Sunday, December 20th 2020

Intel Core i9-11900 and i9-11900K (ES) Alleged CPU-Z Bench Numbers Reveal a 12% IPC Gain

Benchmark numbers of an upcoming Intel Core i9-11900 (non-K) and i9-11900K processor engineering samples allegedly obtained on CPU-Z Bench reveal that the chip will deliver on the company's "double-digit IPC gain" promise for the "Rocket Lake" microarchitecture. The i9-11900 (non-K) sample posted a single-threaded performance score of 582 points, while the i9-11900K ES posted 597 points, which are roughly 12% higher than typical CPU-Z Bench single-thread numbers for the current-gen i9-10900 (non-K) and i9-10900K "Comet Lake-S" processors. The multi-threaded score of the i9-11900 (non-K) ES, at 5262 points, ends up just around 5-10% lower than that of the i9-10900, despite a deficit of two cores. Intel's 11th Gen Core "Rocket Lake-S" story is hence shaping up to be that of increased gaming performance from the IPC gain, while roughly the same multi-threaded performance as the 10th Gen "Comet Lake-S."
Source: HXL (Twitter)
Add your own comment

47 Comments on Intel Core i9-11900 and i9-11900K (ES) Alleged CPU-Z Bench Numbers Reveal a 12% IPC Gain

#26
z1n0x
fancuckerThis was the uarch that was going to decimate AMD. Just imagine it on 10/7nm node. Amazing what intel can do hamstrung.
I'm telling you, it's the international reptilian conspiracy, that's keeping them down. Otherwise we would have 128 cores at 5 watts.
:roll::laugh::roll:
Posted on Reply
#27
Xeplane
The Quim ReaperThat's true, you only have to look at the number of dumb ass's paying $1500 for a 3090 (just for gaming) that's only 10% faster than a card costing half as much...
Value is not the only variable. I wish we can all exist in a world where people can be cognizant of that fact and respect it. I wish you well.
Posted on Reply
#28
Frank_100
R0H1TYou overestimate the intelligence part in making purchase decisions, or indeed underestimate the emotions/e-peen involved.
There was a time when I would let my Xeon 2687w desktop do numerical calculations for days at a time.
That was running optimized Intel Fortran to design Finite Impulse Response Filters. A 10% increase in speed would have been appreciated.
Posted on Reply
#29
DeathtoGnomes
z1n0xI'm telling you, it's the international reptilian conspiracy, that's keeping them down. Otherwise we would have 128 cores at 5 watts.
:roll::laugh::roll:
Whats wrong with you?!!!? You cant divide 128 by 5! it needs an even number #conspiracynumber
:kookoo::kookoo::kookoo:
Posted on Reply
#30
Gmr_Chick
kayjay010101I feel attacked :/
Awww, I'm sorry :( *hugs*
Posted on Reply
#31
AnarchoPrimitiv
Well, Moore's law is dead turned out to be correct, he said months ago that the IPC uplift would be 11-12%, something I completely agreed with based on the fact that Intel was making the ambiguous claim of "double digit IPC gains". If Intel achieved anywhere near 19% as AMD did, they would have been touting that specific figure all over the place, but the fact that they just said "double digit" definitely meant it was the on the low end of what constitutes "double digit".

I bet AMD will release an XT/Zen3+ refresh soon after on an improved 7nm node with 5% performance gains, since Zen3 has been released on the same exact 7nm node as Zen2, which seemed weird to me at first, but makes sense. Make more money now by using the same 7nm process with no increased costs, and have the 7nm+ in your back pocket later on to combat Intel AND justify keeping MSRPs exactly where they are despite increased competition.
Posted on Reply
#32
DeathtoGnomes
AnarchoPrimitivWell, Moore's law is dead turned out to be correct, he said months ago that the IPC uplift would be 11-12%, something I completely agreed with based on the fact that Intel was making the ambiguous claim of "double digit IPC gains". If Intel achieved anywhere near 19% as AMD did, they would have been touting that specific figure all over the place, but the fact that they just said "double digit" definitely meant it was the on the low end of what constitutes "double digit".

I bet AMD will release an XT/Zen3+ refresh soon after on an improved 7nm node with 5% performance gains, since Zen3 has been released on the same exact 7nm node as Zen2, which seemed weird to me at first, but makes sense. Make more money now by using the same 7nm process with no increased costs, and have the 7nm+ in your back pocket later on to combat Intel AND justify keeping MSRPs exactly where they are despite increased competition.
it wouldnt be the first time Intel or AMD has artificially inflated IPC gains for some pre-production hype. I predict actual gains to be single digit.
Posted on Reply
#33
Caring1
DeathtoGnomesWhats wrong with you?!!!? You cant divide 128 by 5! it needs an even number #conspiracynumber
:kookoo::kookoo::kookoo:
That's what they want you to think ;) :roll:
Posted on Reply
#34
londiste
Why would there be doubts about the IPC increase of Rocket Lake at this point? It is backported Sunny Cove, so IPC increase will be very close to what we saw in mobile from Skylake to Ice Lake.
Posted on Reply
#35
yeeeeman
gdallskthe scores are the same as my 9900K...
so, umm.... where's the uplift?
Single core is the same with much lower frequency...
Multi core, again, much lower frequency
And even with much lower frequency, it equals or beats 9900k/10700k.
11900k will be similar to 10900k multicore and better than all 10th gen in single core.
The results are underwhelming indeed, but there is so much you can do on 14nm before going overboard with power consumption.
Posted on Reply
#36
R00kie
yeeeemanthere is so much you can do on 14nm before going overboard with power consumption.
They already did, 9900k was a 95W TDP part, this is now 125W with lower frequency.
Sidegrade at best, in my opinion.
Posted on Reply
#37
CrAsHnBuRnXp
If it's not 10nm, i dont want it. Stop releasing CPU's intel until you get 10nm figured out. And a new naming scheme.
Posted on Reply
#38
djuice
gdallskThey already did, 9900k was a 95W TDP part, this is now 125W with lower frequency.
Sidegrade at best, in my opinion.
The 125w TDP is the 11900K, its a 5.3Ghz boost, and if it's true that its around a 10% IPC increase, that would be an upgrade over the 9900K over gaming, less so for multi-threaded application as it would be slower than the 10 core 10900K.
Posted on Reply
#39
23$Pa
yay double-digit improvement from my i7 5930k @ 4.5 that gets 490 points and that's a CPU from 2014 and its been running at OC all that time.
Posted on Reply
#40
R00kie
djuiceThe 125w TDP is the 11900K, its a 5.3Ghz boost, and if it's true that its around a 10% IPC increase, that would be an upgrade over the 9900K over gaming, less so for multi-threaded application as it would be slower than the 10 core 10900K.
Isnt this is what the news article is about?
The second screen is for the 11900 and the last is 11900k?
Neither of them are any meaningful percentage faster than what was already on the market.
Posted on Reply
#41
TheinsanegamerN
gdallskIsnt this is what the news article is about?
The second screen is for the 11900 and the last is 11900k?
Neither of them are any meaningful percentage faster than what was already on the market.
At the clock speeds they were tested. If these chips have even 200 mhz more headroom they'll be faster then the best 10900ks, and if they can hit 5 ghz, well, it'll be core 2 ->nehalem all over again.
Posted on Reply
#42
djuice
gdallskIsnt this is what the news article is about?
The second screen is for the 11900 and the last is 11900k?
Neither of them are any meaningful percentage faster than what was already on the market.
All screenshots are of the 10900 non K @ 4.4GHz, with a score of roughly 590~ for single thread. Which is on par with a 5.3GHz 10900K according to valid.x86.fr/bench/1 .

Now imagine a 11900K 5.3GHz score.
Posted on Reply
#43
WeeRab
To be fair...These are most likely ES parts, with a likely increase in core speed still to come.
I think it's quite impressive what Intel have achieved here on an old 14nm Uarch. Within spitting distance of AMD's IPC lead.
Having said that - 8 cores isn't going to be enough on the productivity side of the equation, even if they are marginally better in gaming.
Posted on Reply
#44
lu1zp1
Got to give it to inte! No one makes a better 14NM Processor than them ! :roll:
Posted on Reply
#45
r9
It's hard to keep track of Intel manufacturing, is this 14+++ or 14++++ ?
Posted on Reply
#46
Vayra86
kayjay010101I feel attacked :/
Just tell the world you 'create' something and all is fine with Ryzen nine. :roll:

Same goes for the top end GPUs, if you need cognitive dissonance the marketing is happy to provide.
Posted on Reply
#47
Blueberries
QUANTUMPHYSICSI won't be upgrading till DDR5 motherboards are the norm. Maybe 2 years from now. Not sure what generation intel will be at then, but I'm sure the performance will be great.
This is how I feel. DDR5 will to large of a generational leap to ignore, and will likely require a new socket / architecture as well as MOBO.
Posted on Reply
Add your own comment
Nov 26th, 2024 04:39 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts