• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Raptor Lake Leaks + Intel 4 developments

Is there any definition in comfortably? What are you thinking, 50%?

I'm sure we'll revisit this claim next year.

lol 50%? What was the last CPU in the recent generations that was 50% better at something?

If you're gonna use common sense and avoid the hyperbole, 7-10% is realistic. No need to engage if you're gonna struggle to keep a grip on reality though.
 
lol 50%? What was the last CPU in the recent generations that was 50% better at something?

If you're gonna use common sense and avoid the hyperbole, 7-10% is realistic. No need to engage if you're gonna struggle to keep a grip on reality though.
So, by comfortably you mean, slightly faster? I see you got carried away with the fan boy talk, it's okay we all do it sometimes Rob. :cry:

7-10% noted. :)
 
So, by comfortably you mean, slightly faster? I see you got carried away with the fan boy talk, it's okay we all do it sometimes Rob. :cry:

7-10% noted. :)

7% 10% in CPU's is a lot and you know that well. I wish we saw GPU level generational gains but that's not reality. I have no interest in anything but what's the best for my use case. Thus x3d and soon RPL. If anything, I'm an nvidia fanboy but posting in the GPU section is a fast track to the mental asylum.

If Zen4 didn't have a busted mem controller, I'd get one to play with. Ask yourself why your theoretical reads are so far off what they should be? And that's an anchor they'll drag along going forward and painfully so as DDR5 continues to mature.
 
Last edited:
7% 10% in CPU's is a lot and you know that well. I wish we saw GPU level generational gains but that's not reality. I have no interest in anything but what's the best for my use case. Thus x3d and soon RPL. If anything, I'm an nvidia fanboy but posting in the GPU section is a fast track to the mental asylum.

If Zen4 didn't have a busted mem controller, I'd get one to play with. Ask yourself why your theoretical reads are so far off what they should be? And that's an anchor they'll drag along going forward and painfully so as DDR5 continues to mature.
Fair, hear you on the GPU part. :cry:

I just hear comments like destroyed, comfortably etc, but as soon as you dig into the claim it's like 5-10%, albeit under very controlled circumstances and most likely in specific games.
 
7% 10% in CPU's is a lot and you know that well. I wish we saw GPU level generational gains but that's not reality. I have no interest in anything but what's the best for my use case. Thus x3d and soon RPL. If anything, I'm an nvidia fanboy but posting in the GPU section is a fast track to the mental asylum.

If Zen4 didn't have a busted mem controller, I'd get one to play with. Ask yourself why your theoretical reads are so far off what they should be? And that's an anchor they'll drag along going forward and painfully so as DDR5 continues to mature.


You say Zen 4 has a busted memory controller?? What do you mean by that and why and how is it important?? Is it mitigated by really fast DDR5 or no?? Will it also affect the upcoming 3D cache versions of Zen 4??

How do you think 3D cache Zen 4 variants do against Raptor Lake in gaming with a 4090?
 
So, by comfortably you mean, slightly faster? I see you got carried away with the fan boy talk, it's okay we all do it sometimes Rob.

It's also only that amount when you spend more time tweaking a system than you do actually using it, which is fine if you want a hobby, but not realistic and only counts for 0.0000001% (or less) of users of those parts involved.

It's also interesting to see people totally write of a part that isn't out yet, assuming that the interaction between that cache and the underlying CPU architecture will remain the same as the previous generation, not a very scientific approach at all really, and should be frowned upon.
 
It's also only that amount when you spend more time tweaking a system than you do actually using it, which is fine if you want a hobby, but not realistic and only counts for 0.0000001% (or less) of users of those parts involved.

It's also interesting to see people totally write of a part that isn't out yet, assuming that the interaction between that cache and the underlying CPU architecture will remain the same as the previous generation, not a very scientific approach at all really, and should be frowned upon.
Hence my reply tbf, we shall see. I think it’s dangerous to write off VCache.
 
And then, the REALLY hard part is to wait another 8 weeks until CES when Zen4X3D is apparently announced.

To be honest, it's nice to have an era when releases are fast and furious again. I was rather missing the feeling that there's always something new if you just wait a couple of months :)
 
Really with HT off. Isn't HT necessary for highly threaded games to keep the 1% lows smooth?? I mean if way more than 8 cores than yeah HT off. With 8 cores it is a tossup and less than 8 cores needs to be on. But Raptor Lake is still on 8 good cores. And highly threaded games that do not saturate more than 50% of other cores will be much better with HT on I would think. If they saturate almost 100% of all other cores than HT will not help much.

And really you think it beats Zen 5 which is 2 years away??
In most games HT doesn't do anything. In some games it hinders performance by a lot (spiderman for example). I wouldn't turn Ecores off, just the HT.
 
You say Zen 4 has a busted memory controller?? What do you mean by that and why and how is it important?? Is it mitigated by really fast DDR5 or no?? Will it also affect the upcoming 3D cache versions of Zen 4??

How do you think 3D cache Zen 4 variants do against Raptor Lake in gaming with a 4090?
It means that amd with 6000 ddr5 ram has worse reads and writes than I had with ddr4 on my cometlake :D

And the latency, oh boy....

Amd makes great all around cpu's. They are just not great at gaming, never were, never will be. If you want to write off some games as being "unoptimized", then get an amd cpu and join the horde on reddit. Every single game that runs like absolute crap on zen, people just call it unoptimized. I remember back when cyberpunk was released - i still had a 3700x an 8700k and an 11600k. You wanna guess in which of these 3 cpus the game was dropping down to the 30-35 fps? Yeah...
 
Last edited:
You say Zen 4 has a busted memory controller?? What do you mean by that and why and how is it important?? Is it mitigated by really fast DDR5 or no?? Will it also affect the upcoming 3D cache versions of Zen 4??

How do you think 3D cache Zen 4 variants do against Raptor Lake in gaming with a 4090?

Think Suez Canal and the Fabric being your canal. On either side is a big body of data but it’s limited by the canal itself.

AMD failed to make the fabric wider which is a problem since ddr5 is all about big bandwidth via high frequency.

Ultimately, if you want amd, wait for x3d whenever that shows up. Otherwise adl if on a budget or rpl if wanting something newer.
 
Last edited:
lol 50%? What was the last CPU in the recent generations that was 50% better at something?

If you're gonna use common sense and avoid the hyperbole, 7-10% is realistic. No need to engage if you're gonna struggle to keep a grip on reality though.
If you are talking about non 3d zen 4, I think the difference will be way bigger than 7-10%. Considering ADL can already max out current kits out there, and with A die up and coming, a stock 12900k with 7000+ ram probably wipes the floor with any zen 4, and then the 13900k will probably just poop all over the 12900k on top of that.

And this time around the difference will be seen even in reviews, since I assume they will be using the 4090 at 1080p.
What boggles my mind is why intel didn't send reviewers some nasty Adie kits like amd did with the expo kits for reviewing
 
Last edited:
If you turn HT off, you are wasting a sizeable chunk of the core, as you can no longer utilize the whole core with a single thread.
Although a thread may run faster when it has the whole core to itself, but you won't get high utilisation of the core.
 
I put in an RMA for my 7900x yesterday, requested under the 14 days allowance. Think maybe i was a bit hasty in buying as i need to see what both sides offer and prices are high all around, same with boards. I don't have to return it but i likely will and i have a week to do so. Will wait and see what Intel offer and probably keep hold of the 6400 ram i got for 244 for now.

Interested in seeing 13700k performance.
 
I put in an RMA for my 7900x yesterday, requested under the 14 days allowance. Think maybe i was a bit hasty in buying as i need to see what both sides offer and prices are high all around, same with boards. I don't have to return it but i likely will and i have a week to do so. Will wait and see what Intel offer and probably keep hold of the 6400 ram i got for 244 for now.

Interested in seeing 13700k performance.
I think again it comes back to socket support really At least with AM5 you’ll get 2-3 CPU cycles (8-10 years)where as RL is dead socket… it’s tough one.
RL might be faster now, but for those that upgrade CPUs every couple of cycles it’s a new board and CPU again
 
Back
Top Bottom