Monday, June 8th 2020
Intel "Tiger Lake" Beats AMD "Renoir" in Graphics Tests under 3D Mark
Now, take that title with the customary grain of salt, and remember: most mobile configurations aren't directly comparable due to different components, speed of the memory subsystem, and so on. Putting that salt aside, though, one thing remains: Intel beats AMD in the latest purported 3DMark benchmarks - and on the red team's home-field, so to speak: graphics performance. A benchmark posted by renowned leaker and benchmark scavenger rogame on twitter has turned up an Intel Tiger Lake-U (i7-1165G7) scoring 11879 (99.68%) in the Physics and 6912 (112.92%) in the Graphics score compared to AMD's R7 4800U 11917 Physics score and 6121 Graphics score.
For context, this pits a 4-core, 8-thread Intel Willow Cove design paired with Gen12 Xe graphics tech (2.8 GHz base, 4.4 GHz boost) against 8 of AMD's Zen 2 cores and Vega graphics. Also for context, it's expected that Intel's i7-1165G7 runs with a 28 W TDP, compared to AMD's R7 4800U 15 W envelope. Also of note is that 3D Mark isn't exactly the poster-child for CPU parallelization performance, as the benchmark scales up rather poorly as more cores are added. Perhaps more interesting as a comparison, these scores from Intel's Tiger Lake are comparable to the company's current i5-10300H (4C/8T 2.5 GHz base 4.5 GHz boost), which scores 10817 on the Physics side (making the i7-1165G7 9.8% faster with a 200 MHz slower base clock, 100 MHz higher boost & 17 W less TDP (28 W for the Tier Lake and 45 W for the i5-10300H).
Source:
rogame @ Twitter
For context, this pits a 4-core, 8-thread Intel Willow Cove design paired with Gen12 Xe graphics tech (2.8 GHz base, 4.4 GHz boost) against 8 of AMD's Zen 2 cores and Vega graphics. Also for context, it's expected that Intel's i7-1165G7 runs with a 28 W TDP, compared to AMD's R7 4800U 15 W envelope. Also of note is that 3D Mark isn't exactly the poster-child for CPU parallelization performance, as the benchmark scales up rather poorly as more cores are added. Perhaps more interesting as a comparison, these scores from Intel's Tiger Lake are comparable to the company's current i5-10300H (4C/8T 2.5 GHz base 4.5 GHz boost), which scores 10817 on the Physics side (making the i7-1165G7 9.8% faster with a 200 MHz slower base clock, 100 MHz higher boost & 17 W less TDP (28 W for the Tier Lake and 45 W for the i5-10300H).
35 Comments on Intel "Tiger Lake" Beats AMD "Renoir" in Graphics Tests under 3D Mark
RDNA or even better RDNA2 iGPU should have been here already in AMD's cpu's, AMD have the tech to keep Intel eating dust in the laptop segment BUT are very slow using it.
They have already done iGPU with RDNA2 for both SONY and Microsoft consoles, a SoC that actually are in production right now so it is not like they do not have the tech ready for release.
Very weird decision making by AMD. If they are to strike hard at Intel's laptop business the time is NOW before Intel get more competitive.
AMD Laptop cpu@5nm with Ryzen 3 and RDNA2 would send Intel back to the stone age and they should release that at soon as posible, preferably Q1 2021.
If AMD does not take a major step forward in the laptop segment now they will lose the window of opportunity with all the setback Intel had lately and that will probably never will come back.
When Intel get it's duck in a row and catches up it will be too late, just saying.
Basically if AMD goes any higher, they'd end up with this & not many customers!
www.techpowerup.com/268233/another-nail-on-intel-kaby-lake-g-coffin-as-amd-pulls-graphics-driver-support
The Physics test in this runs between 5 and 10 seconds only - this means the CPU will likely be at full turbo for the full time. The combined test runs longer, but is still pretty light for a modern system.
Additionally, there are suggestions that the threading of the physics test in this version of the benchmark is pretty poor - you certainly wouldn't gain from adding more threads, so the 4C8T Tiger Lake is not going to be handicapped against 8C competitors.
Thirdly, 3DMark is one benchmark that Intel does well in.
Also, 28W TDP versus 15W (maybe 25W cTDP) - but again, the benchmark run is so short even a 15W Ryzen APU might still be at full turbo the full run.
Let's wait for real world gaming reviews. Certainly TigerLake is going to be better than IceLake 4C8T, and maybe considerably so (~25%).
Certainly the integrated GPU is far better than previous to beat an AMD iGPU in anything, so that's a massive stride forward. I'm sure AMD have bigger iGPU APUs in the pipeline, but they'll need more memory bandwidth to really shine.
So its likely NOT clock rate, its graphical architectural improvements. I agree we should wait for real world benchmarks first for the same reason as you, but Intel's GPU department hasnt been as sleepy as their CPU department.
It's worth noting as pointed out, it scales poorly when cores are added meaning it's largely based on single core performance but it does illustrate stronger single core performance with the Intel, as expected with the following taken into consideration.
It's only fair to mention 3D Mark 11 is also a legacy bench as indicated on their own site: Futuremark Legacy Benchmarks - UL Benchmarks and was made back when Intel firmly held the lead in 2011 and single core performance was the emphasis at that time.
How this older bench reacts to a newer OS, newer hardware and drivers could be an issue because that bench is almost 10 years old anyway.
To the point:
It's relevant and yet not so much since games simply aren't the same as they were back in 2011. With more emphasis on multi cored performance vs what was made in 2011 it would take a modern game/bench running a modern OS with modern drivers and so on to truly reveal the performance of each respective to each other as you'd expect to see in a typical modern build.
And you damn well know this was rigged to be boosting at a solid 45W.
AMD already has the fastest IGPU for now, they have bigger fish to fry with products that actually makes more money, which is pretty much any other product, and you cry because an unreleased 28 W part beats a 15 W part. AMD has limited resources, just like every other company, and still definitely less so than Intel.
Such drama. :D
It wasn't ever relevant for AMD's APUs and its not relevant now. The baseline perf is fine for ultra casual gaming and basic usage. It won't get bigger than that, unless you like 20 FPS gameplay at low. So now you get 22 FPS. Wooptiedoo. IGPs are lucky if they get enough speed over time to grow along with that bottom line they could once achieve. Games are getting heavier to run much faster than IGPs develop.
Especially now with RT gaining ground, the IGP won't cut it for gaming. And for everything else it just needs to adhere to new standards, like a 4K output.
Yes, APU's is an important product as they go into pretty much every laptop, and there are more laptops sold than desktops.
But of all those AMD APU laptops, how many of them are running on IGP only, while still pushing that IGP to the max? I'd guess it's less than 0.1 % of all laptops.
So, putting a lot of R&D into something that will make those 0,1 % happy (budget buyers), instead of server CPU's ($$$$$) or any kind of graphics cards (gaming/work/compute) makes zero sense.
Don't get me wrong, Intel can take my money if they have a great product, and maybe they have, but this isn't a good comparison.
you can game Valorant, League of legends, overwatch etc on these, you can do a lot of production stuff on them and a lot of people buy laptops with MX250 style gpu but they are now redundant, you know why?
Cause IGP's now beat em.
On the other hand, making every IGP as fast as a RX570 or GTX1060 3gb is not much point either as 70% doesn't care.
So it's a larger portion than you expect but not enough to warrant big fat designs for the masses, and renoir is just that - Budget chip that just happens to be blazing fast and the fastest laptop cpu out there!
AMD has to prioritize, and IGP isn't where the money is.What makes it a budget chip? Compared to what??? Have you seen any review?
Second of all, just because one chip is taped out doesn't mean they can just slice it down into another chip in one move. You expected them to have RDNA2 ready in an APU only 6 months after the launch of RDNA? And with all the driver bugs that RDNA has been suffering since launch?