- Mar 3, 2017
- 1,677
- 6,159
- 136
This is comparing both options with two CCDs.There's always a tradeoff. Cost/power/interconnect penalties etc. AMD has demonstrated the ability to add an extra CCD without paying too high a price for it, other than tacking on a higher MSRP.
So the idea is to replace the second CCD with Zen5c? Questionable. That kind of slams the door on people (eventually) coding games etc. that might benefit from more than 8 cores.This is comparing both options with two CCDs.
8P + 24D or 8P + 32D would be a lot more enticing.I too would like to see AMD offer 8P + 16D at the same time as 8P + 8P, so that we can finally see for ourselves what part of the desktop consumer market is willing to buy dense cores over performance cores.
The number quoted was not just raw silicon cost, but also packaging, etc. I will admit the number may be off (haven’t had time to dig into it), but even if you use half that as a baseline, you are still looking at a considerable markup.That would put the waffer at roughly 50k$, wich wasnt the case, cost is rather in the 30$/CCD at the die level.
Meh, they could just rename Threadripper to EPYC Workstation, introduce a new socket that supports quad channel memory and 32 PCE Gen 5 lanes, then sell Threadripper 8950x, 8960X, and 8970X with 16, 24, and 32 cores @5.6ghz peak. Also release X3D versions. Those chips could occupy the $600-$1,500 price point.8P + 24D or 8P + 32D would be a lot more enticing.
Heck, even just 32D with V-cache could be pretty compelling.
32D with V-cache AND 32GB or 48GB LPDDR5X + RDNA4 = Apple M3 Pro competitor!
In desktop efficient cores are not needed unless they massively increased the core count. 7950X manages ~5240MHz using 32 threads at 230W.That's a little surprising. For me, I want more high-clockspeed cores wherever possible. 16c is already massive overkill for most non-workstation/server builds, and just adding more cores on top of that with "efficiency" cores seems stupid outside of maybe power-constrained scenarios.
…you mean like the 7945hx? Base clocks are a bit slower, but a node shrink would fix that.In desktop efficient cores are not needed unless they massively increased the core count. 7950X manages ~5240MHz using 32 threads at 230W.
Even at 120W in CB It manages ~4327MHz or 3839MHz at ~100W.
On the other hand, in laptop It could be more interesting.
The question is what is the highest frequency where It's still more efficient than a standard one. 3GHz or 3.5GHz?
I wouldn't mind a 16C32T clocked at 3.5GHz consuming only 45W at full load.
If we are still talking about AM5 as context, two CCDs appear to be the max on the current package, so your config is impossible. In laptops I guess we'll first see how popular 2 CCDs Dragon Range really is. The more cores the more niche the product will be.8P + 24D or 8P + 32D would be a lot more enticing.
Do we know yet whether v-cache is even supported on the Zen 4c CCDs?Heck, even just 32D with V-cache could be pretty compelling.
As I understood from folks on this forum, v-cache and dense cores don't play well together, in the sense that the necessary vias are probably sacrificed to increase density.Do we know yet whether v-cache is even supported on the Zen 4c CCDs?
On AMD's webpage, the base clock is 2.5GHz. That's not just a bit slower.…you mean like the 7945hx? Base clocks are a bit slower, but a node shrink would fix that.
Why would it close that door? There's nothing about gaming that demands the same performance on every thread, and there are even some games today that will make use of Intel's E-cores.So the idea is to replace the second CCD with Zen5c? Questionable. That kind of slams the door on people (eventually) coding games etc. that might benefit from more than 8 cores.
Coding games, not playing them ^Why would it close that door? There's nothing about gaming that demands the same performance on every thread, and there are even some games today that will make use of Intel's E-cores.
8P + 24D or 8P + 32D would be a lot more enticing.
Heck, even just 32D with V-cache could be pretty compelling.
32D with V-cache AND 32GB or 48GB LPDDR5X + RDNA4 = Apple M3 Pro competitor!
It’d absolutely work fine on a Zen 5 CPU that has both a normal CCD and a CCD with dense cores.Coding games, not playing them ^
The thing with Zen5c is that it's the same IPC core with less L3, and running at lower clock. E cores on the other hand are way weaker that P cores IPC wise, while having less cache and running at lower clocks. Zen5 hybrid chips would run MTed games much better than any Intel hybrid chip would.It’d absolutely work fine on a Zen 5 CPU that has both a normal CCD and a CCD with dense cores.
There’s a ton of games that utilize e-cores with great performance (The Last of Us & Jedi Survivor are recent ones). Both of those games have e-cores running tasks that aren’t latency critical (streaming assets, decompression, etc). For a modern multi threaded game each thread is given a priority and then handled by the scheduler appropriately.
That makes even less sense. You want much higher MT performance for game development, whether you're rendering, compiling, running multiple instances, etc. then you do for just playing games. And dev tasks (outside of some audio production workloads) are less likely to be latency sensitive.Coding games, not playing them ^
Depends entirely on how ST vs MT sensitive those lower priority tasks are. And for that matter, how Intel's E-cores vs AMD's dense cores evolve over time.The thing with Zen5c is that it's the same IPC core with less L3, and running at lower clock. E cores on the other hand are way weaker that P cores IPC wise, while having less cache and running at lower clocks. Zen5 hybrid chips would run MTed games much better than any Intel hybrid chip would.
I'm not making a claim that the Intel e-cores are more performant. I'm just saying that having a CCD with Zen 5C cores wouldn't wreck its ability to be a solid gaming CPU since there are games out right now that can take advantage of a heterogeneous CPU.The thing with Zen5c is that it's the same IPC core with less L3, and running at lower clock. E cores on the other hand are way weaker that P cores IPC wise, while having less cache and running at lower clocks. Zen5 hybrid chips would run MTed games much better than any Intel hybrid chip would.
If the 7950X3D has taught us anything, it's that allowing threads to stray onto the slower CCD makes the entire game slower. If there are 16-32 dense cores running at half the clocks (or worse) as the main CCD, the last thing you're gonna want is any game from 2023 mistakenly using that second CCD! Unless something major changes soon in how game engines work, I don't expect that situation to change much.Why would it close that door? There's nothing about gaming that demands the same performance on every thread, and there are even some games today that will make use of Intel's E-cores.
Huh? The 7950X3D performs sub-optimally when either the game's performance critical threads are split across multiple CCDs (die to die communication overhead), or the game is statically bound to the wrong CCD. Changing one die to dense cores doesn't impact that at all. You still want the game primarily bound to one CCD, and it's actually simpler as that one is clearly faster in ST under all circumstances.If the 7950X3D has taught us anything, it's that allowing threads to stray onto the slower CCD makes the entire game slower. If there are 16-32 dense cores running at half the clocks (or worse) as the main CCD, the last thing you're gonna want is any game from 2023 mistakenly using that second CCD! Unless something major changes soon in how game engines work, I don't expect that situation to change much.
The 13900k for example can not touch my 7950x's in productivity in what I do. Not to mention the 7950x can use avx-512 which I also use. You comment only applies to the types of applications that the hybrid approach works better on. NOT ALL PRODUCTIVITY APPS. Please keep your Intel specific comments in Intel threads.Huh? The 7950X3D performs sub-optimally when either the game's performance critical threads are split across multiple CCDs (die to die communication overhead), or the game is statically bound to the wrong CCD. Changing one die to dense cores doesn't impact that at all. You still want the game primarily bound to one CCD, and it's actually simpler as that one is clearly faster in ST under all circumstances.
And in practice, the issue you describe doesn't happen on Intel hybrid systems, despite being rather more complex. You don't see any game inexplicably running on E-cores, do you? No reason to believe AMD would have it worse.
Moreover, the top SKUs are not really for gaming. If you just want to game, get the 9800X3D or whatever and call it a day. The top SKUs are good for productivity, and E-peen, both of which will benefit from a hybrid approach.
There is nothing Intel-specific about what I wrote. I've shown you benchmarks before, and will not waste the time to repeat myself. Productivity apps make good use of hybrid, and there's no reason to believe that would not apply to an AMD hybrid offering.The 13900k for example can not touch my 7950x's in productivity in what I do. Not to mention the 7950x can use avx-512 which I also use. You comment only applies to the types of applications that the hybrid approach works better on. NOT ALL PRODUCTIVITY APPS. Please keep your Intel specific comments in Intel threads.
The faster CCD won't reach max frequency when all cores are maxed out though. Ideally at that point the slower CCD fits right in, likely managing a higher frequency for additional cores, possibly even for the same amount of cores in case of dense Zen.
Huh? The 7950X3D performs sub-optimally when either the game's performance critical threads are split across multiple CCDs (die to die communication overhead)
, or the game is statically bound to the wrong CCD.
Moreover, the top SKUs are not really for gaming.