- Mar 3, 2017
- 1,747
- 6,598
- 136
I think there is that, but I also think that AMD has provisioned its roadmap for the future so that there is enough room for improvement that multiple future core µArchs can build on Zen5, and still have something to show as a significant selling point rather than a bare minimum <10% avg gain.I do think Zen 5 will be great architecture relative to Zen 4 but it’s not some mythical core that will obliterate every other CPU microarchitecture for the next couple of years as some leakers touted it be. In that regard they already got proven wrong.
It's rare but it does happen. We had a GB leak last month for STX that was showing like a 70% IPC improvement or something if the reported frequencies were correct, which they obviously weren't.
Edit: there are now multiple STX entries on GB6 though which are showing roughly the same PPC, so it seems like, at least in GB6, STX has a much higher PPC improvement than 16%.
Why it is AMD's fault when the performance doesn't meet some random forum user's expectation/fantacy ?
One person told to me that the >40% zen5 uplift chips had a ton of bugs and could not be launched.
And what's the significance of A0 silicon?
This kinda gels with my speculation that security vulnerabilities in the underlying µArch revealed during silicon bring up may have led to some loss in perf to half baked mitigation strategies.One person told to me that the >40% zen5 uplift chips had a ton of bugs and could not be launched.
AMD has released multiple steppings of each Z5 part, so presumably they have gotten all they can by now.This kinda gels with my speculation that security vulnerabilities in the underlying µArch revealed during silicon bring up may have led to some loss in perf to half baked mitigation strategies.
Also a significant µArch shift to a much wider core is almost certain to bring a number of new errata into the mix, some of which very likely won't be fixed in the Zen5 generation.
No.This kinda gels with my speculation that security vulnerabilities in the underlying µArch revealed during silicon bring up may have led to some loss in perf to half baked mitigation strategies.
It does have some regressions, but not anything you should care about.Also a significant µArch shift to a much wider core is almost certain to bring a number of new errata into the mix, some of which very likely won't be fixed in the Zen5 generation
But they did make B0 and C0 revisions, the question is 1)why weren't they released? 2)why did they wait 1 year+ to launch A0?It means that the very first design AMD pushed to the fab worked, and didn't require any fixing. Usually when the fab sends the A0 back, it's broken and needs to be debugged and fixed. The A0 being a working, sellable part is considered a significant achievement for the design and validation teams, and saves months of work and tens of millions of dollars.
It's rare, but it does happen. IIRC, the last time it happened to AMD was Bobcat. It doesn't matter much for the consumer, other than that the product is out potentially a bit earlier.
Possibly MSI, or whoever made that chart, is simply wrong. Or is it backed up by some other evidence?But they did make B0 and C0 revisions, the question is 1)why weren't they released? 2)why did they wait 1 year+ to launch A0?
But they did make B0 and C0 revisions, the question is 1)why weren't they released? 2)why did they wait 1 year+ to launch A0?
9900X has a 1.4x MT throughput than 7900X on a 67W lower power limit?https://nanoreview.net/en/cpu-list/cinebench-scores
Better than the Golden and Raptor cove, but i would be happy with that 28xx score leaked some months ago.
18% ST 22% MT. Not bad. While disappointing compared to 40% dreams, it wasn’t long ago that 20% took 4 generations of 5%. Thanks AMD!https://nanoreview.net/en/cpu-list/cinebench-scores
Better than the Golden and Raptor cove, but i would be happy with that 28xx score leaked some months ago.
7950X has 41% better score than the 7900X3D in CB R23, so that s relatively accurate, beside the 7950X doesnt get at 230W in CB and in any other bench.9900X has a 1.4x MT throughput than 7900X on a 67W lower power limit?
Questionable.
The only relevant score so far is that P16 STX GB5/6 run since it's definitely some lobotomite reviewer going YOLO on a pre-prod laptop.This is a very dodgy-looking site. You'd have to be a complete fool to trust anything from there.
Especially this lmao:
View attachment 100830
If or until they "update" it again, yes, of course.You mean in tasks where you previously performed so well that CPU Z editor "updated" the bench to get Intel at the first place again under the pretense that Zen 1 was taking advantage of an alleged bug to score better..?..
What does Intel have to do with AMD slides?
if anyone is truly sandbagging it would be Intel this time. Their CPUs don’t launch till September with Lunar and October/November for Arrow. So it’s “okay” for Intel to use the margin of error.
I do think Zen 5 will be great architecture relative to Zen 4 but it’s not some mythical core that will obliterate every other CPU microarchitecture for the next couple of years as some leakers touted it be. In that regard they already got proven wrong.
I’m sure we will see some insane gains in AVX-512 but the actual core performance will have to revealed by SPEC.
I think there is that, but I also think that AMD has provisioned its roadmap for the future so that there is enough room for improvement that multiple future core µArchs can build on Zen5, and still have something to show as a significant selling point rather than a bare minimum <10% avg gain.
It's no good managing some mind blowing, show stopping 30+% IPC gain for Zen5, only to produce a Zen6 or 7 looking so meager that people would rather put a pin in their upgrade, and wait for something more substantial to come along later.
With Zen1 they had no choice but to go all the way to dispel as much of the bad taste of Bulldozer as possible, but now they can afford to be a little more conservative, especially with Intel being so profligate with power spending to match or exceed AMD's efforts.
AMD keep digging with the same(~) uarch wide won't take them back to the game.
Is pretty sad to see AMD doing record profits without a killer CPU core and a not so big GPU like the enthusiasts want.
At now everyone here know my real nick.
18% ST 22% MT. Not bad. While disappointing compared to 40% dreams, it wasn’t long ago that 20% took 4 generations of 5%. Thanks AMD!
Oh, that joke was terrible lolNothing. Try having a sense of humor.
Not really. If they don't know performance without a 20% span by now they are incompetent, or obscuring for some reason. Also, oliterate every other CPU uarch is some serious straw man crap. Nobody claimed that.
It's more in line with typical Zen gains. Mid double digits. Zen 3 was the outlier. People expected more because they (supposedly) widened the core for the first time. We probably won't find a lot until Hot Chips unfortunatly.
Are you admiting to having a duplicate account on these forums? That wouldn't be wise.
Nobody who knows what they were talking about expected 40%. Totally agree with the rest of it though. No more "You will get four cores with minimal IPC gains and like it*" nonsense. And no buying a new motherboard because we changed the socket just to screw you over and sell more chipsets.
* Want hyperthreading? Unlock it for an extra $100!
Oh, that joke was terrible lol
Tip: CB R23One thing that is not logical is that if the 9700X does 25300 at 88W PPT then the 9950X should do 50000 pts since this would require only 180W, so far AMD slide state something like 46300 pts.
It means that the very first design AMD pushed to the fab worked, and didn't require any fixing. Usually when the fab sends the A0 back, it's broken and needs to be debugged and fixed. The A0 being a working, sellable part is considered a significant achievement for the design and validation teams, and saves months of work and tens of millions of dollars.
It's rare, but it does happen. IIRC, the last time it happened to AMD was Bobcat. It doesn't matter much for the consumer, other than that the product is out potentially a bit earlier.
But they did make B0 and C0 revisions, the question is 1)why weren't they released? 2)why did they wait 1 year+ to launch A0?
Also, oliterate every other CPU uarch is some serious straw man crap. Nobody claimed that.