Intel gives us some more information and a first look at their next generation Tiger Lake mobile CPU this week on HotChips 32. The company unveiled last week all of its Tiger Lake processors with a full overview of the 11th generation mobility lineup that you can check here.
Intel Introduces 11th Gen Tiger Lake CPU Dies, Gives Us A Detailed Look At Everything New With Its Next-gen Mobility Processors
The Intel Tiger Lake CPUs are based on the 10nm ++ process node and have the latest Willow Cove cores from the company. The Willow Cove architecture is based on the same foundations as the last generation Sunny Cove architecture. It can be seen as a refinement of Sunny Cove in every possible way, but at the same time, Willow Cove cores also have a redesigned chip hierarchy for faster performance performance.
The 10nm ++ process node is referred to by Intel as the 10nm SuperFin transistor design. Intel itself claims that the SuperFin process delivers the same increase in performance as a proper node operation. Intel stated that it had achieved about 17-18% performance increase over the standard 10nm process node on 10th generation Ice Lake chips, while delivering much faster frequencies. You can read our full article here about the Intel SUPERFin transistor technology.
Intel 11th Gen Tiger Lake CPU block diagram
Coming to the newly unveiled details, Intel has presented both a block diagram and a die-shot of its Tiger Lake CPUs. The block diagram mentions everything that is new about Ice Lake chips & also the parts that are kept the same.
Let’s first look at the similarities between Ice Lake and Tiger Lake CPUs. There are a few that include the DDR4 memory controller that supports 3200 MHz speeds with up to 64 GB capacity. Other controllers such as SGX, Fuse, JTAG, SVID, and OPIO are also comparable to those used on Ice Lake chips. The display USB Type-C controller also includes DP 1.4 which was also last gen.
Tiger Lake CPUs have only a few blocks that are kept similar to Ice Lake. The rest of the chip is all brand new and that can be clearly seen in the block diagram. The main new additions to Tiger Lake CPUs are the Willow Cove cores, Xe LP graphics, and media engine, 12 MB L3 (last level cache), a new display engine with support for 8K display (4 pipelines / 64 GB / s read bandwidth), IPU6 with 6 camera sensors (video up to 4K90, initially at 4K30) and added support for LPDDR5 memory with speeds up to 5400 MHz (32 GB capacity). The rest of the blocks have received major and minor upgrades across Ice Lake.
Intel 11th Gen Tiger Lake CPU Die Shot With Annotations (Image Credit: @Locuza_):
The Intel Tiger Lake CPUs will use a dual-ring interconnect. The Tiger Lake-U CPU array will have up to 4 Willow Cove cores and an Xe-LP integrated GPU engine with 96 EUs and 768 cores in total, running at a frequency of about 1300 MHz. The Tiger Lake CPUs themselves are expected to reach short burst frequencies of up to 5 GHz, which is impressive for its 10nm ++ process node.
The 11th Generation Tiger Lake CPU Family – Built for mobile laptops and gaming notebooks
The Intel Tiger Lake CPUs will be referred to as the 11th generation Core family and would be held exclusively for laptops and gaming notebooks. The lineup would come in three flavors that would include Tiger Lake-Y, Tiger Lake-U, and Tiger Lake-H. There have been several leaks for Tiger Lake-Y and Tiger Lake-U Processors that are being tested internally by various OEMs and laptop manufacturers that would integrate the CPUs into their next-generation devices.
De Tiger Lake-U family would consist of 15-28W TDP CPUs and would have 4 cores and 8 wires, although with much higher clock speeds with boosts closer to 4.50 GHz. These CPUs would also have GT2 tier, Gen 12 Xe GPUs, and would come in the UP3 (BGA 1499) package. Tiger Lake-U will be the first to hit notebooks on September 2 as previously reported.
The manufacturer’s documents mention LPDDR5 support for the named family, while the rest of the setup will use both LPDDR4 (X) memory. We’ve seen laptops with LPDDR4 and LPDDR4X memory in addition to the Tiger Lake-U CPUs, so we can expect multiple configurations with few higher-end & the more expensive variants that offer LPDDR5 support.
Intel Tiger Lake vs. AMD Renoir Mobility CPU Comparison:
| CPU family name | Intel Tiger Lake-U | AMD Renoir U-Series |
|---|---|---|
| Family Branding | Intel 11th Gen Core (U Series) | AMD Ryzen 4000 (U-Series) |
| Node processing | 10nm | 7pm |
| CPU architecture | Willow Cove | Zen 2 |
| CPU cores / wires (max.) | 4/8 | 8/16 |
| Max CPU clock | TBD (Core i7-1185G7) | 4.2 GHz (Ryzen 7 4800U) |
| GPU Core Architecture | Xe Graphics Engine | Vega improve 7nm |
| Max GPU Cores | 96 EU (768 cores) | 8 CUs (512 cores) |
| Max GPU clock | 1300 MHz? | 1750 MHz |
| TDP (cTDP Down / Up) | 15W (12W-28W) | 15W (10W-25W) |
| Start | Mids 2020 | March 2020 |
The Intel Tiger Lake-Y family would consist of 4.5-9W TDP CPUs and would have a maximum of 4 cores and 8 wires. The GPU side would include a GT2 tier, Gen 12 Xe GPU. The Tiger Lake-Y Processors will come in the UP4 (BGA 1598) package. The Tiger Lake-Y family would exclusively support LPDDR4X memory.
Then there is the high performance Tiger Lake-H arrangement consisting of a maximum of 8 core and 16 thread chips based on the new Willow Cove architecture. The CPUs would carry a maximum of 34 MB of cache, which is 24 MB L3 (3 MB L3 per core) and 10 MB L2 (1.25 MB per core). Tiger Lake CPUs will come with an asymmetric 48/32 KB L1 cache and will fully support AVX2 and AVX-512 instructions. Tiger Lake-H CPUs would additionally have two-level memory (2LM) and SGX (Software Guard Extensions). Intel’s Tiger Lake-H family would support DDR4 speeds up to 3200 MHz.
Intel Tiger Lake-H vs. AMD Ryzen H-Series CPU Performance with High Performance:
| CPU family name | Intel Tiger Lake-H | AMD Renoir H-Series | AMD Cezanne-H Series |
|---|---|---|---|
| Family Branding | Intel 11th Gen Core (H Series) | AMD Ryzen 4000 (H-Series) | AMD Ryzen 5000 (H-Series) |
| Node processing | 10nm | 7pm | 7nm + |
| CPU architecture | Willow Cove | Zen 2 | Zen 3 |
| CPU cores / wires (max.) | 8/16 | 8/16 | 8/16 |
| Max CPU clock | TBD | 4.3 GHz (Ryzen 9 4900HS) | TBD |
| GPU Core Architecture | Xe Graphics Engine | Vega improve 7nm | Vega improve 7nm |
| Max GPU Cores | 96 EU (768 cores)? | 8 CUs (512 cores) | 8 CUs (512 cores)? |
| Max GPU clock | TBD | 1750 MHz | TBD |
| TDP (cTDP Down / Up) | 35W (65W cTDP) | 35W (45W cTDP) | 35W (45W cTDP) |
| Start | Q1 2021 | Q2 2020 | Q2 2021? |
Intel also detailed that the Tiger Lake mobile CPUs can scale all the way from 9W to 65W. We know that Intel prepares Tiger Lake CPUs for at least three segments. These include Tiger Lake-Y, Tiger Lake-U, and Tiger Lake-H. The Tiger Lake-U CPUs will first follow the market shelves in various OEM designs next month, followed by Tiger Lake-Y later this year. The Tiger Lake-H lineup is scheduled for the first half of 2020 and TDPs are expected to range from 35W to 45 / W65W (cTDP up). All Tiger Lake CPUs will support PCIe Gen 4, USB 4 and Thunderbolt 4 functionalities.
The 10nm Tiger Lake CPUs will hit AMD’s 7nm Zen 2-based Ryzen 4000 ‘Renoir’ family when they are officially launched on September 2nd.
