thought of putting ARL-U to compare with STX, but the specs are so low I am not sure what is the points . IFS really needs to catch up...
Debugged RDNA3.5 ?Then the question is, how will AMD compensate having only 2/3 of WGPs(CUs) compared to PHX, assuming this is the successor.
With CPUs It's clear -> higher IPC, but what about IGP?
Exactly. The only difference in the AMD dense and regular cores is cache size and speed. The performance is close and the usability is identical . Example, they both have avx-512 for AMD. Intel used to have it only on the P cores, but is now totally gone or disabled.