• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Key Slides from Intel 13th Gen "Raptor Lake" Launch Presentation Leak

The e core seems like a tough pill to swallow for some, making compression harder and non-linear to the point is 'feels not fair'.

I have a filling that by the time AMD will implement it`s e core solution (around ZEN 5) there will not be any argue about it, just as having HT is just better than not having it.

Time will tell...
 
A niche reason to favor p cores only processor is if you use VMware Workstation heavily and need mt performance. Having VMware utilize any e core will make it only utilize e cores.

You have to edit vmx file to tell VMware not to use e cores if you want to utilize p cores. That makes paying for e cores a waste of money if what you're doing in VMware is the reason you need mt performance.
What you're describing is actually a bug in VMWare, not a reason to favor P-cores only. Though if you need that, you can either get a CPU with no E-cores or get any other CPU and disable the E-cores.

Also, what you (and other before you) describe, is more of a use case for a workstation CPU, not a consumer one. And yes, I know we don't have workstation ADL CPUs yet.
 
What you're describing is actually a bug in VMWare, not a reason to favor P-cores only.
It's definitely a reason to favor all p cores if the issue affects you.
Though if you need that, you can either get a CPU with no E-cores
I agree and is what I said.

Also, what you (and other before you) describe, is more of a use case for a workstation CPU, not a consumer one.
Just because someone's computer usage would benefit from HEDT/workstation, it doesn't mean they can afford or convince their employer to purchase one.
 
Just because someone's computer usage would benefit from HEDT/workstation, it doesn't mean they can afford or convince their employer to purchase one.
Just as not being able to afford one doesn't make it Intel's fault for not supporting workstation scenarios better on the consumer CPUs ;)
 
Just as not being able to afford one doesn't make it Intel's fault for not supporting workstation scenarios better on the consumer CPUs ;)
Also, what you (and other before you) describe, is more of a use case for a workstation CPU, not a consumer one. And yes, I know we don't have workstation ADL CPUs yet.
And if Intel comes out with an HEDT cpu that has little cores that need to be disabled for best performance in a particular app you'll be telling us we should be using server CPUs if mt is so important. And if Intel brings little cores to server CPUs you'll say we should be using super computers or cloud/distributed computing. Fan bois always gotta move them goal posts.

There are definitely scenarios that all P core cpus are the best, it's not every case for sure, but when a case comes up, just accept it and move on, there's no need to try to minimize something that was already minimized from the start by myself saying: "a niche reason to favor P cores...", doing so just makes you look blinded by bias.
 
And if Intel comes out with an HEDT cpu that has little cores that need to be disabled for best performance in a particular app you'll be telling us we should be using server CPUs if mt is so important. And if Intel brings little cores to server CPUs you'll say we should be using super computers or cloud/distributed computing. Fan bois always gotta move them goal posts.

There are definitely scenarios that all P core cpus are the best, it's not every case for sure, but when a case comes up, just accept it and move on, there's no need to try to minimize something that was already minimized from the start by myself saying: "a niche reason to favor P cores...", doing so just makes you look blinded by bias.
You seem to be missing the point. If Intel gave you 12P cores tomorrow, in heavy multithreading, that will most likely not be any faster than a 6P+8E cores. Because of thermals, those 12P cores would throtle like crazy, whereas a P+E configuration would have an easier time. That is why demanding "moar P-cores" doesn't make much sense. You want either P cores (few threads, high GHz scenarios) or P+E cores (many threads, lower GHz).

This is easy to test, someone can test an 8P vs 4P+4E cores all at the same configured TDP and see how they do in MT. But I haven't seen such a test yet.
 
Whoever wins, I got some $74 16GB DDR5-4800 standing by...
 
You seem to be missing the point. If Intel gave you 12P cores tomorrow, in heavy multithreading, that will most likely not be any faster than a 6P+8E cores.
You never tried making that point in any previous reply to me, if you had, I would have told you to re-read my post as it was pretty clear that you can only use P cores or E cores in the scenario provided. You cannot use a combination of both under that scenario. In the scenario, all cores of a Zen3/4 would be usable, but only 8 cores of your choice (all E or all P) would be usable from AL/RL. This gives the Zen3 a big advantage that will carry over to Zen4 up until the point VMware can properly utilize P+E cores at the same time.

If a person is considering AL/RL then that's the price range they are in. If they need mt performance in VMware more than gaming performance and can't use E cores (the scenario I posted), they should obviously opt for Zen3/4 for better cost/perf ratio over AL/RL. Suggesting people skip Zen3/4 which would give a free bump to performance, and to instead suggest they buy something at a price point not being considered, is pointless misdirection unless you thought AL/RL were the only options to consider for the scenario and I was suggestiong Intel make their CPUs with only P cores.
 
Weird things I read here. 4 ecores are 1pcore in space. Ecores are for better MT performance in applications. (so more pcores would not contribute to better MT performance in applications?) At the same time people claim how efficient pcores are at lower wattage when playing game etc. and yet they say ecores are better. Hmm, this makes me think, if these Ecores you all say are so great, why Intel even bothered with Pcores in the first place? Because of ST performance? Well then, that is 1 pcore needed not 8 of them.
 
Wrong, there are plenty of games that use more than 8 cores.
Euh... which?

How do background task cores offer better multithreaded performance?

They do in a scenario where your other cores would guzzle more power to do the same work, which they do even solely because of higher clocks and simply being bigger.

The whole reason Intel pushes these smaller cores is because they run into limitations otherwise. It is much like how Hyperthreading used to work in their advantage: they focus on a very strong ST performance, and the E core count is there even now to offload to free up TDP budget for higher P core performance. A HT thread was never 'a full core'. It just used the available headroom on a large core to do more work. E cores are really the hardware based implementation of it.

And this is also why they can keep up right now, which they really couldn't prior to big little. One could say an E core thread is worth more than an SMT thread.
 
Last edited:
Because of ST performance? Well then, that is 1 pcore needed not 8 of them.
Because the vast majority of apps either scale up to 8 cores or up to ncores. So for apps that scale up to 8 cores, having 8 pcores is better than having 1. Foe example photoshop premier autocad and games
 
You never tried making that point in any previous reply to me, if you had, I would have told you to re-read my post as it was pretty clear that you can only use P cores or E cores in the scenario provided. You cannot use a combination of both under that scenario. In the scenario, all cores of a Zen3/4 would be usable, but only 8 cores of your choice (all E or all P) would be usable from AL/RL. This gives the Zen3 a big advantage that will carry over to Zen4 up until the point VMware can properly utilize P+E cores at the same time.

If a person is considering AL/RL then that's the price range they are in. If they need mt performance in VMware more than gaming performance and can't use E cores (the scenario I posted), they should obviously opt for Zen3/4 for better cost/perf ratio over AL/RL. Suggesting people skip Zen3/4 which would give a free bump to performance, and to instead suggest they buy something at a price point not being considered, is pointless misdirection unless you thought AL/RL were the only options to consider for the scenario and I was suggestiong Intel make their CPUs with only P cores.
I have a hard time seeing how demanding changes in silicon instead of software fixes/updates is a good idea.
 
I have a hard time seeing how demanding changes in silicon instead of software fixes/updates is a good idea.
A person can demand hardware changes or software changes all they want but it will have little effect and I haven't suggested it. A person does however have full control over buying the best product that fits their needs within their budget. In this scenario, that would not be a cpu with E cores.
 
A person can demand hardware changes or software changes all they want but it will have little effect and I haven't suggested it. A person does however have full control over buying the best product that fits their needs within their budget. In this scenario, that would not be a cpu with E cores.
A nasty software bug is hardly a scenario.
I would just switch to VirtualBox if I were you. Or get a Threadripper.
 
A nasty software bug is hardly a scenario.
I would just switch to VirtualBox if I were you. Or get a Threadripper.
I gave VirtualBox an honest shot but it just didn't work out, and it has it's own bugs.
 
I gave VirtualBox an honest shot but it just didn't work out, and it has it's own bugs.
Yeah, it's quirky. And quite expensive if you need the extension pack. Serves me well enough though.
 
Euh... which?



They do in a scenario where your other cores would guzzle more power to do the same work, which they do even solely because of higher clocks and simply being bigger.

The whole reason Intel pushes these smaller cores is because they run into limitations otherwise. It is much like how Hyperthreading used to work in their advantage: they focus on a very strong ST performance, and the E core count is there even now to offload to free up TDP budget for higher P core performance. A HT thread was never 'a full core'. It just used the available headroom on a large core to do more work. E cores are really the hardware based implementation of it.

And this is also why they can keep up right now, which they really couldn't prior to big little. One could say an E core thread is worth more than an SMT thread.
Doom Eternal, F1, CSG, etc.
 
Doom Eternal, F1, CSG, etc.
Theyre all single thread limited in the end, and none of them presents a significant CPU load. You dont gain jack from having more cores over 8 but you do from frequency, except you will be GPU limited long before that in most setups.
 
z790 : waste of sand :shadedshu::kookoo:
 
I hate to say it, but if AMD launches with the prices they announced they are in big trouble! The RX 7600x only has 6 cores and is going against Intel's 14 cores, 6 performance cores which means its going to be equal in performance to the 7600x part and better in multithreaded applications. At the same price point why would I buy the 7600x over the 13600k?

I feel like AMD need to lower the pricing of all of their products by $100 in order to be competitive, otherwise it's going to be a massive win for Intel and we know how bad that is going to be for the PC market and for us consumers in general.

Their whole product stack needs to be $100 dollars less. I would hate to go with Intel, but as of right now they seem like the better option.
 
At the same price point why would I buy the 7600x over the 13600k?
Cause of the AM5 upgradability. In 2026 you can probably buy a CPU for 300€ that is similar in performance to the 13600k you can buy today :roll:
 
Cause of the AM5 upgradability. In 2026 you can probably buy a CPU for 300€ that is similar in performance to the 13600k you can buy today :roll:
I think that $300 by that point becomes more like a value of $50 or so, its still valuable, its better than nothing, but if at present the 7600x and 13600k have the same gaming performance, but the 13600k is faster in multithreaded apps and costs the same as the 7600x, there is very little current incentive to go for the 7600x. Yeah the 7600x is probably going to be much more power efficient, but I don't see that as a too big of an issue for most people. Maybe for office computers that need that extra power efficiency, but gamers aren't going to mind at all!
 
more e-waste cores? i'm happy i've sold 12100 and swapped for 5600X. it will be OK for me long days forward lol
 
I think that $300 by that point becomes more like a value of $50 or so, its still valuable, its better than nothing, but if at present the 7600x and 13600k have the same gaming performance, but the 13600k is faster in multithreaded apps and costs the same as the 7600x, there is very little current incentive to go for the 7600x. Yeah the 7600x is probably going to be much more power efficient, but I don't see that as a too big of an issue for most people. Maybe for office computers that need that extra power efficiency, but gamers aren't going to mind at all!
7600X more power efficient? No way in hell.
 
Back
Top