Intel Tiger Lake (Xe) Graphics Details Slide Leaks
Another day, another leak. This round information on Intel leaked like melting ice cream, yep Ice Lake in specific, well actually Tiger Lake, but that's more or less the same. If you can remember it, that's the dreaded 10nm node that is causing so many issues. Intel is continuing though, on the ultra-thin and ultra-low-power designs.
So first things first, Tiger Lake is an Intel CPU microarchitecture based on the third-generation 10nm process node (named "10nm++"). It will replace Ice Lake, representing the Optimization step in Intel's Process-Architecture-Optimization model. In specifically that segment you can expect Tiger Lake processors, SoC's we should say up-to four cores with TDPs ranging from 9 running to 28 Watt. Tiger Lake is going to be a 4+2 design, and that means 4-cores and 2 Gen12 iGPU clusters. Tiger Lake would feature up to 768 Shading Units. BTW the slides confirm VP1 decoding support.
In the ultra-thin and ultra-low-power it could take on 4-core Renois parts, AMD's upcoming Ryzen 4000 APUs based on ZEN2.
Intel Tiger Lake | Intel Ice Lake | AMD Renoir | |
---|---|---|---|
Die Size | 146 mm2 | 122 mm2 | 156 mm2 |
Core Design | Willow Cove | Sunny Cove | Zen2 |
Node | 10nm | 10nm | 7nm |
CPU | 2 / 4 | 2 / 4 | 4 / 6 / 8 |
GPU Shaders | 384 / 512 / 640 / 768 | 256 / 384 / 512 | 384 / 448 / 512 |
GPU Architecture | Gen12 (Xe) | Gen11 | GCN 5.0 (Vega) |
PCIe Gen | 4.0 | 3.0 | 4.0 |
Thunderbolt Gen | 4.0 | 3.0 | 3.0 |
TDP | U (28/15W) / Y (9W) | U (28/15W) / Y (9W) | U (15W), H (45/35W), G (65W) |
Integrated Xe Intel Tiger Lake GPU outperforms Vega at 7nm from Ryzen 4000 for notebooks - 03/25/2020 10:07 AM
Intel Lake Tiger-U processors are yet to be found on the market, but Intel is testing them, and tests have appeared in 3DMark that reveal the first performance numbers of these chips and their integra...
Ivy Bridge heat problem is caused by Intel TIM choice - 05/13/2012 08:52 AM
Japanese website Impress Personal Computer Watch has proven what many have suspected. They
Senior Member
Posts: 231
Joined: 2019-04-15
It is rather amusing to see OpenGL v4.5 listed as "Latest APIs Support". It was released in 2014 while OpenGL v4.6 is from 2017 and it is not even listed. There is no mention of Vulkan either which is puzzling for support of latest APIs. It would be a much better option than OpenGL of any version.
Senior Member
Posts: 118
Joined: 2013-11-10
I am curious if the high end Intel IGPUs will be faster then the fastest AMD IGPU. Judging by the shader count they will, but frequency is the other factor in the calculation.
Senior Member
Posts: 3117
Joined: 2016-08-01
shader count between different architectures they say literally nothing see for example Cuda vs Gcn vs navi 1 shader counts .... they are not 1 to 1 ....at all
Senior Member
Posts: 8408
Joined: 2008-07-31
Even if the frequency is higher and the shader count is higher that doesnt mean the performance will be higher. Different architects mean comparing those two things itself is basically meaningless numbers.
Senior Member
Posts: 3299
Joined: 2006-04-25
It looks like both the Intel and AMD solutions are going to be really close to entry level, bargain video cards. Close enough where that entire segment may disappear entirely?