uzzi38
Platinum Member
- Oct 16, 2019
- 2,607
- 5,821
- 146
Well considering the post I was replying to was about the ST score specifically...
Gideon
Golden Member
- Nov 27, 2007
- 1,619
- 3,643
- 136
Andrei made a great point:
And I agree that the same problem exists (to a lesser degree) with Zen 3 vs Zen 2. AES IMO is nowhere near general enough to be such a part of the score (the main loser here BTW is mostly M1 vs Intel and AMD).
Anyway according to Andrei, Geekbench 6 will remove Crypto score from total, which makes sense.
Andrei F said:
And I agree that the same problem exists (to a lesser degree) with Zen 3 vs Zen 2. AES IMO is nowhere near general enough to be such a part of the score (the main loser here BTW is mostly M1 vs Intel and AMD).
Anyway according to Andrei, Geekbench 6 will remove Crypto score from total, which makes sense.
DrMrLordX
Lifer
- Apr 27, 2000
- 21,608
- 10,802
- 136
It's almost as if using a single synthetic benchmark to compare CPUs is a bad idea . . .
The leaked Geekbench result is already discussed in December:
videocardz.com
www.tomshardware.com
forums.anandtech.com
Intel Core i7-11700K beats Ryzen 9 5950X by 8% in Geekbench 5 single-core benchmark - VideoCardz.com
It appears that a retail version of the Core i7-11700K might have just made a debut on Geekbench. Intel Core i7-11700K spotted, Z490 motherboard, 5.0 GHz clocks The new 11th Gen Core processor has been spotted in the Geekbench database with clock speeds reaching 5.0 GHz. This processor is based...
videocardz.com
Intel Core i7-11700K Appears to Beat AMD in Geekbench 5
Is AMD's performance supremacy going to be short-lived?
www.tomshardware.com
Page 338 - Discussion - Intel current and future Lakes & Rapids thread
Page 338 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
forums.anandtech.com
DisEnchantment
Golden Member
- Mar 3, 2017
- 1,599
- 5,767
- 136
Maybe for clients not but most server loads uses SSL for basically everything, and speedup due to AES makes a lot of difference how much time the request takes to be routed to the worker nodes from the proxy / balancer.
Additionally, walking the iptables and ebtables can be greatly accelerated with AVX type instructions.
If you are doing something in the server, you want these things.
Last edited:
Asterox
Golden Member
- May 15, 2012
- 1,026
- 1,775
- 136
A smart move from Intel, which lately has only been redirecting customers to the AMD camp. 
videocardz.com
Intel confirms B460 and H410 motherboards will not support 11th Gen Core Rocket Lake CPUs - VideoCardz.com
No Rocket Lake support on B460/H470 motherboards Intel will force budget-oriented customers to buy 500 series motherboards if they chose Rocket Lake CPUs. Intel has now confirmed that existing entry-level H410 and mid-range B460 motherboards will not be compatible with upcoming Intel Rocket Lake...
videocardz.com
Dave2150
Senior member
- Jan 20, 2015
- 639
- 178
- 116
A smart move from Intel, which lately has only been redirecting customers to the AMD camp.
Intel confirms B460 and H410 motherboards will not support 11th Gen Core Rocket Lake CPUs - VideoCardz.com
No Rocket Lake support on B460/H470 motherboards Intel will force budget-oriented customers to buy 500 series motherboards if they chose Rocket Lake CPUs. Intel has now confirmed that existing entry-level H410 and mid-range B460 motherboards will not be compatible with upcoming Intel Rocket Lake...
Z490 chipsets and CPU's are still cheaper than Ryzen, plus there's actually stock of the Intel parts, so no, AMD continue to redirect customers to Intel.
Wonder if there is any technical reason for this, or is it pure segmentation again?A smart move from Intel, which lately has only been redirecting customers to the AMD camp.
Intel confirms B460 and H410 motherboards will not support 11th Gen Core Rocket Lake CPUs - VideoCardz.com
No Rocket Lake support on B460/H470 motherboards Intel will force budget-oriented customers to buy 500 series motherboards if they chose Rocket Lake CPUs. Intel has now confirmed that existing entry-level H410 and mid-range B460 motherboards will not be compatible with upcoming Intel Rocket Lake...
Clutching for a technical reason, I guess as probably the last 14nm product, Rocket Lake might mean power hungrier than Comet Lake but still if Z490 supports Rocket Lake that would only make sense if Intel gave the mobo manufacturer the wrong spec for B460/H410 versus Z490.
uzzi38
Platinum Member
- Oct 16, 2019
- 2,607
- 5,821
- 146
Nope. In fact, we now have "H410" boards that actually come with a H470 chipset just to get around this.
遠坂小町@Komachi on Twitter: "🙄 https://t.co/GH06EcjG1w" / Twitter
Pretty sure H410 and possibly even B460 is just a rebrand of original Coffee Lake's chipset and might not even be 14 nm. I am a little surprised they aren't going to support the 11th Gen Comet Lake, makes me wonder if the chipsets are going to be EOLed. Still very unusual for Intel to do this.
The H470 is using the updated one and despite the typo on the page does seem like it is supported.
AlderLake-S ADP-S DDR5
Base Frequency 2.19 GHz
Maximum Frequency 27.2 GHz
I wonder if it's 2.72 Ghz maximum frequency.
Base Frequency 2.19 GHz
Maximum Frequency 27.2 GHz
I wonder if it's 2.72 Ghz maximum frequency.
uzzi38
Platinum Member
- Oct 16, 2019
- 2,607
- 5,821
- 146
1T score is on par with the 9900K, and 27.2/8 = 3.4, so that's where my guess is. This score makes the most sense between the 3.4-3.6GHz range.
Makes more sense but even with 3.4-3.6 Ghz it's a big IPC boost when I look to 4.7 Ghz i7-1165G7, this score is oddly high for 3.4-3.6 Ghz.
B460 and H410 (Comet Point-V) are actually 22nm Kaby Point-H dies (originally 200 Series). The other 400 Series chipsets are 14nm Comet Point-H, and the 500 Series are 14nm Tiger Point-H. Intel just happened to use the same stepping designation (A0) for CMP-H as KBP-H, which I'm sure wasn't intentional at all.
But we've known all along that CMP-V wasn't going to support Rocket Lake, so I'm not sure why this is news now.
Hulk
Diamond Member
- Oct 9, 1999
- 4,212
- 2,001
- 136
The ADL single core score is right about where a 1185G7 is.
Assuming the TGL 1185G7 was single core turbo'ed all the way to 4.8GHz, as long as the ADL was running under about 3.8GHz we could be looking at a 20% IPC increase over TGL. Lots of assumptions though so who knows...
More assumptions.
If we look at multithread scores, then this ADL result is comparable to about 9.2 TGL cores.
But 16 cores/24 threads seems to indicate 8 Big HT cores, and 8 Small non-HT cores. I can't make sense of it. One interpretation (that doesn't make sense) is that the ADL cores are equal to TGL cores (in this bench) and the 8 little cores are equal to 1.2 TGL cores? Perhaps the clocks of the small cores are really low? Or cores not enabled?
These "leaks" remind me of a treasure hunt where every clue leads to more questions and draws one no closer to the treasure. But I guess the adventure is fun so that's why I'm here
Assuming the TGL 1185G7 was single core turbo'ed all the way to 4.8GHz, as long as the ADL was running under about 3.8GHz we could be looking at a 20% IPC increase over TGL. Lots of assumptions though so who knows...
More assumptions.
If we look at multithread scores, then this ADL result is comparable to about 9.2 TGL cores.
But 16 cores/24 threads seems to indicate 8 Big HT cores, and 8 Small non-HT cores. I can't make sense of it. One interpretation (that doesn't make sense) is that the ADL cores are equal to TGL cores (in this bench) and the 8 little cores are equal to 1.2 TGL cores? Perhaps the clocks of the small cores are really low? Or cores not enabled?
These "leaks" remind me of a treasure hunt where every clue leads to more questions and draws one no closer to the treasure. But I guess the adventure is fun so that's why I'm here
Last edited:
Ajay
Lifer
- Jan 8, 2001
- 15,406
- 7,833
- 136
Geez, that would be a pretty poor all core frequency. Ofc, it's probably an engineering sample on non optimized BIOS.
uzzi38
Platinum Member
- Oct 16, 2019
- 2,607
- 5,821
- 146
There's something up with the memory tests too. Despite reporting as DDR5, the bandwidth and latency looks more like low clocked but very tightly tuned DDR4.
Intel Corporation Alder Lake Client Platform vs Micro-Star International Co., Ltd. MS-7B17 - Geekbench Browser
But in any case, at 3.4GHz compared to 5GHz it's roughly a 1.45x IPC improvement over Skylake. At 3.6GHz it's roughly a 40% IPC improvement. Definitely on the higher side of things, but can't say it's totally unexpected. I was expecting more of the second though if I'm honest - more like a 40% improvement, but hey, both are good.
Ajay
Lifer
- Jan 8, 2001
- 15,406
- 7,833
- 136
Intel will still need higher clocks to take on AMD, in addition to the IPC boosts. Even though I think Geekbench is like using a pair of binoculars to study the rings of Saturn.
Just a reminder, Intel told us H2 2021. The sources put it to Sep 2021. That's over 6 months to the launch. This means we are looking at an early ES using an early MOBOs, super-conservative RAM modules, etc.
uzzi38
Platinum Member
- Oct 16, 2019
- 2,607
- 5,821
- 146
Well I certainly wouldn't expect Alder Lake to cap out at the mid-3GHz range ahahahahaha.
No chance of the final version going any lower than 4.3GHz, because then we're talking about potentially lower 1T perf than TGL-H actual.
SAAA
Senior member
- May 14, 2014
- 541
- 126
- 116
No doubt it's going faster than 3.4 or such when we got one leak at 4.7GHz already.
That's Alder-P so mobile too, I believe the desktop line-up will definitely reach 5+ thanks to 10ESF process, just like Tiger H vs quads.
5GHz and 40% IPC over Comet by the end of this year would be fabulous, they are only missing HEDT class but there will be Sapphire Rapids for that, without big-little and probably up to 18 cores, albeit some time later.
Hulk
Diamond Member
- Oct 9, 1999
- 4,212
- 2,001
- 136
Golden Cove isn't going to have +40% IPC. Keep expectations a bit more grounded.
Edit: Misread. Thought that was supposed to be over Sunny Cove.
Edit: Misread. Thought that was supposed to be over Sunny Cove.
Last edited:
Looking back & found this slide which was posted few days ago in this thread I realize,
I don't know who made this slide but from the info I read it seems he was a bit conservative about ADL's performance, something fishy: the first 10%-20% over TGL is a large range, adding up the gap of RKL lead against SKL/CML(~15%), it could translate to ADL has ~25%-35% advantage against SKL/CML, which is pretty conservative compare to rumored ~40%.
Best scenario if the final RKL has 20% average lead to CML, adding up and still a floating range of 30%-40% for a gap between ADL to CML.
As for the leak about ADL's perf it's sad that we always saw something like Geekbench which has so many unstable factor, I don't understand why those leakers addicted to GB lol.
I don't know who made this slide but from the info I read it seems he was a bit conservative about ADL's performance, something fishy: the first 10%-20% over TGL is a large range, adding up the gap of RKL lead against SKL/CML(~15%), it could translate to ADL has ~25%-35% advantage against SKL/CML, which is pretty conservative compare to rumored ~40%.
Best scenario if the final RKL has 20% average lead to CML, adding up and still a floating range of 30%-40% for a gap between ADL to CML.
As for the leak about ADL's perf it's sad that we always saw something like Geekbench which has so many unstable factor, I don't understand why those leakers addicted to GB lol.
TRENDING THREADS
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)
- Started by DisEnchantment
- Replies: 9K
-
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 7K
-
Facebook
Twitter