Associate
- Joined
- 11 Jun 2021
- Posts
- 1,024
- Location
- Earth
-
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.
Alder Lake-S leaks
- Thread starter Grim5
- Start date
Yeah, this is the point i have been making, ADL P cores are 3X the size of Zen 3 cores and AMD have the option of a 6nm that's about 40% more dense or 5nm that is twice as dense, they also have some significant IPC gains coming, they have the technology to cram 4X as many P Cores on their CPU's vs Intel.
AMD don't even have to try and they would crush ADL.
Associate
- Joined
- 29 Oct 2019
- Posts
- 1,003
It's nice to see Intel competitive on the benchmarks again, even if it's at the expense of insane power consumption. Overall the 12900k still seems inferior to the 1 year old 5950x, but competition is good and hopefully AMD will be forced to lower their prices.
Associate
- Joined
- 11 Jun 2021
- Posts
- 1,024
- Location
- Earth
So am I right in thinking (after watching 5-6 reviews) that Alderlake is now the dogs ******* in everything but efficiency, including on DDR4 / PCIE-4 ? Some anomalies with Win10 but looking solid on Win11 for day 1.
Haha.
Welcome back.
Good to be back. I was only going to get to wear that 12900K Golden Plastic Wafer Gaming Performance Crown for a few months till V-Cache Ryzen 9's came out anyway!! Short term pain for a 2022 gain!! AMD will likely hold the crown once V-Cache arrives till Raptor Lake launches at the end of 2022. I'll get to wear a crown for a whole 9-10 months instead of only 2!! LOL
No, the 12900K is trading blows with the 5950X in productivity at twice the power.
Its about 10% faster overall in games at 720P.
Its good, its just not as good as it needed to be.
Anandtech's review is out. Alder lake's multithreaded scaling of SPECfp's memory intensive tasks are much better compared to Zen 3, but only in DDR5. Doubling of channels did pay off. Hopefully we'll see the same with Zen 4.
Each P core can consume ~65w of power. Each E core can consume ~10w power. The E cores are only slightly slower than Skylake. These are not efficient chips in any, shape or form.
Definitely a return to form from Intel, but the value proposition isn't there, given the prices of Z690 and DDR5 right now. By the time DDR5 prices become competitive we'll have Zen 4 in the market as well, so will be an entirely different scene.
Each P core can consume ~65w of power. Each E core can consume ~10w power. The E cores are only slightly slower than Skylake. These are not efficient chips in any, shape or form.
Definitely a return to form from Intel, but the value proposition isn't there, given the prices of Z690 and DDR5 right now. By the time DDR5 prices become competitive we'll have Zen 4 in the market as well, so will be an entirely different scene.
Reviews should really focus on 1440p.
Time to update my signature:
Intel Core i5-12600K | ASUS Z690M PLUS D4 | 2x16GB DDR4 @ 3200MHz | RTX 3080 FE | Samsung 980 PRO 1TB | Fractal ION+ 860W Platinum | Fractal Focus G Mini Case | Gigabyte 27" 1440p 144Hz 1ms IPS Monitor | Logitech THX 2.1 Speakers
Time to update my signature:
Intel Core i5-12600K | ASUS Z690M PLUS D4 | 2x16GB DDR4 @ 3200MHz | RTX 3080 FE | Samsung 980 PRO 1TB | Fractal ION+ 860W Platinum | Fractal Focus G Mini Case | Gigabyte 27" 1440p 144Hz 1ms IPS Monitor | Logitech THX 2.1 Speakers
FP scaling in memory intensive tasks in great on DDR5, but poorer than Zen 3 on DDR4. So it's not like that. Performance is roughly equal to Zen 3 (wins some, loses some), but efficiency is just much worse.
Nothing in real world.
I would pick the motherboard and platform you like most and one that is easy for you to setup with a cooler you can actually get or an adapter for your current cooler.
There is really nothing in it. Have a good look at the motherboards and see which you like most and the one that offers you the things you need now and things you may update later, eg more m.2 slots, wifi, pcie slot placements, amount of usb ports etc.
Make a list of advantages against the disadvantages and pick the winner.
If you have the, try Cyberpunk 2077 or WD:L with raytracing maxed out and 100% detail in WD:L, but DLSS ultra performance, and go to an actual in-game area and drive around (say Japan Town in CP, and Westminster for WD:L). Curious to see results.
Zen3 and ADL are both dual channel memory platforms/cpus. There is no doubling of channels that would make it a quad channel cpu/platform.
Like x99 or x299 as an example are quad channel.Zen3 and ADL are both dual channel memory platforms/cpus. There is no doubling of channels that would make it a quad channel cpu/platform.
In DDR5, each DIMM has two channels.
Associate
- Joined
- 11 Jun 2021
- Posts
- 1,024
- Location
- Earth
That's ranks not channels. Also DDR4 has dual rank ram too, my current 64gb 2x32 are dual rank. Not the same thing as channels mate.
Also heads up DDR5 also has single rank ram too.
I've not got Cyberpunk or Watchdogs Legions unfortunately so wont be able to test them , feel free to share your Steam account with me though and will happily give them both a shot
i will say if in games i have to use DLSS in i normally have it set to quality at max as i even at quality i can notice it isn't quite as nice as native . ultra performance would be a no go but suppose its useful for stressing the cpu with higher fps 
Deleted member 209350
D
Deleted member 209350
Go and read the reviews, you're very wrong. Gaming power consumption is very, very good on these chips.
Look, igor even says so here (via Google translate)
MODEL BOYS you hear!
Who on earth games at 720p
No, that's channels. Literally. Use Google:
https://techxplore.com/news/2020-07-ddr5-sdram-standard-boost-dual-channel.html
>> Each DIMM can handle two 32-bit memory channels instead of only a single 64-bit channel. Since each bank operates independently of each other, the burst length can be doubled and greater efficiency can be achieved. That means, for instance, DDR5 SDRAM can perform two 64-byte operations in the same time it takes DDR4 SDRAM to perform just one operation
In case you weren't aware, this was a huge bottleneck in multithreaded FP operations in all DDR4 systems which made bandwidth utilisation very inefficient on non-server grade platforms.
So yeah, not the same thing as dual rank, "mate"