A Geekbench score of the Intel Core i7-1065G7 'Ice Lake' running on the HP Spectre x360 convertible laptop is now available and seems to indicate very strong single-core performance that can rival processors much higher TDP such as the AMD Ryzen 9 3900X. Intel's 10nm could offer a significant advantage over AMD at least in the laptop segment with even more improved IPCs.
https://www.notebookcheck.net/10th-gen-Ice-Lake-Core-i7-1065G7-Geekbench-score-reaffirms-Intel-s-hegemony-in-laptops-offers-single-core-performance-on-par-with-an-AMD-Ryzen-9-3900X.427965.0.html
Told you this on last passmark leak. This little CPU scores like a 7700K (4.5Ghz, while running at 3.8Ghz). This is the 18% IPC increase Intel was talking about. Too bad they have issues with 10nm.
You are comparing Desktop chips from AMD with Laptop chips from intel. Can't you wait until AMD releases APUs targetted for laptops?
Quote from: sridhar kondoji on July 22, 2019, 15:54:46
You are comparing Desktop chips from AMD with Laptop chips from intel. Can't you wait until AMD releases APUs targetted for laptops?
Do you really think Zen 2 APUs would be able to take on Intel? Desktop, yes. AMD has a winner. Laptops, I think not yet.
Quote from: fasddsadasdas on July 22, 2019, 15:32:15
Told you this on last passmark leak. This little CPU scores like a 7700K (4.5Ghz, while running at 3.8Ghz). This is the 18% IPC increase Intel was talking about. Too bad they have issues with 10nm.
Issues were mostly because they were trying to outwit Moore's law (as discussed in yesterday's article). 10nm desktop CPUs would be interesting at this rate.
It's a silly article with a click bait subject line. We always knew the Intel has had better single core performance even with laptop chips. Even the 8950hk chips performs better than all ryzen CPU in single core benchmark. So it's nothing new... But it's not the only reason people buy a 3900X.
Just when Razer came out with their workstations and I thought the wait was over...
So, are we just going to ignore that the Apple A12x has a higher multi-core score than this chip? Or that the "105W" 3900x basically has 3x the multicore score?
Interestingly, compared to my i7-8650, the biggest gain is the memory score, which goes to 5819.
Hello guys, the article FORGOTTEN to mension about multi-thread performance. Who uses these days single? So, AMD 3 times more !!!
I bet this result achieved using Turbo boost. So, CPU work with real TDP around 35-40 watt(why everybody forget about this?) . Like my 8650U CPU.
15 watt TDP only for base clock.
AMD CPU real tdp 105 watt. So, if we remove 2/3 amd cores we can have around 105/3=35 TDP on quad core chip... And comparable multithread and single thread performance and perfomance per watt.
@mikhail - it isn't who uses single core, it is more of the fact that most programming languages were not created with multicore/multithread in mind.
When multithread was added to programming languages, many implementations have been messy patches. Add to the fact that most programmers learn to write code for single thread (because that was how it was always done). And things get complicated...
Add to the fact that the webbrowser is mostly single thread and that is why we've been stuck dependent on single thread.
The bright side is that things are improving on both the programming space and the browser space. But it'll still take some time for things to sink in.
And how much will it score after having to patch it to solve security issues?
Wow OK. I was going to buy a Ryzen 3900X, but I guess I better wait for this chip instead. /s
Comparison is apples and oranges. Plus why not just say it's as fast as an i7-8700K in single core instead?
And also, the TDP comparison is flawed. How much power is the 3900X drawing when stressing single core? Definitely nowhere near 105W.
Quote from: Matteo Mazzone on July 22, 2019, 18:03:51
So, are we just going to ignore that the Apple A12x has a higher multi-core score than this chip? Or that the "105W" 3900x basically has 3x the multicore score?
Exactly - the scores for A12X are 5053/18258 at presumably a fraction of the TDP. I bet Apple is going to release an ARM macbook next year - Intel can't compete.
Geekbench on PC and IOS is not even running the same load, so the scores can not be compared.
I call shenanigans, and I think what you meant to say was "take these benchmarks with a grain of salt." There's no way a laptop chip is coming close to a desktop chip.
Quote from: Vaidyanathan on July 22, 2019, 16:01:19
Quote from: sridhar kondoji on July 22, 2019, 15:54:46
Do you really think Zen 2 APUs would be able to take on Intel? Desktop, yes. AMD has a winner. Laptops, I think not yet.
Do you not even know that AMD's 7nm Zen 2 APUs have not been launched yet? Why don't you wait until then?!
Quote from: Wout on July 23, 2019, 14:50:42
Quote from: Matteo Mazzone on July 22, 2019, 18:03:51
So, are we just going to ignore that the Apple A12x has a higher multi-core score than this chip? Or that the "105W" 3900x basically has 3x the multicore score?
Exactly - the scores for A12X are 5053/18258 at presumably a fraction of the TDP. I bet Apple is going to release an ARM macbook next year - Intel can't compete.
That's dumb! When your silly Mac launches on ARM, you will have a whole world of pain when software built on x86/x64 stops running or ported versions crashes all the time!
Else any emulated versions of said software will run significantly slower & laggier compared to running natively on x86/x64. As the Qualcomm experience already has proven.
You can't even tell the difference between a crappy toy OS like iOS and real x86/x64 software used by corporates. Sheesh...
Quote from: A on July 22, 2019, 21:37:02
@mikhail - it isn't who uses single core, it is more of the fact that most programming languages were not created with multicore/multithread in mind.
When multithread was added to programming languages, many implementations have been messy patches. Add to the fact that most programmers learn to write code for single thread (because that was how it was always done). And things get complicated...
Add to the fact that the webbrowser is mostly single thread and that is why we've been stuck dependent on single thread.
The bright side is that things are improving on both the programming space and the browser space. But it'll still take some time for things to sink in.
Are you trying to impress us with 20 year old legacy programming truisms & history?
Are you so outdated you have not noticed that the latest most cutting edge software like games are already implementing spade loads of multi-threaded parallel programming? Even ray tracing requires shed loads of parallel programming goodness.
Single threaded performance is so passe. Its the old paradigm where Intel tells you to shut down all other apps before you can run a game (else it stutters like hell). Or that pop up that tells you to shut down everything else even when installing a piece of software!
Besides, you are also extremely outdated (at least 10 years) in terms of real world consumer usage of computing devices. Real world users run multiple software apps simultaneously all the time. On my PC I have up to three modern multi-threaded browsers each running dozens of tabs, streaming video or news, not to mention messaging and productivity apps running constantly in the background. That is multi-threaded processing at work, you dinosaur!
Even gaming is hardly single threaded. Modern gaming involves STREAMING the gameplay live at the same time. That requires multi-core multi-threaded performance GEDDIT?! That is where AMD shines and blows Intel completely away.
There is a reason why Android takes up close to 90% of the world's market share! It is an OS that has prioritised multi-threaded performance with multi-core (typically hexacore or octacore) CPUs since day one. AMD is using that same understanding of workloads the way ARM/Qualcomm are - building many efficient cores to perform a lot of tasks and running lots of software simultaneously in as fast/efficient manner as possible.
Hi....I tried a test with a new cooler. This CPU is only 2 wks. old for me.
5725 is not the highest number ever recorded, i know.
No links for me as a noob I guess.
icro-Star International Co., Ltd. MS-7B79
Single-Core Score Multi-Core Score
5725 47070
Geekbench 4.3.4 Tryout for Windows x86 (64-bit)
Result Information
Upload Date July 26 2019 02:50 PM
Views 6
System Information
System Information
Operating System Microsoft Windows 10 Pro (64-bit)
Model Micro-Star International Co., Ltd. MS-7B79
Motherboard Micro-Star International Co., Ltd. X470 GAMING PLUS (MS-7B79)
Memory 32768 MB DDR4 SDRAM 1672MHz
Northbridge AMD Ryzen SOC 00
Southbridge AMD X470 51
BIOS American Megatrends Inc. A.A0
Processor Information
Name AMD Ryzen 9 3900X
Topology 1 Processor, 12 Cores, 24 Threads
Identifier AuthenticAMD Family 23 Model 113 Stepping 0
Base Frequency 3.80 GHz
Maximum Frequency 4.54 GHz
Package
Codename Matisse
L1 Instruction Cache 32.0 KB x 12
L1 Data Cache 32.0 KB x 12
L2 Cache 512 KB x 12
L3 Cache 16.0 MB x 4
Single-Core Performance
Single-Core Score 5725
Crypto Score 7013
Integer Score 5338
Floating Point Score 5578
Memory Score 6496
AES
7013
5.28 GB/sec
LZMA
5713
8.92 MB/sec
JPEG
5219
42.0 Mpixels/sec
Canny
4932
68.4 Mpixels/sec
Lua
4687
4.82 MB/sec
Dijkstra
6428
4.35 MTE/sec
SQLite
4792
132.9 Krows/sec
HTML5 Parse
4469
20.3 MB/sec
HTML5 DOM
5977
5.42 MElements/sec
Histogram Equalization
4101
128.2 Mpixels/sec
PDF Rendering
5101
135.5 Mpixels/sec
LLVM
8083
555.8 functions/sec
Camera
5590
15.5 images/sec
SGEMM
4634
98.0 Gflops
SFFT
5520
13.8 Gflops
N-Body Physics
5738
4.29 Mpairs/sec
Ray Tracing
5453
796.3 Kpixels/sec
Rigid Body Physics
5738
16798.1 FPS
HDR
5303
19.2 Mpixels/sec
Gaussian Blur
5901
103.4 Mpixels/sec
Speech Recognition
6051
51.8 Words/sec
Face Detection
6016
1.76 Msubwindows/sec
Memory Copy
7278
20.2 GB/sec
Memory Latency
5383
80.4 ns
Memory Bandwidth
6997
37.4 GB/sec
Multi-Core Performance
Multi-Core Score 47070
Crypto Score 20287
Integer Score 59193
Floating Point Score 59893
Memory Score 7254
AES
20287
15.3 GB/sec
LZMA
80775
126.2 MB/sec
JPEG
70637
568.4 Mpixels/sec
Canny
41576
576.5 Mpixels/sec
Lua
69208
71.1 MB/sec
Dijkstra
40177
27.2 MTE/sec
SQLite
51807
1.44 Mrows/sec
HTML5 Parse
55348
251.3 MB/sec
HTML5 DOM
71405
64.7 MElements/sec
Histogram Equalization
53508
1.67 Gpixels/sec
PDF Rendering
36955
981.8 Mpixels/sec
LLVM
100784
6.93 Kfunctions/sec
Camera
68754
190.6 images/sec
SGEMM
48436
1.02 Tflops
SFFT
65799
164.0 Gflops
N-Body Physics
79395
59.3 Mpairs/sec
Ray Tracing
40655
5.94 Mpixels/sec
Rigid Body Physics
55697
163046.5 FPS
HDR
55974
202.9 Mpixels/sec
Gaussian Blur
68402
1.20 Gpixels/sec
Speech Recognition
62885
538.0 Words/sec
Face Detection
71891
21.0 Msubwindows/sec
Memory Copy
9332
25.9 GB/sec
Memory Latency
5291
81.8 ns