- Mar 3, 2017
- 1,680
- 6,177
- 136
What application are you using to measure IPC?Yea it is, that L1/L2 lat regression hurts.
SIR 1-copy.What application are you using to measure IPC?
That's not clk-normalized.There are IPC gains. Minor yes, but not a regression.
Just to be sure, There aren't any public leaks about this yet, right? As far as the plebs are concerned, this is still only forum / twitter rumor-mill?I suppose this is confirmation of Z5 being the highest IPC core in the world.
Pretty clear that they have made a major breakthrough on the front end, conveniently that is the one area of the core all leaks have said nothing about.
Only shame is about the SMT yield. The fact the thing clocks comfortably over 5GHz is what will make the ARM boys head scratch.
Doubt.I suppose this is confirmation of Z5 being the highest IPC core in the world.
Nah we chillin'The fact the thing clocks comfortably over 5GHz is what will make the ARM boys head scratch.
I mean, Zen 4 does DDR-5 8000 today, even the APUs. It just doesn’t help much due to the architecture.There are no "next-gen boards" (and why would they be called 800-series when current boards are 600-series?)
Also DDR5-8000, 2400 FCLK, X3D launching day one and 8+16 SKUs are pure fantasy from him.
People always say that x86 can’t get any faster/wider/better…and then it does.Agreed. Should hopefully finally dispel the belief that x86 cannot scale as wide / have the same IPC as flagship premier ARM cores.
If adroc is to be trusted, then apparently not? Though he just seems to imply things, rather than tell them straight, for whatever reason.Doubt.
Zen 5 will have to bring a >50% IPC gain. That is very unlikely, eh?
Probably because saying X770 out loud makes it sound like "exeven seventy"and why would they be called 800-series when current boards are 600-series
Just to be sure, There aren't any public leaks about this yet, right? As far as the plebs are concerned, this is still only forum / twitter rumor-mill?
Why I ask, is that being a chronical Eastern European I love being needlessly sceptical and wrong, instead of being too hopeful and end up disappointed.
Case in point:
I was super hyped about RDNA3, once this Skyjuice leak surfaced, as it seemed quite obviously real and it was! These decent architectural gains on paper, excellent process shrink (TSMC N5) and promises by the vendor (50% perf/watt improvement) seemed to indicate a slam-dunk. Especially as AMD delivered on RDNA1 and RDNA2 perf/watt promises and had excellent track-record with chiplets.
Turns out, even awesome looking specs & architecture details don't mean much until they are backed up by actual benhmark numbers.
April can't come soon enough.
This is not Instructions Per Cycle (IPC). This is Geekbench 6 Score Per Clock (GB6SPC).3000 GB6 ST
5.7 GHz
IPC: 526.3157
The restricted VOPD scheme has been known from the open source ecosystem in advance. Yet, there has been clowns confirming stuff like: "n33 6nm 128bit gddr6 perf>6900xt".I too was pretty excited for RDNA3. Those ALU counts were wrong though because it didn't factor in that the actual spec was half the ALU count but they were dual issue (on very rare occasions).
It was also the 1st time for a long time AMD missed a publicly released perf/w uplift, even though they tried to make it seem like a hit with their shady 300w 7900XTX Vs 300w 6900XT comparison which was utter BS. To be honest whoever decided to run with those cooked numbers really needed to lose their job because it was such a step back for how transparent AMD usually are with marketing performance figures.
I read about the doom and gloom of x86 since I could access internet. It'd always something else sending it to the grave and the damn Michael Jackson keeps doing the thriller. Until arm windows ecosystem is thriving, there is no chance in hell I will switch, as I'm too invested into the software side.People always say that x86 can’t get any faster/wider/better…and then it does.
What's the normalized score of the Ryzen at 4GHz? I imagine it should be a bit higher, but no idea how much.Doubt.
Apple M3 Max
3200 GB6 ST
4.05 GHz
IPC: 790.1234567901
Ryzen 7950X
3000 GB6 ST
5.7 GHz
IPC: 526.3157
Possibly, but could also be interpreted as Zen 5 in ALL of its consumer products will be ready by H2, then technically speaking that statement would be correct if one assumes that Strix Point launches Summer 2024.So is Zen 5 client postponed until H2 after all?
Postponed? wasn't 20-24 months after Zen 4 always the plan?So is Zen 5 client postponed until H2 after all?
According to who? Hasn't AMD historically launched a new Zen generation in 18 months? Zen 4 was the exception due to COVID, new platform, new node.Postponed? wasn't 20-24 months after Zen 4 always the plan?
2 to 3 was 16mo or therein.Hasn't AMD historically launched a new Zen generation in 18 months?
According to every AMD chart which showed it simply in 2024, not 2024H1.According to who? Hasn't AMD historically launched a new Zen generation in 18 months? Zen 4 was the exception due to COVID, new platform, new node.
Zen 5 doesn't have to deal with any of that.
Zen3 is a new and mostly different core and was 16m.Planning for more validation time isn't surprising given it's supposedly an actual redesign of the core. Not like Zen1234 which were all evolutions.
2 to 3 was 16mo or therein.
I figured I do the homework and just lay it out for everyone's benefit:According to every AMD chart which showed it simply in 2024, not 2024H1.
Planning for more validation time isn't surprising given it's supposedly an actual redesign of the core. Not like Zen234 which were all evolutions.
Zen+ is a misdirection. Zen 4 was late September 2022. The real average of Zen1->2->3->4 is about 22 months. In any case, the point is AMD never said anything about 2024H1 for Zen 5. The idea that it is postponed comes from people making a trend from the exception to the rule (Zen 3) or thinking optimistically.I figured I do the homework and just lay it out for everyone's benefit:
Zen: March 2017
Zen+: April 2018 (13 months later)
Zen 2: July 2019 (15 months later)
Zen 3: November 2020 (16 months later)
Zen 4: November 2022 (24 months later)
That's on average 17 months between releases. More changes lead to longer gaps between releases and vice-versa. I'll let everyone make their own conclusions as to where Zen 5 *should* land.