Intel "Tiger Lake" vs. AMD "Renoir" a Pitched Battle on 3DMark Database

[btarunr]

Intel's 11th generation Core i7-1165G7 "Tiger Lake-U" processor armed with 4 "Willow Cove" cores and Gen12 Xe graphics fights a pitched battle against AMD Ryzen 7 4800U "Renoir" (8 "Zen 2" cores and Radeon Vega 8 graphics), courtesy of some digging by Thai PC enthusiast TUM_APISAK. The 4800U beats the i7-1165G7 by a wafer-thin margin of 1.9% despite double the CPU core-count and a supposedly advanced iGPU, with 6331 points as against 6211 points of the Intel chip, in 3DMark 11. A breakdown of the score reveals fascinating details of the battle.

The Core i7-1165G7 beats the Ryzen 7 4800U in graphics tests, with a graphics score of 6218 points, against 6104 points of the 4800U, resulting in a 1.9% lead. In graphics tests 1, 2, and 3, the Gen12 Xe iGPU is 7.3-8.9% faster than the Radeon Vega 8, through translating to 2-4 FPS. The Intel iGPU crosses the 30 FPS mark in these three tests. With graphics test 4, the AMD iGPU ends up 8.8% faster. Much of AMD's performance gains come from its massive 55.6% physics score lead thanks to its 8-core/16-thread CPU, which ends up beating the 4-core/8-thread "Willow Cove," with the 4800U scoring 12494 points compared to 8028 points for the i7-1165G7. This CPU muscle also plays a big role in graphics test 4. This battle provides sufficient basis to speculate that "Tiger Lake-U" will have a very uphill task matching "Renoir-U" chips such as the Ryzen 7 4800U, and the upcoming Ryzen 9 4900U (designed to compete with the i7-1185G7).



View at TechPowerUp Main Site

 

[Flanker]

I thought we should shift away from benchmarks

 

[ARF]

The Renoir CPU score in Physics is 55% higher.
Intel has only slightly faster graphics units which goes to say that AMD's Vega is too old and needs a replacement with something faster and more modern soon.

 

[watzupken]

I am not expecting Intel to pull a miracle with Tiger Lake, and this kind of gives an indication of what to expect. However, there are a lot of unknowns as to the test bed for the respective processors in this leak, and in particular, it will be interesting to know what are the memory speed used for each of the test systems, which is expected to significantly shift performance for the GPU.

 

[londiste]

The Renoir CPU score in Physics is 55% higher.

True but keep in mind this is with 100% more cores.

 

[R0H1T]

And a 15W TDP, can't say much about TGL atm.

 

[Valantar]

This looks promising, though previous data from Intel's GPUs (and their drivers) shows that 3DMark is one of very few applications where they perform well. Typically their Iris Pro/Plus iGPUs perform well in 3DMark, with performance outside of a few popular esports titles (DOTA and the like) not living up to the relative performance in 3DMark. Still, the upcoming years look really good for iGPUs across the board. Looking forward to what comes next, particularly if AMD shows up with RDNA2 and LPDDR5 next year - that would be interesting indeed.

 

[Tom Yum]

This looks promising, though previous data from Intel's GPUs (and their drivers) shows that 3DMark is one of very few applications where they perform well. Typically their Iris Pro/Plus iGPUs perform well in 3DMark, with performance outside of a few popular esports titles (DOTA and the like) not living up to the relative performance in 3DMark. Still, the upcoming years look really good for iGPUs across the board. Looking forward to what comes next, particularly if AMD shows up with RDNA2 and LPDDR5 next year - that would be interesting indeed.

Exactly, it always seems like Intel just doesn't invest effort into driver development for actual games, just benchmarks. No point being 15% ahead in 3DMark if you are 30% behind in actual games people play compared to your competition.

 

[Dante Uchiha]

Exactly, it always seems like Intel just doesn't invest effort into driver development for actual games, just benchmarks. No point being 15% ahead in 3DMark if you are 30% behind in actual games people play compared to your competition.

That's the point. They say that vega 8 was outperformed based on the 3dmark result.. It's pretty dishonest.

 

[londiste]

The result here is from 3DMark 11 which is very old. By the way, Graphics Test 4 seems to be heavy on tessellation.

There has been a difference in allocation of different types of resources in iGPUs - Intel does better where fixed-function stuff is more of a bottleneck, AMD does better where compute is more of a bottleneck. It is not always that straightforward and depends on what the test, benchmark or game relies on most. With that being said, we do not know very well how Xe performs across larger amounts of applications - it seems to have more focus on shaders/compute so it could be a good contender.

About this particular comparison - there are two things missing to really know how comparable they are:
- RAM speed which affects iGPU results a lot since both the iGPUs here are bandwidth-starved)
- Power usage. Both CPUs nominally being 15W does not mean they are configured as such in this particular instance

 

[Valantar]

That's the point. They say that vega 8 was outperformed based on the 3dmark result.. It's pretty dishonest.

Dishonest? By who? The person testing a preproduction chip on 3Dmark and not publishing it? The one discovering this and publishing it? Or the one reporting on the find? Neither of these are doing anything even remotely dishonest. Nobody is claiming that this single preproduction chip test result is the definitive measure of tiger lake performance. Nobody is claiming that this equals faster game performance than the 4800U. Sure, the reporting could stand to add a tad of nuance, but the omission is entirely understandable as there is no other data to base a comparison on.

 

[Dante Uchiha]

Dishonest? By who? The person testing a preproduction chip on 3Dmark and not publishing it? The one discovering this and publishing it? Or the one reporting on the find? Neither of these are doing anything even remotely dishonest. Nobody is claiming that this single preproduction chip test result is the definitive measure of tiger lake performance. Nobody is claiming that this equals faster game performance than the 4800U. Sure, the reporting could stand to add a tad of nuance, but the omission is entirely understandable as there is no other data to base a comparison on.

Anyone who uses this result to create hype, I see it a lot on tech sites, "Intel finally got it. Tiger Lake has graphic power to beat renoir."

See the reality, IceLake beats Picasso in synthetic tests and some CPU bound games, but loses in 90% of games.

About this particular comparison - there are two things missing to really know how comparable they are:
- RAM speed which affects iGPU results a lot since both the iGPUs here are bandwidth-starved)
- Power usage. Both CPUs nominally being 15W does not mean they are configured as such in this particular instance

Yeah, I haven't seen any renoir laptop using LPDDR4X @ 4266. I suppose it's due to cost, people(OEM) still think amd's is only for cost effective products... although it wins the competitor in all respects: https://www.notebookcheck.net/Intel...r-RTX-2080-laptops-anytime-soon.462278.0.html

 

[ARF]

True but keep in mind this is with 100% more cores.

When in the same TDP, it's either less cores at higher clocks or more cores at lower clocks.
This said - do you honestly believe that an 8-core Tiger Lake in 15-watt will show any better results?

These mobile chips normally get better bins - for instance my Polaris 21 GPU is rated ASIC 60.0%.
Other cards I have seen with lower 40-45% ASICs.

The result here is from 3DMark 11 which is very old. By the way, Graphics Test 4 seems to be heavy on tessellation.

There has been a difference in allocation of different types of resources in iGPUs - Intel does better where fixed-function stuff is more of a bottleneck, AMD does better where compute is more of a bottleneck. It is not always that straightforward and depends on what the test, benchmark or game relies on most. With that being said, we do not know very well how Xe performs across larger amounts of applications - it seems to have more focus on shaders/compute so it could be a good contender.

About this particular comparison - there are two things missing to really know how comparable they are:
- RAM speed which affects iGPU results a lot since both the iGPUs here are bandwidth-starved)
- Power usage. Both CPUs nominally being 15W does not mean they are configured as such in this particular instance

I think it's very rare if you see an OEM reconfiguring the chips to run at lower or higher TDP...
Most of the time they are just left in their original and default power envelopes.

 

[HugsNotDrugs]

I'm happy to see Intel take graphics seriously after being stalled so long on the HD620 /HD630.

 

[Valantar]

I think it's very rare if you see an OEM reconfiguring the chips to run at lower or higher TDP...
Most of the time they are just left in their original and default power envelopes.

These days it's actually quite common. There are lots of 25W U-series configurations out there. Of course there are more 15W ones, but 25W isn't the rarity it once was.

Anyone who uses this result to create hype, I see it a lot on tech sites, "Intel finally got it. Tiger Lake has graphic power to beat renoir."

See the reality, IceLake beats Picasso in synthetic tests and some CPU bound games, but loses in 90% of games.

That is true, but Ice Lake is not Tiger Lake, and Intel are bound to put more effort into their driver development with the launch of Xe. We obviously don't know how that will pan out, but we can't assume a 1:1 correlation with the past either. We simply don't know.

 

[medi01]

Bad news for NV, as it means Intel no longer needs MX for its notebooks graphics not to suck.

Xe barely beating Vega... thank god, Raja is no longer at AMD...

 

[ARF]

I'm happy to see Intel take graphics seriously after being stalled so long on the HD620 /HD630.

Yes, this is kind of a complex for Intel. When they had been constantly beaten in the graphics department and they start focusing mainly on it.

 

[Punkenjoy]

having a future product matching performance of an actual product is like the bare minimum.