Raptor Lake Leaks + Intel 4 developments

[Bencher]

But you are missing is that most people at 'stock' was 241watt not the 125watt setting. So Intel with their spec sheet yes it would be more efficient out the box at 125w, however real world settings stock for 95% of people with the 12900k is 241w because of the mobo spec setup and so that is real world stock vs spec sheet stock that is being discussed.

Nobody moved the goalposts because everyone was talking about the real world using the machine etc. Not what Intel has decided is spec. The same point being out the box that means the Intel 12900k is pulling 241w and the 7950x pulling the 170w+ whatever happens to do based on the thermal limit ability to boost.

So to get to Intel stock settings you need to set the 12900k via PL1 in Bios 95% time and thus that means if you are doing that then you should compare the 105w or 65w of the 7950x or whatever Eco options are there then, because that is also click and play and not even in bios.

Yes Anandtech indeed are showing their score as 27k at 125w for the 12900k.

The 13900k would need to be showing a 50% improvement in efficiency to get there. Looking at the slides from Intel and what is happening then that isn't likely to change this gen from Intel just to try and get back on topic. Power has gone to 340w in boost ability to get this extra performance with it boosting to 5.8Ghz with leaks originally suggesting around 40k for that. I haven't checked if that still right but seems these are less efficient this time around to get max performance but likely similar efficient to the previous 12900k then at the lower power draw settings?

Lets just focus on raptolake in this topic. There is no way in hell the 13900 is going to be less efficient than the 12900k. I'm talking iso wattage, or the 241 / 253 watts pl2 limit. Minimum difference in efficiency should be around 25%. Actually in one of their slides they showed 65w 13900k matching 241w 12900k.

Its a more mature process with extra cores, it really cannot be less efficient than alderlake

 

[Bencher]

Is that oced? If stock, the st is freaking impressive. My 12900k oced to hell for st does 927.

 

[Bencher]

Here we see the 13th gen jumping the shark.

Trying very hard not to highlight the performance of the 5800X3D, which clearly wins in some games.

No mention of the just released Zen 4 CPUs:

You know whats the best way to hide the 3d? By not having it at all in the slides. Just like amd did

 

[uscool]

But you are missing is that most people at 'stock' was 241watt not the 125watt setting. So Intel with their spec sheet yes it would be more efficient out the box at 125w, however real world settings stock for 95% of people with the 12900k is 241w because of the mobo spec setup and so that is real world stock vs spec sheet stock that is being discussed.

Nobody moved the goalposts because everyone was talking about the real world using the machine etc. Not what Intel has decided is spec. The same point being out the box that means the Intel 12900k is pulling 241w and the 7950x pulling the 170w+ whatever happens to do based on the thermal limit ability to boost.

So to get to Intel stock settings you need to set the 12900k via PL1 in Bios 95% time and thus that means if you are doing that then you should compare the 105w or 65w of the 7950x or whatever Eco options are there then, because that is also click and play and not even in bios.

Yes Anandtech indeed are showing their score as 27k at 125w for the 12900k.

The 13900k would need to be showing a 50% improvement in efficiency to get there. Looking at the slides from Intel and what is happening then that isn't likely to change this gen from Intel just to try and get back on topic. Power has gone to 340w in boost ability to get this extra performance with it boosting to 5.8Ghz with leaks originally suggesting around 40k for that. I haven't checked if that still right but seems these are less efficient this time around to get max performance but likely similar efficient to the previous 12900k then at the lower power draw settings?

I have no idea what he is on, I've watched plenty of reviews of both out of the box stock

12900k 241w CB23 score of around 27k
7950x 225w CB23 score of around 38k

And if you start tweaking the 7950x wipes the floor

I stopped debating I seen for myself and that's enough for me just unbelievable the lengths of twisting just to defend I couldn't care for either brand not sure what you get out of it

 

[g67575]

You know whats the best way to hide the 3d? By not having it at all in the slides. Just like amd did

This is such a silly slide, they are claiming a percentage based performance advantage in some games (e.g. 6 percent), when you can see that the X3D is ahead in several titles.

It's sort of like saying 'well, the X3Ds don't count, they cheat with extra cache'

 

[uscool]

This is such a silly slide, they are claiming a percentage based performance advantage in some games (e.g. 6 percent), when you can see that the X3D is ahead in several titles.

That's against the 5950x it's all the marketing game

 

[LtMatt]

You know whats the best way to hide the 3d? By not having it at all in the slides. Just like amd did

They should have done that tbf, because if you can't beat the 5800X3D, then you are in trouble when Zen 4 3DVcache hits.

 

[arknor]

Do you have some kind of personality problem?

Getting upset about technology?

oh look personal attacks because he can't handle one mans opinion

 

[Dave2150]

They should have done that tbf, because if you can't beat the 5800X3D, then you are in trouble when Zen 4 3DVcache hits.

Zen4 with 3dvcache should have no issues reaching 110C, perhaps that's why they didn't release it to begin with!

 

[Deleted User 456458]

Is that oced? If stock, the st is freaking impressive. My 12900k oced to hell for st does 927.

Stock. 6hz ST. 5.6 all core. I’ll prob wait for KS knowing it’s coming much sooner than in the past.

With HT off, for gaming, 5.9/6.0 all core daily stable should be doable. Esp given how cool rpl runs compared to adl.

 

[Journey]

Zen4 with 3dvcache should have no issues reaching 110C, perhaps that's why they didn't release it to begin with!

It's already hit 110c in this thread with all that hot air you keep spouting.

 

[uscool]

You know whats the best way to hide the 3d? By not having it at all in the slides. Just like amd did

Tbf the 12 game average at 1080p reviews do have the zen4 chips slightly ahead of the 3d chip overall these are 3rd party reviews

In that intel slide it's 10 games either intel is sandbagging or its not doing too well

But 12900k wins overall so don't see them releasing something that will be slower

 

[Curlyriff]

Lets just focus on raptolake in this topic. There is no way in hell the 13900 is going to be less efficient than the 12900k. I'm talking iso wattage, or the 241 / 253 watts pl2 limit. Minimum difference in efficiency should be around 25%. Actually in one of their slides they showed 65w 13900k matching 241w 12900k.

Its a more mature process with extra cores, it really cannot be less efficient than alderlake

Okay the only slides I have seen showed 13900k scoring 40k at 350watt was where I got that from in not being any more efficient or not by a huge margin but as stated PL2 241watt side of things if that score of 40k is right then there doesn't appear to be huge efficiency improvements. The 65w matching the PL2 241w is good. But that still puts it way off the 7950x then since that scores 31300pts at 65w compared to you was saying 27k points you'd be looking at for the 13900k at 65w then if that same as PL2 12900K.

 

[Bencher]

Tbf the 12 game average at 1080p reviews do have the zen4 chips slightly ahead of the 3d chip overall these are 3rd party reviews

In that intel slide it's 10 games either intel is sandbagging or its not doing too well

But 12900k wins overall so don't see them releasing something that will be slower

Its using 4800 / 5600 ram for ald and rpl. Makes sense the numbers are the way they are.

 

[g67575]

oh look personal attacks because he can't handle one mans opinion

If it seemed harsh, I was only reacting to what seemed to be an overly defensive point of view regarding Intel.

 

[Bencher]

Okay the only slides I have seen showed 13900k scoring 40k at 350watt was where I got that from in not being any more efficient or not by a huge margin but as stated PL2 241watt side of things if that score of 40k is right then there doesn't appear to be huge efficiency improvements. The 65w matching the PL2 241w is good. But that still puts it way off the 7950x then since that scores 31300pts at 65w compared to you was saying 27k points you'd be looking at for the 13900k at 65w then if that same as PL2 12900K.

In 65w zen4 might win but around 150 to 180w is a more reasonable comparison. Generally speaking you dont run these huge cpus at that low power draw.

Also heavy mt isn't the only metric either. Ald is still beating zen 4 in lightly threaded stuff, not all, but a lot of them. Things like adobe premiere autocad and the likes. Raptorlake will be the nail in the coffin in these

 

[Journey]

Generally speaking you dont run these huge cpus at that low power draw.

Incorrect, a couple of months ago one of the companies I work with/for shipped over 120 1U systems with Intel 12900T's in them which are 35w parts. Same number of cores and threads, lots of low power S/T class CPU's are used, and I've used normal parts power limited many times. Don't apply your personal circumstances to the rest of the world.

 

[Curlyriff]

In 65w zen4 might win but around 150 to 180w is a more reasonable comparison. Generally speaking you dont run these huge cpus at that low power draw.

Also heavy mt isn't the only metric either. Ald is still beating zen 4 in lightly threaded stuff, not all, but a lot of them. Things like adobe premiere autocad and the likes. Raptorlake will be the nail in the coffin in these

Indeed but why are you now picking arbitary wattage figures. We have 65w, 105w, 125w, 170w and 241w as like fixed delta points to compare. At none of those does the 13900k look to compete with Zen4 if you do direct power to performance from anything noted. Yes it is better than the 12900k but if the first point that 65w is 5k points behind then it takes 350w to get similar to the 7950k at 170w+ boost wattage extra, so around 240w max. It will still at least in this one synthetic test that everyone has been talking about show it off.

Steve shows that the 5950x is still the most efficient though in heavy mt workloads relative to performance. Would be interesting to compare mind what having it locked down in one of the eco modes does. It just seems the stock side is stupid high and way past efficiency curve. Which goes back to needing 350w from13900k and same issue for it to match AMD.

Yes software and all sorts otherwise are hugely different. I would be interested in why Zen4 struggles with some things like AutoCAD when it seems to boost fine and have better IPC but still fail. Is there still instructions or issues on AMD side because things have been so heavily optimised for Intel for last decade+ in that software as example? It relative to all the figures and IPC we know shouldn't be loosing out unless there are critical instruction paths that Ryzen doesn't work with as well compared to Ald & Rpt.

 

[Nickolp1974]

Anyone know what OC boards will be available such as Apex, unify at launch, would prefer 2 dimm boards

 

[uscool]

In 65w zen4 might win but around 150 to 180w is a more reasonable comparison. Generally speaking you dont run these huge cpus at that low power draw.

Also heavy mt isn't the only metric either. Ald is still beating zen 4 in lightly threaded stuff, not all, but a lot of them. Things like adobe premiere autocad and the likes. Raptorlake will be the nail in the coffin in these

wins in every productivity tests here

AMD Ryzen 9 7950X Review & Benchmarks vs. 5950X, 12900K...

Support us on Patreon: https://www.patreon.com/hardwareunboxedJoin us on Floatplane: https://www.floatplane.com/channel/HardwareUnboxedBuy relevant products ...

www.youtube.com