Interesting. Zen5c looks less dense relative to Zen5 than Zen4c is to Zen4. And it Feels like it's not a coincidence that Zen5c and Zen5 are pretty much the same the same width now. But since we know the c and non-c cores can coexist in the same CCX (eg. Pheonix 2), and CCXs can now go up to 16 in the Zen5 generation, I just have a hard time understanding this layout.
On the one hand, why not put all 12 cores on the same CCX?
And on the other, why not turn those middle Zen5c cores around and place them on the 16MB CCX with the big cores? Better performance, especially for games that scale well to 8 cores, which is a lot of them. Meanwhile the remaining 4 Zen5c cores on the 8MB CCX would still be more than sufficient for light and background tasks.
Depends if you think AI is a dog.
There's almost no reviews, and every current Computex article is a AI muffle mouth mess. Read the comments on the Intel XE2 article, other readers noted it straight away.
Apparently there’s frame gen turned on or something. Results seem too good to be true, because the results put the 7900XTX into 4090 territory I think.
Apparently there’s frame gen turned on or something. Results seem too good to be true, because the results put the 7900XTX into 4090 territory I think.
Zen5 should be looked at from a different perspective. It is a completely redesigned and expanded Front-End. Front-end has 2x greater instruction throughput.
Zen 5 is another, after Zen1, thoroughly redesigned and expanded core. This is the starting point and gateway to higher IPC in subsequent generations. Now I understand what Mike Clark meant about Zen5, although many are disappointed with only a 16% increase in average IPC.
As I wrote earlier, a breakthrough in core design does not necessarily mean a breakthrough in IPC profit.
Well, I tried for almost 2 days now to decide on what to say about Zen 5's results.
I still don't know what to say, so I'm throwing my hands in the air.
Wait for reviews for details. This is a very strange, and very disappointing affair.
Zen5 should be looked at from a different perspective. It is a completely redesigned and expanded Front-End. Front-end has 2x greater instruction throughput.
Zen 5 is another, after Zen1, thoroughly redesigned and expanded core. This is the starting point and gateway to higher IPC in subsequent generations. Now I understand what Mike Clark meant about Zen5, although many are disappointed with only a 16% increase in average IPC.
As I wrote earlier, a breakthrough in core design does not necessarily mean a breakthrough in IPC profit.
Well, I tried for almost 2 days now to decide on what to say about Zen 5's results.
I still don't know what to say, so I'm throwing my hands in the air.
Wait for reviews for details. This is a very strange, and very disappointing affair.
Hey Mahboi, remember that meme video you made about Zen5 before Computex?
You included the scene of the Zen5 hype train crashing. Initially I was doubtful about showing this, because what if Zen5 did indeed live upto the hype and brought 40% ST improvement?
Well, I tried for almost 2 days now to decide on what to say about Zen 5's results.
I still don't know what to say, so I'm throwing my hands in the air.
Wait for reviews for details. This is a very strange, and very disappointing affair.
IC: Finally, what should AMD users look forward to?
MC: It's going be great! I wish I could tell you of all what's coming. I have this annual architecture meeting where we go over everything that's going on, and at one of them (I won't say when) the team and I went through Zen 5. I learned a lot, because of nowadays as running the roadmap, I don't get as close to the design as I wish I could. Coming out of that meeting, I just wanted to close my eyes, go to sleep, and then wake up and buy this thing. I want to be in the future, this thing is awesome and it's going be so great - I can't wait for it. The hard part of this business is knowing how long it takes to get what you have conceived to a point where you can build it to production.
Hey Mahboi, remember that meme video you made about Zen5 before Computex?
You included the scene of the Zen5 hype train crashing. Initially I was doubtful about showing this, because what if Zen5 did indeed live upto the hype and brought 40% ST improvement?
Someone also told me that they "didn't know if it's to mock the 50% believers or the 5% believers", for some reason. That vid was not meant to mock anyone. It's my honest representation of every AMD hype train. It goes too fast, goes too hard, and loses all control. I fit it (the whole Gamer Gitler part) to the Zen 5 situation, but it's more or less applicable in every single hype train.
It's meant to be read the other way actually: you have "Doc AMDHype", "Clara Nonbeliever", and the protagonist/reader/us, and the Train itself. Doc Hype is meant to hype you, Nonbeliever is meant to be the linchpin (can we convince her or not), and they're both amping the intensity of the Train ride.
The Train loses control and crashes every time. It doesn't crash because it's meant to fail or succeed, it crashes because it has to crash. Imagine if Zen 5 was everything untrustworthy leakers claimed it would be: we'd be very happy, sure. But the Hype train ends anyway. The numbers are final, the fantasy has become reality, we can't apply more or less imaginary expectations anymore. The train crashes on the goal every single time, in this case Computex, and disappears into the void of history, while the protagonist jumps forward to Reality, past the Hype Train.
Doc can convince Nonbeliever and they can fly away happy together. Or maybe he can't. It doesn't matter. The train crashes either way, with the protagonist that traverses through the hype and survives to see the result.
Although I hate to use quantic metaphors because quantic is hell, as long as the product isn't observed, the Train accelerates. When the product is out, whether it is everything projected or fantasized by the Train or a total bust, the Train ends its mad rush and crashes. The fantasy hits reality (the END OF LINE sign), the protagonist leaves the Hype Train, the quantic equation resolves in a flash of lightning as fantasy dissociates from reality.
It's the same mental logic that creates this football fan mentality that some have where you "gotta support your semiconductor club". People don't go LETS GO JENSEN because they just love the RTX 3080 so dang much, they support him because they fantasize the possibilities that'll come with the 3080 and beyond it.
If Zen 5 was all 45% improvement, we still wouldn't be fantasizing about all the magical things that it would bring. We'd just have a number.
And past the hype train fun, to tell you how little I care about Zen 5, even if it was a 70% improvement, my 5600x runs every single app I need flawlessly and all that extra speed wouldn't get me to spend 200€ on a new CPU any time soon. Hype trains are a very fun ride. But I never cared about having a 9600X in the first place.
I disagree. I think it's more serious than that. I tried yapping my thoughts on it for like 30 mins over 2 days but couldn't get past just lining up all the oddities with this result.
I don't think it's "just mid", there's a lot of little things that I feel don't make much sense.
Not the least of which is the fact that pricing isn't out and that Lion Cove is just a nudge below, giving maybe 5% more over the RPL vs Zen 4 situation. It's like AMD aimed juuuuuuust slightly above their Z4/RPL situation.
Skymont being a glorious success on paper is great, but I'll be waiting for actual benchmarks on it. That "38% better on workloads" thing isn't reassuring.
More than ever:
But if all the numbers prove accurate, while my AMD fanboi self is heavily disappointed, it is actually an incredibly glorious situation for people. I was genuinely afraid that we'd see Inverted Skylake for 5 years, instead, it seems we've reached a near perfect balance with AMD having high perf server, high perf mobile, and gaming, while Intel has low power mobility and multicore perf.
The competition may prove extreme, and it's all the better for the whole of us. It may be the absolute best result for consumers.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.