Intel Core i7-1185G7 in Review - First Tiger Lake Benchmarks
Test System
We tested an "Intel Reference Design System", which was build for Intel. It is based on the MSI Prestige 14 (in review with Comet Lake) barebone notebook, which will not be available for retail. It is a comparatively slim 14-inch laptop, which is slightly raised when you open the lid and sucks in air from the bottom. The fan noise was moderate during our benchmark tests. The aluminum chassis did get pretty warm, but it was not too bad. We are unfortunately not allowed to post any pictures of the internals.
We could select three different performance levels for our benchmarks:
- Better Battery: 15 Watt PL1 (sustained load, HWInfo: 14.8 W during Blender) and 54W PL2 (28s peak load, HWInfo 54.5W during Blender)
- Better Performance: 28 Watt PL1 (sustained load, HWInfo: 27.9 W duirng Blender) and 54W PL2 (28s peak load, HWInfo: 54 W during Blender)
- Best Performance: 28 Watt or 36 Watt with Dynamic Tuning (sustained load, HWInfo 35.3W during Blender) and 54 W PL2 (HWInfo 54.7 W duirng Blender)
Intel recommended the mode "Better Performance", because it will probably be the most representative for production units. We therefore performed the majority of our tests with 28W and only checked individual tests with 15W and 36W (Dynamic Tuning). Dynamic Tuning maintains the 36W PL1 as long as the internal temperatures and the "skin" temperatures are not getting too high, otherwise it will drop down to 28W.
CPU Performance - Ice Lake with Turbo
The tested Core i7-1185G7 is the high-end model for the launch of the Tiger Lake U generation and offers four processor cores (Willow Cove micro architecture), which can run at up to 4.8 GHz. This means the peak clock is much higher compared to Ice Lake and should result in higher single-core performance. Compared to AMD's current top models, Intel still uses four cores and prefers to use the chip die for a more powerful graphics unit as well as (partially) integrated Wi-Fi as well as Thunderbolt 4.
If we only look at the processor performance, there is an impressive advantage over the old Core i7-1065G7 (like in the Dell XPS 13 9300, for example). Compared to the fast AMD Renoir processors, the test system is sitting between the two 8-core chips Ryzen 7 4700U and 4800U - impressive. The high single-core performance really makes the difference here. The i7-1185G7 can beat all the rivals in the Cinebench Single-Core tests. Even a desktop Core i9-10900K or Ryzen 7 3800XT are slightly slower. If you look at the multi-core benchmarks, like the time consuming Blender test, AMD's 8-core chips still have a big advantage. However, Intel can close the gap to the Ryzen 5 4500U 6-core CPU, which once again shows the performance potential of the four Tiger Lake cores.
Compared to Intel's own Comet lake-U and H-series chips, Tiger Lake once again performs very well. The U-series is clearly beaten - despite 6 cores (which hardly pay off due to the low TDP envelope). The H-series with 8 cores still wins the multi-core tests, but the smaller quad-core chips like the Core i5-10300H are beaten.
Performance Rating - Percent | |
Average Intel Core i9-10980HK | |
Average Intel Core i7-10875H | |
Lenovo Yoga Slim 7-14ARE | |
Intel Reference Design Laptop 35W Dynamic Tuning -4! | |
Average Intel Core i7-10750H | |
Intel Reference Design Laptop 28W | |
Average AMD Ryzen 7 4700U | |
Average Intel Core i7-1165G7 | |
Xiaomi RedmiBook 16 R5 | |
Average Intel Core i5-10300H | |
Average AMD Ryzen 5 4500U | |
Intel Reference Design Laptop 15W | |
Average Intel Core i7-10710U | |
Dell XPS 13 9300 4K UHD | |
Average Intel Core i7-10810U | |
Average Intel Core i7-1065G7 | |
Average AMD Ryzen 3 4300U |
* ... smaller is better
The sustained performance is stable on Intel's reference laptop, and we can neither see the usual drop after the initial benchmark runs in the 28W or 15W mode.
As we have already reported during the launch, Tiger Lake does not use fixed TDP limits anymore, but manufacturers can choose between 15 and 28W (depending on the cooling solution). Including Dynamic Tuning, the CPU can even consume up to 36W for longer periods. Our benchmarks clearly show that these TDP settings have a massive effect on the performance. The single-core performance is not affected that much, because the 15W limit is already sufficient for high clocks, but the CPU performance will drop by 27 % or 31 %, respectively, compared to the maximum setting. This means there will still be significant performance differences from notebook to notebook. You cannot just look at the model designation like i7-1185G7 for the performance you will get.
Blender - v2.79 BMW27 CPU | |
Average AMD Ryzen 3 4300U (868 - 907, n=2) | |
Intel Reference Design Laptop 15W | |
Average Intel Core i7-1165G7 (468 - 1138, n=80) | |
Average AMD Ryzen 5 4500U (533 - 748, n=11) | |
Intel Reference Design Laptop 28W | |
Xiaomi RedmiBook 16 R5 | |
Average Intel Core i5-10300H (479 - 555, n=14) | |
Intel Reference Design Laptop 35W Dynamic Tuning | |
Average AMD Ryzen 7 4700U (407 - 537, n=8) | |
Average AMD Ryzen 5 4600U () | |
Average AMD Ryzen 7 4800U (322 - 374, n=3) | |
Average Intel Core i7-10875H (259 - 535, n=30) | |
Lenovo Yoga Slim 7-14ARE | |
Average Intel Core i9-10980HK (239 - 299, n=13) |
* ... smaller is better
Application Performance - PCMark 10
We still use the regular PCMark 10 for the application performance, which includes free software like LibreOffice, GIMP, Firefox, and Chromium. Thanks to the fast 1TB PM981a SSD from Samsung, the MSI barebone with Tiger Lake shows a great performance. The system only drops back a bit in the Content Creation part, but the other two sub tests produce even higher scores compared to fast gaming notebooks with 8-core CML-H CPUs (i9-10980HK). The overall score is lower on the i7-1185G7 system, but it is alightly ahead of the Lenovo Yoga Slim 7 with Ryzen 7 4800U CPU. The result is an indicator how well Tiger Lake based systems should (or could, respectively) handle office and productive tasks.
PCMark 10 / Score | |
Average Intel Core i9-10980HK (5277 - 6988, n=12) | |
Average Intel Core i7-10875H (4884 - 6739, n=24) | |
Intel Reference Design Laptop 28W | |
Lenovo Yoga Slim 7-14ARE | |
Average AMD Ryzen 7 4800U (5233 - 5383, n=3) | |
Average Intel Core i7-1165G7 (4531 - 5931, n=72) | |
Average AMD Ryzen 7 4700U (4527 - 5307, n=8) | |
Xiaomi RedmiBook 16 R5 | |
Average AMD Ryzen 5 4600U () | |
Average Intel Core i5-10300H (4077 - 5301, n=13) | |
Average AMD Ryzen 5 4500U (4393 - 5024, n=12) | |
Dell XPS 13 9300 4K UHD | |
Average AMD Ryzen 3 4300U () |
PCMark 10 / Essentials | |
Intel Reference Design Laptop 28W | |
Average Intel Core i9-10980HK (9312 - 10829, n=12) | |
Average Intel Core i7-1165G7 (8026 - 11505, n=72) | |
Lenovo Yoga Slim 7-14ARE | |
Average Intel Core i7-10875H (7547 - 10458, n=24) | |
Average AMD Ryzen 7 4800U (9188 - 9637, n=3) | |
Average AMD Ryzen 5 4600U () | |
Dell XPS 13 9300 4K UHD | |
Average AMD Ryzen 7 4700U (8303 - 9559, n=8) | |
Xiaomi RedmiBook 16 R5 | |
Average Intel Core i5-10300H (7148 - 9874, n=13) | |
Average AMD Ryzen 5 4500U (8204 - 9124, n=11) | |
Average AMD Ryzen 3 4300U () |
PCMark 10 / Productivity | |
Intel Reference Design Laptop 28W | |
Average Intel Core i9-10980HK (7641 - 9337, n=12) | |
Average Intel Core i7-10875H (7534 - 9070, n=24) | |
Average AMD Ryzen 7 4800U (7631 - 8091, n=3) | |
Lenovo Yoga Slim 7-14ARE | |
Average Intel Core i5-10300H (6732 - 8058, n=13) | |
Average Intel Core i7-1165G7 (5570 - 9317, n=72) | |
Average AMD Ryzen 7 4700U (5348 - 7945, n=8) | |
Xiaomi RedmiBook 16 R5 | |
Average AMD Ryzen 5 4500U (6400 - 7752, n=11) | |
Dell XPS 13 9300 4K UHD | |
Average AMD Ryzen 3 4300U () | |
Average AMD Ryzen 5 4600U () |
PCMark 10 / Digital Content Creation | |
Average Intel Core i9-10980HK (5177 - 10217, n=12) | |
Average Intel Core i7-10875H (4563 - 9448, n=24) | |
Lenovo Yoga Slim 7-14ARE | |
Average AMD Ryzen 7 4800U (5123 - 5758, n=3) | |
Intel Reference Design Laptop 28W | |
Average AMD Ryzen 5 4600U () | |
Average Intel Core i7-1165G7 (4083 - 6957, n=72) | |
Average AMD Ryzen 7 4700U (4459 - 5495, n=8) | |
Xiaomi RedmiBook 16 R5 | |
Average Intel Core i5-10300H (2928 - 6003, n=13) | |
Average AMD Ryzen 5 4500U (4189 - 4808, n=11) | |
Average AMD Ryzen 3 4300U () | |
Dell XPS 13 9300 4K UHD |
GPU Performance - Xe Architecture promises a lot of performance
The Tiger Lake processors are the first chips with the new Xe graphics generation. Intel only calls it "Intel Iris Xe Graphics" on our Core i7-1185G7, and it is supported by 16 GB LPDDR4x-4266 RAM in a dual-channel configuration. However, the iGPU does not have its own graphics memory (except for caches). Intel advertises a big performance advantage over the Ice Lake Generation, so it is a big improvement after a long time of familiar iGPUs.
The new Xe GPU performs extremely well in the synthetic graphics benchmarks. It is on par with a dedicated AMD Radeon Pro 560X in the Apple MacBook Pro 15 and the fastest Ryzen iGPUs (like the RX Vega 8 in the Ryzen 4000) are clearly beaten. Mobile graphics cards from Nvidia like the GeForce MX350 are also much slower and the score is more comparable to the new MX450.
The compute performance of the GPU in V-Ray is not that impressive, but still on par with the older GeForce MX250/MX330. The AMD GPUs are more powerful in this test.
Basemark GPU 1.2 - 1920x1080 Vulkan Medium Offscreen | |
Average NVIDIA GeForce MX450 () | |
Intel Reference Design Laptop 28W | |
Apple MacBook Pro 15 2018 (2.6 GHz, 560X) | |
Apple MacBook Pro 15 2018 (2.6 GHz, 560X) | |
Intel Reference Design Laptop 15W | |
Average AMD Radeon RX Vega 8 (Ryzen 4000/5000) (76.9 - 289, n=9) | |
Average AMD Radeon RX Vega 7 (104.6 - 289, n=10) | |
Average Intel Iris Plus Graphics G7 (Ice Lake 64 EU) () | |
Average AMD Radeon RX Vega 6 (Ryzen 4000/5000) (82.6 - 104.6, n=2) | |
Xiaomi RedmiBook 16 R5 |
V-Ray Benchmark Next 4.10 - GPU | |
Average NVIDIA GeForce GTX 1050 Mobile () | |
Average AMD Radeon RX Vega 8 (Ryzen 4000/5000) (31 - 83, n=47) | |
Lenovo Yoga Slim 7-14ARE | |
Average AMD Radeon RX Vega 7 (14 - 70, n=47) | |
Average AMD Radeon RX Vega 6 (Ryzen 4000/5000) (30 - 52, n=15) | |
Intel Reference Design Laptop 35W Dynamic Tuning | |
Average NVIDIA GeForce MX350 (33 - 41, n=10) | |
Average NVIDIA GeForce MX450 (28 - 40, n=17) | |
Xiaomi RedmiBook 16 R5 | |
Intel Reference Design Laptop 28W | |
Intel Reference Design Laptop 28W | |
Average NVIDIA GeForce MX250 (24 - 34, n=5) | |
Intel Reference Design Laptop 15W | |
Average Intel Iris Plus Graphics G7 (Ice Lake 64 EU) (17 - 28, n=5) |
Gaming Performance
Thanks to the much higher GPU performance, the iGPU is also much better suited for games compared to previous iGPUs. We tested a couple of titles and even Doom Eternal ran smoothly with good frame rates, but we did not record the result due to a graphics error caused by the beta drivers.
The demanding title "The Witcher 3" shows the performance of the iGPU very well. It is almost on par with the GeForce GTX 1050 Max-Q and playable at 1080p and high settings at 35 FPS. The GeForce MX350 is much slower, but the situation changes in GTA 5, where the MX350 can manage a clear lead. The AMD Vega 7 in the Acer Swift 3 SF314 is clearly beaten both times.
Our test of an older version of DOTA 2 also places the Xe behind the MX350 and on par with the Vega 7, while the Vega 8 in the Slim 7 is even a bit faster.
Real-world gaming tests once again show that the performance depends on drivers, optimizations, and graphics memory. You cannot always say that dedicated GPUa like a MX350 with fast GDDR5 VRAM are always beaten, but you can see the performance potential of the Iris Xe and that it makes entry-level dGPUs unnecessary.
Verdict
Our initial benchmarks with a reference system from Intel with the new high-end Tiger Lake-U chip, the Core i7-1185G7, are very promising and the chip leaves an excellent impression. Intel shows what a well-configured version with sufficient cooling is capable of. The performance is very impressive for a quad-core CPU and both Intel's as well as AMD's 6-core chips are beaten. The fast Ryzen 7 4800U with 8 cores is still ahead in many multi-core tests for this TDP segment, but the real highlight is the excellent single-core performance thank to high clocks. The efficient mobile processor can even beat high-end desktop models like the i9-10900K. This also results in excellent results in application benchmarks, which is shown by PCMark 10.
The integrated Iris Xe GPU is another highlight. Intel manages to beat dedicated entry-level GPUs like the GeForce MX330 or 350 in synthetic tests, but we have to wait and see how the real-world gaming performance will be. Our initial tests are promising in this respect.
The performance of Tiger Lake is still decent at 15 Watts, so we also hope that there will be fast 6W models. However, the model designation alone does not show the performance, but it really depends on the TDP configuration and the cooling performance.