Thanks Kepler!Medusa Desktop
no.Is the news of 4WGP/8CU RDNA 4 on the IOD true for Olympic Ridge?
no,If true, alongside the improved memory controller, AMD will be unifying the Ryzen X series and the G series into a single product.
That makes more sense given previous rumors.no.
no,
They're gonna reuse the CCD but SOC tile cost targets are way way way different.
no that's a completely schizophrenic concept to begin with.I thought it would be a unification of the G series into the standard line-up.
I don't think soIs the news of 4WGP/8CU RDNA 4 on the IOD true for Olympic Ridge?
Its most likely a base Mx comp SKU. gotta dethrone the macbook air and Microsoft does want that
no that's a completely schizophrenic concept to begin with.
DT/luggable parts have dGFX to mooch off.
DelusionaldGFX are petty much ultra premium item and on their way to extention at this point, they need to start providing faster starting IGP perf.
Unless you've tried *really* hard to implement both ISA, it's mostly impossible to conclude definitely in a purely objective way.An Interview with Zen Chief Architect Mike Clark
Zen is one of the most important microarchitectures in the history of the x86 ecosystem.www.computerenhance.com
Mentions not seeing much difference/serious advantage from hypothetically going ARM. According to Clark, the clocks/performance/power qualities of x86 versus ARM (basically the higher clocks, higher power "nature" of x86) SoCs mostly have to do with markets that they target, not with the ISA.
well they did.Unless you've tried *really* hard to implement both ISA, it's mostly impossible to conclude definitely in a purely objective way.
They didn't release anything (for good reasons, but that's another discussion). So in my book, there's no proof.well they did.
K12 was Zen1 but aa64.
They did very much work on it tho so idk your point.They didn't release anything
You think AMD faked the K12 custom ARM core? That's a "Stanley Kubrick faked the moon landing" level conspiracy theory.They didn't release anything (for good reasons, but that's another discussion). So in my book, there's no proof.
Intel also does a lot of work and loses to AMD in many benchmarks, and that's only of x86 vs x86. Do you really think Intel engineers are incompetent? No one knows for sure how K12 Arm would have behaved.They did very much work on it tho so idk your point.
Was it released?You think AMD faked the K12 custom ARM core? That's a "Stanley Kubrick faked the moon landing" level conspiracy theory.
No, as I'm sure you know, K12 was canceled.Was it released?
like Zen but aa64.No one knows for sure how K12 Arm would have behaved.
This is what Jim Keller was saying too. It goes both ways. ISA really doesn’t matter.An Interview with Zen Chief Architect Mike Clark
Zen is one of the most important microarchitectures in the history of the x86 ecosystem.www.computerenhance.com
Mentions not seeing much difference/serious advantage from hypothetically going ARM. According to Clark, the clocks/performance/power qualities of x86 versus ARM (basically the higher clocks, higher power "nature" of x86) SoCs mostly have to do with markets that they target, not with the ISA.
But they did ship the Seattle cores in 2016.They didn't release anything (for good reasons, but that's another discussion). So in my book, there's no proof.
Well, ISA DOES matter a bit. If it didn't matter at all, we wouldn't be seeing APX in a few years in an x86 processor near you.This is what Jim Keller was saying too. It goes both ways. ISA really doesn’t matter.
But design matter. Let’s see who creates the widest core but also has the highest clocks, no one’s done that yet that’s far more interesting than ISA arguments.
He mentions they basically worked around the x86 variable-length encoding by adding uOp cache. Nowadays AMD uOp caches are rather huge with notable share of R&D pouring into them. Wouldn't it be easier to not having to?Mentions not seeing much difference/serious advantage from hypothetically going ARM. According to Clark, the clocks/performance/power qualities of x86 versus ARM (basically the higher clocks, higher power "nature" of x86) SoCs mostly have to do with markets that they target, not with the ISA.
Wow cool. Found this: https://old.hotchips.org/wp-content...eattle-White-AMD-HotChipsAMDSeattle_FINAL.pdfBut they did ship the Seattle cores in 2016.