Intel Core i7-1165G7 "Tiger Lake" Mauls Ryzen 7 4700U "Renoir" in Most Geekbench Tests

[yeeeeman]

It's not even the comparison against low power Zen 2 that's surprising, but the fact that those low powered chips are also just as good as intel desktop cpu, while being far more efficient. If those results are accurate, it's a tremendous improvement, and if they can make 8 core cpu based on that, then you have to wonder why did they even bothered to made the desktop 10th gen instead of that ?

Margins are very small on desktop

 

[HenrySomeone]

Yup, it's happenning! Just like some of us have been saying for a while now, the next Core 2 / Sandy Bridge moment is coming fast!

 

[ARF]

Intel's upcoming Core i7-1165G7 4-core/8-thread processor based on the 10 nm "Tiger Lake-U" silicon packs a mean punch in comparison to the AMD Ryzen 7 4700U processor, despite half the number of CPU cores.

The thing is that single-core performance doesn't depend on the overall number of cores.
There can be a single-core CPU with 10000zillion times higher performance, and yet super slow in multi-threading environment.

Yup, it's happenning! Just like some of us have been saying for a while now, the next Core 2 / Sandy Bridge moment is coming fast!

Not so funny joke.

 

[HenrySomeone]

A (hypothetical, obviously) single core CPU with "10000 zillion" times higher performance will still be (10000 zillion times)/n faster than an n-core cpu made out of 10000 zillion times slower cores in multi threaded apps, unless n is at least 10000 zillion as well, but probably a lot more.

 

[ppn]

Yup, it's happenning! Just like some of us have been saying for a while now, the next Core 2 / Sandy Bridge moment is coming fast!

Haswell/ sky /kaby delivered 25% better ipc so that i72600K performed more like i5 broadwell - 400 cpuz score at 3,2gHz, when sandy/ivy needed 4,0ghz.

The next big thing is Meteor lake 80% ipc boost over comet. Accept nothing less.

 

[ARF]

It's 8C-8T vs 4C-8T. So basically apples vs apples, almost.
Except that Zen 3 is coming, so that 1165G7 will still be slower than the newest AMD chips.

 

[HugsNotDrugs]

Exciting results but I'm apprehensive as to whether the Intel chip consumes far more power to hit those performance targets.

 

[jayjr1105]

Take it with a grain of salt. Could be a golden sample, we don't know anything about cost, power consumption, TDP, whether they get 2 or 12 good CPU's from a wafer, etc.

 

[phanbuey]

Doesnt even matter tbh... pricing will be the ultimate decider of these.

The majority of the starting ultra-books come with slaughtered on 4Gb-8GB on the board, for some of these, like the Dell Inspiron 7000 14, 8GB is the only option. These machines hang like it's going out of style when having a zoom/hangouts/skype call and having some tabs open in chrome + some other apps running.

Basically the ultrabooks that this is going into are so massively constrained on other aspects (screen/ram/ssd/bat/graphics) that the raw processor performance is typically way, way down on the list of importance. Knowing intel they will probably charge a premium, and as result will be a worse buy than the AMD systems at any given price point.

 

[Vayra86]

Take it with a grain of salt. Could be a golden sample, we don't know anything about cost, power consumption, TDP, whether they get 2 or 12 good CPU's from a wafer, etc.

Im taking my several cubic metres of salt right here tbh

Geekbench, missing info and miracles do not generally happen in the real wotld. If this miracle was real, it would have been much bigger than s half assed Geekbench leak.

People need to get their act together tbh. The amount of spin and how easily some buy it... pfff

 

[ARF]

Im taking my several cubic metres of salt right here tbh

Himalayan pink salt - it's healthier

 

[Dave65]

That's clearly Geekbench 4 by the way not 5 though and on different versions at that.

I pity whoever unironically believes a 25W CPU has similar performance to a 125W+ desktop CPU, imagine being this gullible. Intel's 10nm went from barely any better than 14nm to providing 5-6 times the efficiency ? Right.

I have to agree.

 

[king of swag187]

I wonder what the wattage and thermal requirements will be of the intel chips, those ryzens may have kickstarted the mobile segment again

Shame given how poor the tweaking and power management is, thats the catch with it, theres always one with AMD.

 

[watzupken]

Overall I'm very pleased with the TGL results. Intel has seemingly solved their 10nm woes as the TGL CPUs are boosting up to 4.7GHz. Hopefully desktop parts will come before 2022. Here's a complete list of leaked TGL results.

20% faster in single threaded mode than the Ryzen 3700X. Floating point performance is simply outstanding (see SFFT test). The 8-core CPU is just 50% faster despite having twice as many cores. Overall, it's a huge achievement considering that the 3700X consumes up to 91W while this TGL CPU is limited to 25W.



Yet GB5 represents real life performance quite well unlike GB4 which had certain glaring issues, e.g. it favoured RAM speeds (the test was completely removed in GB5) and AES performance quite a lot (it has a much less performance impact on the final scores in GB5).



GeekBench is an independent company which has never been known for favouring any CPU/GPU vendor. Your criticism is completely unwarranted.

I won't conclude this early when we don't know if this is tested in a laptop, or some bespoke test bench. Intel's 10nm should have improved after so many years, but I don't expect a miracle between first gen and second gen 10nm. With my existing Ice Lake based laptop, my observation is that these boost clockspeed generally can't be sustained for more than a second. Based on my observations, the usual sustained clockspeed is usually around 3 to 3.5Ghz, depending on manufacturer and the cooling solution used.

As for single core performance, I am expecting Tiger Lake to be better. Based on my observations, Zen 2 is generally close to about 10% faster than Intel's Skylake clock for clock. When we move to Ice Lake (Sunny Cove), its about approx 18% uplift from Skylake, clock for clock. So with Willow Cove, this should take the IPC gain even further.

Geekbench is an independent company, yes. But it is easy to game a benchmark since it is very predictable and easy to manipulate the score. Its worst when the leaks are mostly Geekbench results. Best is still to look at actual results in various tests.

 

[Fujikoma]

I wouldn't characterize any of it as a mauling. Article would have benefitted from CPU TDPs and, if available, CPU prices. Processor speed is pretty much the only thing saving Intel, aside from AMD limiting cache size on the 4800U.


AMD 4700U Vs Intel 1165G7

AMD 4700U 2.00 GHz
Single: 4923
Multi: 23314 (1457/Thread/GHz)

Intel 1165G7 2.80 GHz (+40%)
Single: 6737 (+37)
Multi: 23414 (+0.43%) (1045/Thread/GHz)


AMD 4800U Vs Intel 1165G7

AMD 4800U 1.80 GHz
Single: 5210
Multi: 28651 (995/Thread/GHz) cache size/architecture issue

Intel 1165G7 2.80 GHz (+56%)
Single: 6737 (+29%)
Multi: 23414 (-18%) (1045/Thread/GHz)

 

[yeeeeman]

It's 8C-8T vs 4C-8T. So basically apples vs apples, almost.
Except that Zen 3 is coming, so that 1165G7 will still be slower than the newest AMD chips.

Zen 3 renoir parts will come next year at this time. So a looooooooong time from now. In conclusion, Zen 3 parts will have to fight with tigerlake successor. Which is rumoured to be based on a new core, golden cove and about 25% better ipc than tigerlake (tigerlake will be more or less similar in IPC with Zen 3).

 

[HenrySomeone]

Precisely, Intel's superiority in the mobile segment at least is not really challenged at all and certainly won't be in the near future.

 

[Vayra86]

Zen 3 renoir parts will come next year at this time. So a looooooooong time from now. In conclusion, Zen 3 parts will have to fight with tigerlake successor. Which is rumoured to be based on a new core, golden cove and about 25% better ipc than tigerlake (tigerlake will be more or less similar in IPC with Zen 3).

Quite right but mobile wont cover their datacenter market. We remember what happens when you cannot compete in HPC and so far 10nm is not the answer it needs to be. Will a 7nm part do better.... we shall see.

AMD right now has a pretty neat single die design that they can tweak for years. Intel was already tweaking to keep up and has diversified immensely. Lots of IP means lots of designs for essentially the same thing.

Mid-long term these chips dont mean a thing unless they can scale them up towards performance segment. Laptops die fast. The margins arent great. And with even the minor competition from AMD in this space, that margin is under pressure and AMD does carry a major cost advantage per chip. Intel relies on high volume to keep a stack populated with TGL chips. AMD has virtually full binning freedom.

 

[HenrySomeone]

Margins on high-end laptops (which are almost exclusively Intel) are pretty decent and I'd imagine that at least some of that gets to the cpu maker as well