Overclock.net banner

1 - 16 of 16 Posts

·
sudo apt install sl
Joined
·
7,305 Posts
Discussion Starter #1
Source: https://hothardware.com/news/lenovo-thinkpad-core-i7-1165g7-tiger-lake-cpu-xe-gpu-leak

There has been a flood of Tiger Lake leaks over the past several weeks (and months, really), and it does not look like they are slowing down. Case in point, a Lenovo ThinkPad laptop sporting an Intel 11th generation Core i7-1165G7 processor with Xe graphics in tow has made another cameo in the Geekbench database.

The Core i7-1165G7 is an unreleased Tiger Lake-U processor for ultrathin and light laptops. It is built on Intel's enhanced 10-nanometer manufacturing process (10nm+) and features four Willow Cove processor cores with HyperThreading support (so eight threads).

In the single-core test, the Core i7-1165G7 result is nearly 22 percent faster than the Ryzen 7 4800U. Bear in mind that Intel's chip is also running faster, at 2.8GHz to 4.7GHz versus 1.8GHz to 4.2GHz. Still, it is an impressive showing.
Odd they used an Arch Linux submission to compare. It is the best result for the 4800u, Intel also gains a decent performance boost when using Linux as well.

Best Windows multi-threaded result for the LENOVO 82DM(4800u): 25039 https://browser.geekbench.com/v4/cpu/15604942
Best Windows single-threaded result for the LENOVO 82DM(4800u): 5033 https://browser.geekbench.com/v4/cpu/15582880
 

Attachments

·
PC Evangelist
Joined
·
46,708 Posts
What is Ice Lake in this benchmark?
 

·
sudo apt install sl
Joined
·
7,305 Posts
Discussion Starter #3

·
Vandelay Industries
Joined
·
1,924 Posts
Need to know if it's running at 15w or 28w. We know the renoir is running at 1.8ghz while the tiger lake is running at 2.8ghz.

Context matters...w/o knowing, the information is basically useless.
 

·
PC Evangelist
Joined
·
46,708 Posts

·
sudo apt install sl
Joined
·
7,305 Posts
Discussion Starter #6
Relative to Ice lake seems more like refined 10nm + clk speeds + L2.
Yep, Willow Cove cores are rumored to have a 7% IPC improvement over Sunny Cove cores. Golden Cove will be a 25% increase over Willow Cove. So we're looking at a 16% IPC increase over AMD's Zen 2 with Willow Cove.
 

·
FOCAL ARIA FTW
Joined
·
9,344 Posts
Hmmm with a 1 GHz clock difference we have the below:

Code:
Multi core : AMD has an advantage of 17.61%
Single Core : Intel has an advantage of 21.8%
I am curious to see a test that normalizes clocks and brings the same no of threads on both sides.
 

·
sudo apt install sl
Joined
·
7,305 Posts
Discussion Starter #8
Hmmm with a 1 GHz clock difference we have the below:

Code:
Multi core : AMD has an advantage of 17.61%
Single Core : Intel has an advantage of 21.8%
I am curious to see a test that normalizes clocks and brings the same no of threads on both sides.
The 1Ghz clock difference will most likely became a factor during the multi-threaded benchmark. At most the single core difference is probably 300Mhz because I highly doubt both touched or kept their max boost during GeekBench.

I just ran GeekBench 4 on my 3500U(base 2.10Ghz) laptop with a 15w TDP it mostly stayed at 3068Mhz and jumped up to 3.2Ghz during the single core benchmark. I never saw 3.5Ghz which is the max boost for this laptop.

Single Core: 3134
MultiCore: 8759
 

·
there is no easy way out
Joined
·
3,253 Posts
Assuming a core isn't starved by low memory bandwidth and uncore/IMC/L3 cache speed, the performance increment should generally be linear with increasing core clock speed. Then I use this rough formula to calculate the IPC difference.

GB 4.4.2:
i7-1165G7: 6737 single core score / 4.7 GHz = 1433 score per GHz.
R7-4800U: 5531 single core score / 4.2 GHz = 1317 score per GHz.

1433 / 1317 = 8.8% higher IPC with i7-1165G7. only 8.8%!? Dayum!

(scratching my head)This is weird. As far as I have read that the Ice Lake itself has at least 10% better IPC than even the Zen 2 desktop part. Either the Intel one isn't reaching 4.7 GHz turbo or it's the 4800U going more turbo than 4.2 GHz? [or] the 4800U is running with faster RAM or the benchmark is biased toward Intel?

Also why use GB4 when GB5 is here? Is there any special reason to not use GB5?
 

·
there is no easy way out
Joined
·
3,253 Posts
One more thing. It definitely looks like Intel gets advantage in GB4 results as my Haswell at 4.1 GHz gives 5100 single core(4.1 GHz core, 2314 MHz memory) GB4 when Zen 3 gives 5033 at 4.3 GHz. This must not be happening. Haswell can not be faster than Zen 2 in at least IPC, let alone Zen 3. The GB4 is biased toward Intel or it doesn't get the same level of optimization to be fairly used on Ryzen.
https://browser.geekbench.com/v4/cpu/15582880
https://browser.geekbench.com/v4/cpu/15609883

Edit: corrections.
 

·
Zen
Joined
·
1,107 Posts
One more thing. It definitely looks like Intel gets advantage in GB4 results as my Haswell at 4.1 GHz gives 5100 single core GB4 when Zen 2 gives 5033 at 4.3 GHz. This must not be happening. Haswell can not be faster than Zen 2 in at least IPC. The GB4 is biased toward Intel or it doesn't get the same level of optimization to be used on Ryzen.
https://browser.geekbench.com/v4/cpu/15582880
https://browser.geekbench.com/v4/cpu/15609883
I don't like Geekbench 4 score that much since it causes too much score inflation on latency and bandwidth.
 

·
Vandelay Industries
Joined
·
1,924 Posts
Again...and I cannot believe I have to repeat this......Context ppl......marketing information is useless. I'll put it bluntly...if you put any relevance into information w/o context you are ignorant.
 

·
there is no easy way out
Joined
·
3,253 Posts
I don't like Geekbench 4 score that much since it causes too much score inflation on latency and bandwidth.
That's what synthetic benchmarks are. They don't make much sense in the real world but we still like them anyhow.
 

·
FOCAL ARIA FTW
Joined
·
9,344 Posts
This is my Razer Laptop.
I don't have to throw it away just yet ...
 

Attachments

·
Not a linux lobbyist
Joined
·
2,303 Posts
I found a plausible benchmark for this chip. Looks like being power throttled and single core performance is running away from the capabilities of Skylake and Ryzen. Ram latency is bad, but this is power throttled mobile so that will happen. GPU drivers look janky as well. Don't know if this has been posted elsewhere yet.

https://www.userbenchmark.com/UserRun/29555622
 

Attachments

·
there is no easy way out
Joined
·
3,253 Posts
I don't like Geekbench 4 score that much since it causes too much score inflation on latency and bandwidth.
Also, it's not latency and bandwidth only. With 8 GB RAM it give 5100 and with 16 GB RAM it gives 5173 (dual channel in both cases). Core, uncore speed, all voltages, memory speed, timings are exactly identical in both cases.
 
1 - 16 of 16 Posts
Top