Monday, September 13th 2021

First Tentative Alder Lake DDR5 Performance Figures Leak

The first indication of what to expect in terms of early DDR5 memory performance on Intel's Alder Lake platform has leaked, courtesy of a Chinese forum and Twitter. It's not the first time we've seen DDR5 performance figures, but this time the CPU is clocked at much higher speeds compared to the figures leaked by Longsys back in March.

Although the memory is still running at 6400 MHz, performance is up significantly with read speeds of over 90 GB/s. As we're most likely still looking at an ES sample and an early UEFI build, there should be room for improvement here. The leaked performance numbers appear to be from a Dell system of unknown model.
The not so great news is that the memory latency is still very high at 92.5 ns, although the 40-40-40-85 timings of the DDR5 modules used most likely has something to do with this. The CPU used for the test is said to be a Core i5 12600K and as an extra bonus, single core CPU-Z numbers of it were also posted, where it scores 785.6 points.

At this point in time, DDR5 still doesn't look very exciting, but as always when we transition to new memory technologies, there is usually an overlap where the fastest memory of the older generation always tends to outperform the first couple of JEDEC rated speeds of the new generation of memory. Once the memory module manufacturers get to play around a bit with DDR5, we expect to see much better performing modules in the market.
Source: @harukaze5719
Add your own comment

57 Comments on First Tentative Alder Lake DDR5 Performance Figures Leak

#51
Chrispy_
HenrySomeoneYup, Ryzens won't stay competitive much longer, especially considering it looks like even Meteor Lake will be out before Zen4...yikes! :D
Oh no! AMD's 2020 product won't be as fast as intel's unreleased Q4'2021 or 2022 product, AMD are doomed for sure, better sell all my AMD stock before they crash and burn!
/s
Posted on Reply
#52
GorbazTheDragon
ZoneDymothey are, core count matters more for those who dont know and because of that calling 6 performance cores and 4 eco cores simple "10 cores" compared to a 5800X 8 performance cores, is basically a loophole to false marketing.
8 cores on a desktop won't give equivalent performance to 8 cores on a laptop, that feels like such a weak argument... Imagine 15w renoir 8 core vs a 5800x...
TheLostSwedesuggests the small cores are actually doing a lot more than expected.
How were people expecting them to do? I'd expect the 4c cluster to at least match a single GNC core in performance... That's the very least.

8c GNC with 20% IPC over SNC/WLC puts you at around 9.6 core equivalent SNC/WLC (and in turn around that for zen 3), and just off the top of my head I'd say the little cores should have around 50-100% better performance per area compared to the big cores, so that'd be around 3.6-4.8 zen 3 core equivalent, which puts you just shy of a 5950x (assuming that is clocking the same as a 5800x).
HenrySomeoneYup, Ryzens won't stay competitive much longer, especially considering it looks like even Meteor Lake will be out before Zen4...yikes!
It'll just be back to price competition.
Posted on Reply
#53
TheLostSwede
News Editor
GorbazTheDragonHow were people expecting them to do? I'd expect the 4c cluster to at least match a single GNC core in performance... That's the very least.

8c GNC with 20% IPC over SNC/WLC puts you at around 9.6 core equivalent SNC/WLC (and in turn around that for zen 3), and just off the top of my head I'd say the little cores should have around 50-100% better performance per area compared to the big cores, so that'd be around 3.6-4.8 zen 3 core equivalent, which puts you just shy of a 5950x (assuming that is clocking the same as a 5800x).

It'll just be back to price competition.
What's GNC/SNC/WLC? Too tired to play guessing games and if you're going to use acronyms that aren't commonly used, at least explain them once.

I expected them to perform like the Atom cores they are, i.e. not so great and no at what looks to be very close to a full fat Core i core.
I presume you clicked on the link and looked at the graphs? According to that, Intel is making AMD look bad in most benchmarks and the 11x00 series look like a joke.
Somehow it doesn't quite add up with what I expected at least, although it also seems like not everything is behaving as expected, especially X265 encoding and the "digital content" test in PCMark, where the new CPUs are really falling behind.
Again, I would take all those graphs with a both an unhealthy dose of sodium chloride and ammonium chloride.
Posted on Reply
#54
GorbazTheDragon
TheLostSwedeGNC/SNC/WLC
golden sunny willow cove
TheLostSwedeI expected them to perform like the Atom cores they are
Intel can call them gaming cores, infinity cores, whatever they like... They wouldn't stick 4 of them in place of a single big core if they could get more performance out of the single core (either assuming equal area or equal logic since they both would take up one ring stop). In reality the performance per transistor of a complex superscalar, out of order core like this drops off a cliff as you increase the complexity, a lot of structures need to be scaled exponentially to get performance uplifts.

The graphs on that link are definitely questionable, but I don't think you should be surprised to see the top 8+8 parts getting within touching distance of a 5950x (that's not to say they won't draw more power).
Posted on Reply
#55
HenrySomeone
GorbazTheDragon8c GNC with 20% IPC over SNC/WLC puts you at around 9.6 core equivalent SNC/WLC (and in turn around that for zen 3), and just off the top of my head I'd say the little cores should have around 50-100% better performance per area compared to the big cores, so that'd be around 3.6-4.8 zen 3 core equivalent, which puts you just shy of a 5950x (assuming that is clocking the same as a 5800x).

It'll just be back to price competition.
Yeah, I expect a lot of sullen faces when 12900k with the so called "weak/fake/cheating cores" (you name it) will be beating 5950x more times than not for a lower price and comparable power consumption (possibly even quite a bit lower in a lot of scenarios). And oh, regarding AMD being competitive on prices - they were able to do that when they were still using the sub-par GLOFO, not so much now on the crowded 7nm TSMC (and it will be worse for 5nm, that's for sure). They really should have used this last couple of years better, but looking at their net profit figures, they haven't and it won't get better from here. The future honestly isn't all that bright for team red...
Posted on Reply
#56
GorbazTheDragon
HenrySomeoneYeah, I expect a lot of sullen faces when 12900k with the so called "weak/fake/cheating cores" (you name it) will be beating 5950x more times than not for a lower price and comparable power consumption (possibly even quite a bit lower in a lot of scenarios). And oh, regarding AMD being competitive on prices - they were able to do that when they were still using the sub-par GLOFO, not so much now on the crowded 7nm TSMC (and it will be worse for 5nm, that's for sure). They really should have used this last couple of years better, but looking at their net profit figures, they haven't and it won't get better from here. The future honestly isn't all that bright for team red...
The WSAs with TSMC were locked in long term and I don't see any indication of TSMC being reluctant to give AMD more capacity... Zen 2 was sold at lower prices per mm^2 when the node was less mature...

Also, the lowest cost per chiplet of zen 3 at current is the 5900X, cheaper per chiplet than even the 5600X. Not to mention supply of the Cezanne parts is plentiful while those have over 2x the die area of the single CCD parts (worse yields on top of 2x the raw silicon cost), yet costs less. Cezanne may not be selling gangbusters but I don't think there is any reason to believe AMD is at all limited by how much N7 they can get from TSMC or the prices it's being moved at...
Posted on Reply
#57
jbgtly
MakaveliI score this in the Cpu z benchmark. And we will need to see a full review with more than 1 app. ADL will most certainly have a ST lead as it should. However its abit soon to be writing off the 3d stacking chips based on a single leak benchmark.

This is my 5800X Cpu-z bench. This is after about 30min with curve optimizer when I first got the processor. Its actually hit 700/7000 a few times in the past when ambient temp was a little cooler, usually when crybaby family was over that think 80deg F is hot.
I honestly think the only times ADL will beat the current AMD cpus is when its using Sh** Tons more power.

Posted on Reply
Add your own comment
May 10th, 2024 22:11 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts