- Mar 3, 2017
- 1,747
- 6,598
- 136
have you seen Z5 power curves? no? funny.
have you seen Z5 power curves? no? funny.
I would argue Phoronix has the most useful benchmark suite of them all, since it's based almost entirely on real world workloads. It is also open! The only drawback is that it is Linux based. I'm sure you could probably run it on Windows with some effort, but I haven't looked at it in quite a while.
Unless Intel has decided to completely mislead, The "E" cores are about as fast as Zen 4 +-10%.
I think X3D chips will be the champs of gaming. I don't see Intel overcoming that. I think Intel will probably win this gen overall (minus gaming) unless AMD does something uncharacteristic with X3D, which is a very real possibility.
If intel is right, the E Cores are faster than Zen 5 at the same power envelope (20% faster than Raptor Cove at low power). Discounting AVX512 workloads of course.
Edit: in lunar lake will be a bit weaker due to no l3, so maybe faster than zen5c and close to zen5.
Law of large numbers is not the right term here, indeed. I'd describe it more in terms of diminishing returns. A more interesting point is what you say about relative vs. absolute value increase. I think we will have to adjust in the future to use absolute improvements more. Expecting consitent improvements in percentages is clearly not realistic as it assumes that the rate of improvement goes up with time. E.g. 10% growth per year means that the performance from year 10 to year 11 has to increase by 2.5x as much as from year 1 to year 2. These are effects we see in economy and biological systems, but I am not sure that they are equally applicable to CPU technology. For a long while we where in a growth region where linear gains could be approximated by multiplicative constants, but we have now reached a level of complexity and diminishing returns where this is no longer feasible.
SKT seems to be Z4c without the SIMD goodies at ? area.We seem to be throwing IPC and performance around synonymously which is causing confusion. If Intel slides turn out to be correct, then Skymont should be more or less on par with Zen4 IPC-wise (which is still fairly astounding and a great accomplishment by Intel). However, there is still a sizeable gap in clock speed with Zen4 DT and Skymont so I wouldn't expect the performance to be close.
And yet, videocardz says that this result is from stock 9600X. Something doesn't add up in their leaks.That was the reply to the clock speed. 9600x production CPU has higher boost clock speed than 5.1 GHz.
And yet, videocardz says that this result is from stock 9600X. Something doesn't add up in their leaks.
AMD Ryzen 5 9600X engineering sample gets overclocked to 5.7 GHz across all cores - VideoCardz.com
AMD Ryzen 5 9600X pre-release SKU under overclocking There’s an update to the leaked Ryzen 5 9600X engineering sample. Following the leaked AIDA64 score, the same leaker has now provided updated CPU-Z scores, confirming it can boost well beyond the last-gen CPU. The Ryzen 5 9600X is a 6-core...videocardz.com
Skymont/Raptor Cove IPC is Zen-4 like. But Skymont is 20%-80% faster than Lion Cove in its power range, which means it is very likely faster than Zen 5 in that power range, unless Zen 5 is vastly more efficient in that regime, which may be, but I don’t see evidence of that.We seem to be throwing IPC and performance around synonymously which is causing confusion. If Intel slides turn out to be correct, then Skymont should be more or less on par with Zen4 IPC-wise (which is still fairly astounding and a great accomplishment by Intel). However, there is still a sizeable gap in clock speed with Zen4 DT and Skymont so I wouldn't expect the performance to be close.
What they call 9600X ES is probably running @ 5.1Ghz in ST CPUz benchmark.Yeah, strange. And there is still the FMax mystery.
Again, just because IPC in Zen4-like, does not mean that it cannot outperform Zen 5 in the power band it is optimized for.SKT seems to be Z4c without the SIMD goodies at ? area.
In this case WCCFTECH have better reporting than videocardzAnd yet, videocardz says that this result is from stock 9600X. Something doesn't add up in their leaks.
AMD Ryzen 5 9600X engineering sample gets overclocked to 5.7 GHz across all cores - VideoCardz.com
AMD Ryzen 5 9600X pre-release SKU under overclocking There’s an update to the leaked Ryzen 5 9600X engineering sample. Following the leaked AIDA64 score, the same leaker has now provided updated CPU-Z scores, confirming it can boost well beyond the last-gen CPU. The Ryzen 5 9600X is a 6-core...videocardz.com
SKT seems to be Z4c without the SIMD goodies at ? area.
20-80% over lion cove has nothing to do with die shrinkWell yeah it gotta outperform Z4c first.
yeah wonders of shrink.
hello from 3 years ago.
View attachment 100926
20-80% over lion cove has nothing to do with die shrink
Intel numbers are usually accurate enough, that footnote is just-in-case stuff.Since it is +/- 10%, it could be Zen 3 to Zen4c territory, with 1 or 2 node advantage, 1-3 year behind Zen 3 / Zen 4c
That's not remotely iso comparison since LNC requires lighting the ring on fire.Ah, they compared to both. I’m talking about the lion cove comparison:
View attachment 100929
You said you have Zen 5 power scaling performance data, so I don’t know if 20-80% isn’t enough to keep up, but I think it may be close. I doubt Zen 5 is more than 20-80% efficient than lion cove.
Eh we’ll see. Even using the 20% performance iso power and 40% lower power at iso performance will be hard for Zen 5 to beat, outside avx512lmao. I was going to jump in, but what Adroc says is right. Skymont is cool but cmon guys. Read the Intel slides.
You can always compare PHX2 versus RPL-U 282; not that hard.Even using the 20% performance iso power and 40% lower power at iso performance will be hard for Zen 5 to beat
Yeah I just did. Best comparison I’ve found so far is AMD’s claim of something like 80% better PPW compared to 155h (vs 8840u), though that surely includes some less than favorable circumstances. Stacking another 20% on that… that does seem hard to beat. We’ll see.You can always compare PHX2 versus RPL-U 282; not that hard.
Can you read?Yeah I just did. Best comparison I’ve found so far is AMD’s claim of something like 80% better PPW compared to 155h (vs 8840u), though that surely includes some less than favorable circumstances. Stacking another 20% on that… that does seem hard to beat. We’ll see.
I'd say, not much )In this case WCCFTECH have better reporting than videocardz
And indeed, without fit/cca/whatever throttling Zen shows perfect linear scaling in CPU-z both ST and multi.Ryzen 5 9600X (5.7 GHz OC ES)
Ryzen 5 9600X (5.05 GHz Stock ES)
No need to be insulting. Go ahead and share the data.Can you read?
I've said RPL-U and PHX2.
That's the most apples to apples stuff out there, 2+8 versus 2+4, both backed by an on-ring LLC and both do it well at 15W.