• 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.

Intel Cannonlake, Cascade Lake, Ice Lake, Tiger Lake & Sapphire Rapid thread

cascade is doa v TR3 looking at the results, the 18 core is not much faster than the 3900x lol.

Depends what you use it for - true workloads require may require extra lanes so the 3900x is never even considered regardless of the performance it has, it's a home user desktop chip.
 
Depends what you use it for - true workloads require may require extra lanes so the 3900x is never even considered regardless of the performance it has, it's a home user desktop chip.

Don't forget that the 3900X can support the lanes with double the bandwidth, because of PCIe 4.
 
3900x doesn't get much overclocking
Most games perform exactly the same between pbo and auto oc on or off

meanwhile the cascade lake x parts will get significant gains by overclocking the core and mesh

I'm not really referring to gaming, i'm talking rendering and encoding, tasks that can push these CPUs. If you can squeeze out 300Mhz per core on a 12 core cpu, on these tasks that's pretty good.
 
I think Sandys was referring more to the HEDT chips like the 7980XE/9980XE. If you have the cooling / power delivery* they can clock quite well on the all core frequency (although not as much if you compare to the 1-2 core max boost clocks and starting from a lower point thus have more theoretical headroom).

*portable nuclear power plant on standby.. :p

Yup I am all about watercooled HEDT here, screw the environment, power savings ...meh :D
 
Yup I am all about watercooled HEDT here, screw the environment, power savings ...meh :D

Amen to this. The day all the doom tree-hunging prophets drop their mobile phones & tablets for the good of the environment, then I would consider to scale down my PC :D

Similarly I would scale down from the 2 x V6 cars :D
 
that's why I can't take the results seriously for games.

i need to see a review where someone has done proper overclocking on the core, mesh and memory

They did overclock the Mesh, they clearly said they did and whats more if you look at the slide again you can see it, overclocked the 7800X gained 25% performance, you don't get that from overclocking the cores alone to 4.6Ghz.

Its simply that even with the Mesh overclocked its still not fast enough, the penalty for Intel's Mesh is far higher than it is for Infinity Fabric.

One other thing.

36 Games 9900K @ 5Ghz vs 3900X @ PBO + OC, probably running at 4.4Ghz all core, if even that high, a 14% difference in clock speed to the 9900K.

Overall performance difference 5% to the 9900K, that's about a 10% IPC deficit.

There are outliers at the bottom, anything over 10% to the 9900K is IPC to it, that's 5 out of 36 games, from 1 to 5%.

Ok so heres the thing about that, this test was done before they started 'DRam Calc' messing about with Timings on Ryzen 3000, later on GN and HUB did a small sample review on DRam Calc and Rysen 3000 performance, it didn't have a huge impact on all game but some of them near 20%.

I did this myself, 86 FPS Min in FC5 XMP RAM, i cross checked that with Toms Hardware who got 84 FPS min with their 3600. after tweaking the RAM speed, timings and sub timings my min FPS shot up to 96 FPS, on the same list from toms Hardware a stock 9700K was getting 94 FPS min, 129 Avr to my 121 but my GPU topped out at 160 for a while at the end of the run holding back averages.

Take note of what reviewers, especially HUB are not doing, with Zen + they made quite a big thing about using DRam Clac to improve Zen+ performance vs Coffeelake, given that in that way you could bring up gaming performance Coffeelake was still the clear winner.

They have made no comparisons to DRam Calc tuned Ryzen 3000 vs Coffeelake at all and when you look at the results they got when they did this in a review with Ryzen 3000 alone and how close Ryzen 3000 is to Coffeelake despite the lower clock speeds you have to wonder why.

I don't blame them, if the performance with DRam Calc tuned RAM it swung in Ryzen 3000 faviour i don't think HUB would want that attention for sticking thier head above the parapet, with the comunity being the way that it is they would be the target of a lot of controversy, i think they have decided the time is not right yet, so they are sitting on it.

BtI1ZwO.png
 
Last edited:
^^ Plus I think many of the charts/graphs used on review sites, then used by us enthusiasts, get muddled by timeline by AGESA and BIOS updates. (Not that yours are humbug).

Any net based website are paid by clicks. So many reviews so far, are upheld by us enthusiasts and used in forum discussions.

Soon to be AGESA 1.0.0.4 release should maybe have been the first BIOS reviews were allowed to make a decision for purchase for non/early adopters. But as its the internet these days, there is a rush by online review sites to make a definitive set of findings to inform the masses that a 3800X isn't worth the price premium over the 3700X, to attain their revenue, via the clicks. I would if I ran one.

Maybe the early test/review will be representative of the future. But many contributors in here are fixed on what they read as non owners, and not keep up with the gains from latest updates that enthusiast owners who are usually the most informed and have the experience with said CPU's.

AS a 3800X owner I have a vested/bias interest in that it does pull away from the 3700X once the AGESA and BIOS's are matured, or just maybe not (beta) immature. But there really is too much old, AGESA and BIOS graphs used currently that muddy the water for those looking to buy, particularity when non owners are using anything to 'help' buyers, use old and sometimes irrelevant metrics and reviews.

I get that there is nothing solid yet and AMD will have some egg if the 3800X really doesn't out perform the 3700X by it's %? cost uplift.

Just my 2 penneth at not at anyone.
 
Amen to this. The day all the doom tree-hunging prophets drop their mobile phones & tablets for the good of the environment, then I would consider to scale down my PC :D

Similarly I would scale down from the 2 x V6 cars :D

Don't forget that you are an inseparable part of the environment. If you screw it, you screw yourself.
But in general, your post is pure trolling and criminal.
 
Meh, those are margin of error numbers, i wouldn't read too much into it.

3950X will beat it tho.

They should, at least, have tried to camouflage the rebranding by calling it something else, for example i8-10918X or anything else different to just replacing the 9 with a 10.
 
Ya, and AMD will do a Threadripper 3000 whose entry level will be a 24 core at the same price as Intel's 10980XE that will absolutely destroy it.

That will be more expensive. Even the old now 24-core Threadripper costs £1100.

My basket at Overclockers UK:
Total: £1,108.69 (includes shipping: £8.70)​

Worst things humans can do is produce more humans. They are the only animals that require fuel. Save the planet - don't have kids.

Wrong. The problem is the quality of our governments and the rich people whose main interest lies in the digging the ground for oil, instead of looking at the sky for the sunlight...
 
2970wx can be had for under £900, was even £700 very briefly a while back.

Pretty sure AMD will put the new 24 core Threadripper at the 999 bracket to directly compete and destroy the 10980XE
 
Back
Top Bottom