Now we’ve moved past that to all the reasons Strix Point COULD have been great:Stages of an AMD Microarchitecture's Hype
- Wild exuberance and extreme optimism toward rumors
- Dismissing of less positive rumors
- Core is announced; looks mid
- AMD must be sandbagging the numbers; the real ones will be higher
- ES numbers leak
- Perf is mid, it must just be because it's an ES
- Core launches and is reviewed; proves to be mid
- Compiler/application optimization will improve it! Anyway, the next one will be amazing!
- Go to step 1
Wow, David Huangs blog review just embarassed every big site review to date with such in depth technical testing and explanation of results.Interesting data, I wonder how this comparison would change with proper full-width cores.
Where is the proof?M4 can be pushed further. Apple chose not to, to keep it within the efficient part of the v/f curve.
Hey man. No offense, you seem to have a strange obsession with Apple.Where is the proof?
For me, 9950X is the proof for Zen 5's frequency scaling. Where is Apple's proof of their silicon's frequency scaling superiority?
Strix Point literally fails to sustain its boost clocks @ 28W (look up Geekbench runs in Zenbook, like this one)Yes but if Apple silicon is so great, why can't it scale all the way to 5.5 GHz and solve all the world's problems? Strix Point can be pushed further. AMD chose not to, to keep it within the efficient part of the v/f curve. We also don't know what Apple's design choices cost them. Using huge caches must not be good for their yields. They are not doing anything like V-cache so cache defects lead to discarding the entire die. They have some ways to go before they can be compared to "real world" CPUs that the majority of the world uses. Not select few with oversized pockets or tendency to get themselves into cc debt.
They don't necessarily need to while their chips make a mockery of the competition.Here's what Apple needs to do before I would consider it fair for their CPUs to be used in these comparisons:
Make them price competitive with the competition.
I, a lifetime Windows user, did 2 years ago. And I'm not even from a first-world country, far from it. I wish Strix Point gave me a reason to go back. It did not.Until the above is a reality, these comparisons are nonsensical and just fantasies. No one is going to switch to Apple based on these graphs.
The FP improvements are truly impressive. INT is kind of a nothingburger.Here’s the spec2017 values at full speed, single-core:
View attachment 104158
Also Zen5 vs. Zen5c:
View attachment 104160
As a person from WCCF, you know quite well that 99.9999% of readership does not care about intricacies of arch design and performance (and I find it amusing that we were promised by the very trustworthy leakers an 'int monster' while in reality it's the other way around, as per usual for AMD), they care about games and who wins in useless benchmarks like GB or C23/C24 that their chosen vendor performs well (and if the situation changes, new 'best use case' is invented etc).Wow, David Huangs blog review just embarassed every big site review to date with such in depth technical testing and explaination of results.
SPEC doesn't have specific code paths, it's even one of the rules of benchmark selection not to have any (except of course for portability, such as type lengths, etc.).That depends on who is running SPEC, since some people compile the test with default x86 baseline that depending on compiler version and might not even allow for AVX2, others are using march native that would allow AVX512 usage if compiler knew about the core but would fall back to default if it was not recognized, and some are ensuring specific sets are enabled (like AT with avx2+fma).
I think Geekerwan, Anandtech and David Huang are all using different compilers and different flags to boot.
Now it won't matter if SPEC has a specific code paths for each instruction set and is doing a runtime dispatch based on detected processor features but if we are talking about the autovectorization then I think only David is using flags that would make use of AVX512 possible.
I'm still optimistic that Granite Ridge will post better INT numbers vs Strix Version of Zen5The FP improvements are truly impressive. INT is kind of a nothingburger.
Also interesting to note the distance in INT between 5 and 5c. Looks like there's a general INT weakness compared to Z4.
Hah, Apple silicon also features a v/f curve, am I right?Yes but if Apple silicon is so great, why can't it scale all the way to 5.5 GHz and solve all the world's problems? Strix Point can be pushed further. AMD chose not to, to keep it within the efficient part of the v/f curve.
Where is the proof?
For me, 9950X is the proof for Zen 5's frequency scaling. Where is Apple's proof of their silicon's frequency scaling superiority?
I see people claiming this a lot, but I have yet to see any good arguments for Apple cores reaching higher frequencies. Looking at other chips, the X elite caps out at 4.2GHz, and the Cortex-X4 is even lower. It seems to me that the design of these cores (with their mobile roots) has a lower frequency ceiling than x86 chips.Hey man. No offense, you seem to have a strange obsession with Apple.
The simple answer is that Apple has different targets/goals, and they do things the way they see fit.
I'll say no more, as this is the Zen 5 thread.
There are no tech sites anymore, just a bunch of clowns pushing clickbaits. Why would anyone "skilled" be employed at a tech site? Sadly, it's not 2000s.Wow, David Huangs blog review just embarassed every big site review to date with such in depth technical testing and explaination of results.
This is my personal grudge with Zen 5. It shouldn't be called Zen. It's not a balanced design anymore.The FP improvements are truly impressive. INT is kind of a nothingburger.
Also interesting to note the distance in INT between 5 and 5c. Looks like there's a general INT weakness compared to Z4.
A 9950X at 5.7GHz will use a lot of power in an ST test, like >25 watts and it will still deliver less INT score than a lower clocked M4.
Computerbase tested on a fanless Macbook Air!Computerbase test show that the Apple core doesnt manage to hit its max frequency more than 10s, after wich it throttle at half the power, so the ST perf is actually 25-30% lower than what is displayed, so much for the 9950X delivering less perf.
Edit : That s why Geekbench do pauses between each test, that s a blatant help to Apple as their core wouldnt yield good numbers if the tests would be run without delays.
The Hype Train was supposed to crash and blow up, not keep spinning its wheels after it derailed, lol.I'm still optimistic that Granite Ridge will post better INT numbers vs Strix Version of Zen5
I have a mild copium that a lot of microcode updates will be coming over the next year. It doesn't make much sense for a CPU nowadays to somehow bottom out FP performance but be lenient on INT. It's outright stupid even, which makes me believe that it likely wasn't their goal. It's possible that their new look-ahead branch predictor thing is demanding a lot of tuning and is behaving pretty poorly for now. Or that the new scheduler design has somehow not worked out as planned. It may Finewine yet, we'll see.This is my personal grudge with Zen 5. It shouldn't be called Zen. It's not a balanced design anymore.
I thought Apple's cores were a lot bigger than Zen's, that could explain the difference in perf, but clearly not.Strix Point literally fails to sustain its boost clocks @ 28W (look up Geekbench runs in Zenbook, like this one)
Regarding the yields, the Apple's P-cores aren't even that big. The Avalanche (the M2's P-core) has an area of 2.60mm^2 without L2 (that essentially plays the role of L3 cache) on N5. Zen 5 classic has an area o 3.47mm^2, and it scores like 10% higher in GB6, and M2 is a 2-year old chip.
Apple's cores are straight up better on client at this point.
No reason to think that it won't. It'll do a lot better in bandwidth, but as for the CPU cores, it's planned to be the same CPU cores.Hopefully STX Halo isn't as gimped.
Only on the MacBook Air which is passively cooledComputerbase test show that the Apple core doesnt manage to hit its max frequency more than 10s, after wich it throttle at half the power, so the ST perf is actually 25-30% lower than what is displayed, so much for the 9950X delivering less perf.
Edit : That s why Geekbench do pauses between each test, that s a blatant help for Apple as their core wouldnt yield good numbers if the tests would be run without delays.
Computerbase tested on a fanless Macbook Air!
No, that's a blatant proof that Geekbench was initially developed for embedded platforms.Edit : That s why Geekbench do pauses between each test, that s a blatant help for Apple as their core wouldnt yield good numbers if the tests would be run without delays.
No, that's a blatant proof that Geekbench was initially developed for embedded platforms.
You've never seriously used a recent Mac Book. I've yet to hear my MBP 16" M1 max turns its fans on despite being heavily used for compilation and running simulations. That's a huge advantage over my Lenovo P1 which always turns its fans on as soon as it's plugged in or when I browse on battery. And that Mac is more than twice faster than the P1 (both are ~3 years old), while being silent and not needing to be plugged in for hours.That s Apple choice to go fanless as a mean to boast their laptops as silent, so they are selling false perfs because people dont know that those chips are throttling even in ST.
I'm mostly talking about STX's horrendous uncore. 1MB L3 per Z5c core, and Meteor Lake's P-to-LP-core level cross-CCX latency.I thought Apple's cores were a lot bigger than Zen's, that could explain the difference in perf, but clearly not.
I'm stumped as to why they work so well.
No reason to think that it won't. It'll do a lot better in bandwidth, but as for the CPU cores, it's planned to be the same CPU cores.
Smartphone usage is often spiky.Because in embedded plateforms there will be no case where the ST loading exceed a few seconds, or with pauses that are shorter than in GB.?
The longer you stay there, the more screwed you will get. Need to invest again in max 7 years or maybe earlier when RAM/SSD limitations start troubling you or when the SSD writes get exhausted. No incremental upgrades. There's no guarantee that your laptop will even work flawlessly that long unless you are "subscribed" to AppleCare. RAID 1? What's that, asks Apple? Cheaper workstation laptops give you that option. ECC? Again, what's that, Apple asks? You use what we give you, now shut up and go back to saving money for your next whole laptop upgrade, Apple tells you. Yes, I'm sure it's Paradise on your side of the world.I, a lifetime Windows user, did 2 years ago. And I'm not even from a first-world country, far from it.