News Intel's Core i9-13900KS Rips Ryzen 9 7950X In Early Benchmark

[Admin]

Intel's Core i9-13900KS leads single-thread, multi-thread CPU-Z benchmarks among desktop CPUs.

Intel's Core i9-13900KS Rips Ryzen 9 7950X In Early Benchmark : Read more

 

[-Fran-]

Given how much time left there is for its launch, it could be different OC teams playing around with pre-production chips. Remember that der8auer has gotten eng samples at times and puts them to the test, almost literally, to get some information for this own OC attempts, etc.

Still nice to see such massive jumps, so the performance wars continue.

Regards.

 

[svan71]

If it beats the 7950 in multithreaded work loads I'll eat my hat.

 

[Integr8d]

One, tiny, little core can eat 350w? That’s insane…

 

[Neilbob]

I can't help but shake my head at the grandiose clock speed claims being made these days. For perspective, 6.0Ghz is an approximate 3.5% increase over 5.8Ghz. That's so small as to be almost unnoticeable, but for what would undoubtedly be a comparatively huge power consumption increase.

Like overclocking, it all seems very pointless now.

 

[TerryLaze]

One, tiny, little core can eat 350w? That’s insane…

That limit is for the whole CPU, all cores, the speed increase you get in multi is still because each core runs faster (so single thread) .

 

[SunMaster]

If it beats the 7950 in multithreaded work loads I'll eat my hat.

I'll take care of the leftovers.

 

[Lafong]

For perspective, 6.0Ghz is an approximate 3.5% increase over 5.8Ghz. That's so small as to be almost unnoticeable, but for what would undoubtedly be a comparatively huge power consumption increase.

Yet, that 3.5% is still touted widely and bragged about.

What does that tell you about the touts?

What does that tell you about prospective buyers?

What does that tell you about the touts' opinion of the prospective buyers?

Would you buy "Milquetoast Harmless 3600 DDR4 RAM, 2 x 8 kit"?

 

[King_V]

I can see it now...

"OMG AMD is TERRIBLE for increasing power draw on their chips!!"

"OMG Intel is AMAZING with their performance! 18-25% more performance is totally worth it for an extra 52% of power draw!! That's ONLY 120W more!!"

 

[Eximo]

Or you just buy the locked version, set a reasonable power limit, and call it day. At this point it is almost pure marketing.

Ryzen 6800H vs the 12800H, both at 45W, they come out darn close to each other. And the desktop 12700 only gets you about 25% more performance for "65W" For another 10% you have to go to the 190W 12700k.

 

[spongiemaster]

"OMG Intel is AMAZING with their performance! 18-25% more performance is totally worth it for an extra 52% of power draw!! That's ONLY 120W more!!"

That would still be better than the 7950x. GN reported it being 25% faster than a 5950x while using 100% more power. Ironically, about 120W more for 25% more performance. Just like you said.

 

[hannibal]

This is actually interesting!

View: https://youtu.be/7JiYAwKIHRY

 

[TerryLaze]

I can see it now...

"OMG AMD is TERRIBLE for increasing power draw on their chips!!"

"OMG Intel is AMAZING with their performance! 18-25% more performance is totally worth it for an extra 52% of power draw!! That's ONLY 120W more!!"

We have been seeing it for 5 years now....

"Oh you can't change settings you have to use default out of the box settings..."

"Oh hey look...eco mode"

 

[Chung Leong]

Like overclocking, it all seems very pointless now.

Overclocking, in the old days, means buying a cheap processor and getting to perform like a more expensive one. It was a way for you to beat the system. Nowadays, it's way for the system to squeeze a few more bucks out of you.

 

[TerryLaze]

Overclocking, in the old days, means buying a cheap processor and getting to perform like a more expensive one. It was a way for you to beat the system. Nowadays, it's way for the system to squeeze a few more bucks out of you.

There are a few b660 mobos, even in the $200 range, that allow you to BCLK overclock locked i3, i5 and i7 CPUs and we are talking several tens of % improvements.
All the usual suspects youtubers have made videos about it.

 

[KyaraM]

We have been seeing it for 5 years now....

"Oh you can't change settings you have to use default out of the box settings..."

"Oh hey look...eco mode"

Also always fun: "No, you can't take undervolting into consideration, that's cheating and not actual power consumption!"

Now: "Oh, you can make Ryzen 7000 more efficient by undervolting!"

Same for temperatures.

 

[King_V]

That would still be better than the 7950x. GN reported it being 25% faster than a 5950x while using 100% more power. Ironically, about 120W more for 25% more performance. Just like you said.

I was comparing the 13900KS to the 7950X. No, 350W is NOT better than 230W to get 18-25% faster than the 7950X. Not sure what kind of math you're talking about.

 

[NightLight]

If it beats the 7950 in multithreaded work loads I'll eat my hat.

You want some ketchup to go with that?

 

[ingtar33]

remember, when intel claims "boost" clock speeds they're talking about the two most "overclockable" cores, the rest of the cores will run at much slower speeds. still 350W?

how is that even possible to cool?

as an aside, when AMD talks Ryzen 4 chips boost clocks they're talking about an all core boost. it's not an apples to apples comparison.

 

[TerryLaze]

remember, when intel claims "boost" clock speeds they're talking about the two most "overclockable" cores, the rest of the cores will run at much slower speeds. still 350W?

how is that even possible to cool?

as an aside, when AMD talks Ryzen 4 chips boost clocks they're talking about an all core boost. it's not an apples to apples comparison.

When intel talks about boost they also usually mean "for a duration of TAU" if it only uses 350W until the cooling is saturated and then goes back to normal boost you need much lower cooling.

Also it's not the most "overclockable" but the best binned cores that hit the highest clocks with the least power and volts.

Also also usually all of the intel cores can hit the same clocks if you don't care about power at all and have the cooling for it.

as an aside, when AMD talks Ryzen 4 chips boost clocks they're talking about an all core boost. it's not an apples to apples comparison.

No they don't, unless you mean an "up to" which means whatever.
Amd goes by the same two best cores principle that intel uses.
https://www.techpowerup.com/261377/...ster-dont-correspond-to-cppc2-preferred-cores

 

[KyaraM]

remember, when intel claims "boost" clock speeds they're talking about the two most "overclockable" cores, the rest of the cores will run at much slower speeds. still 350W?

how is that even possible to cool?

as an aside, when AMD talks Ryzen 4 chips boost clocks they're talking about an all core boost. it's not an apples to apples comparison.

Is that so? Must have missed the part where the 7950X runs at 5.7GHz all-core, then. Didn't they claim it hits that frequency? So where is it? It's the exact same as with Intel. Always expect the all-core boost a couple 100MHz lower than single core,which is given in the PR-slides.

 

[johnnyboy5520]

This isn't going to do much to help against the Zen4 v-cache chips considering a 5800X3D with much lower clocks can beat RL in some games and hang with it in many.

 

[TerryLaze]

This isn't going to do much to help against the Zen4 v-cache chips considering a 5800X3D with much lower clocks can beat RL in some games and hang with it in many.

If a game tops out to the max with cache why would you think it would scale even more with core clocks?
Zen4 with vcache could come out and possibly only match the 5800x3d because the game engine can't go any higher.
Zen4 could also possibly gain even more performance with vcache but either scenario is just as possible.

Also gaming is irrelevant when talking about compute performance since the 5800x3d isn't any better than any other zen3 in productivity.

 

[Alpha_Lyrae]

If a game tops out to the max with cache why would you think it would scale even more with core clocks?
Zen4 with vcache could come out and possibly only match the 5800x3d because the game engine can't go any higher.
Zen4 could also possibly gain even more performance with vcache but either scenario is just as possible.

Also gaming is irrelevant when talking about compute performance since the 5800x3d isn't any better than any other zen3 in productivity.

A game will either be CPU-limited or GPU-limited. Some don’t scale well, but many do. So, you’re usually only hitting a wall based on one of those two components. 1080p is almost always CPU-limited without ray tracing. Less GPU waits on frames from CPU, more frames/sec are possible.

Naturally, if a game only sees 2-5% improvement with a new CPU, there’s a compute scaling issue in the engine itself and the result should be excluded as it’s an outlier.

 

[TerryLaze]

A game will either be CPU-limited or GPU-limited. Some don’t scale well, but many do. So, you’re usually only hitting a wall based on one of those two components. 1080p is almost always CPU-limited without ray tracing. Less GPU waits on frames from CPU, more frames/sec are possible.

Naturally, if a game only sees 2-5% improvement with a new CPU, there’s a compute scaling issue in the engine itself and the result should be excluded as it’s an outlier.

So how does that relate to anything I said?
If cache isn't a limit then why are games faster with v-cache...is it magic?