Intel Core Ultra Processors – The Heart of Every PC
The family of Intel Core Ultra Processors utilize a 3D Hybrid architecture that includes the integration of a CPU, GPU and a dedicated NPU engine for AI acceleration all on the same die. Initially introduced as a mobile processor, the Intel Core Ultra 100 Series code named Meteor Lake, brought to market a powerful, energy efficient AI optimized processor that could run ML, SLM, and inferencing applications on a client platform.
The Intel Core Ultra 200 Series features several new processors designed to meet the specific needs of various compute platforms including desktop AI PCs (Arrow Lake S), performance AI notebooks (Arrow Lake H/HX) and premium class AI notebooks (Lunar Lake V).
Intel Core Ultra 200S (Arrow Lake S for Desktop PCs) Technical Notes
Click on the circles below to learn more about each feature.
With the new Lion Cove P-Cores, Intel reduced the Max Turbo Speed by 100MHz but increased the base frequency by 500-700MHz. The result of this change is a reduction in power demand which lowers the overall system temperature by as much as 13 C.
With the new Skymont E-Cores, Intel eliminated Hyper-Threading from this family of processors but they increased the Max Turbo mode speed by 200 to 600MHz and increased the base frequency by 600MHz to 1GHz which actually allows for multi-threaded applications to run more efficiently on the new E-Core processors.
Intel has completely redesigned the cache hierarchy to boost performance and improve efficiency. Arrow Lake now has a maximum of 76MB of cache which includes 40MB of total L2 cache and 36MB of total L3 cache. The P-Cores now have 3MB of L2 cache for each processor which is an increase from 2.5MB on the 14th Gen Processor. Additionally, the E-Cores have now been grouped in to quad clusters with each cluster having access to 4MB of shared L2 cache specifically for the E-Cores. The L3 cache has been redesigned with a maximum of 36MB of shared cache which now gives the E-Cores access to L3 cache.
Intel Arrow Lake S supports standard DDR5 UDIMMs as well as new CUDIMMs which is Clock Unbuffered Dual In Line Memory. This new memory includes an integrated CKD (Clock Driver) that is an IC located on the memory module that improves stability at higher speeds by regenerating the clock signal used by chips. This allows for much higher memory frequencies reaching up to 8000MHz. CUDIMM uses the same 288 pin DDR5 UDIMM slot so it is compatible with standard UDIMM slots.
The Intel Core Ultra 200S uses the same NPU as Meteor Lake (Intel Core Ultra 100) which is NPU 3.0 that combines with the CPU and GPU to deliver up to 36 platform TOPS (13 TOPS from the NPU). This differs from Lunar Lake which uses the NPU 4.0 and has a higher TOPS rating of 48.
The Intel Core Ultra 200S uses the same Xe Intel Graphics engine as Meteor Lake (Intel Core Ultra 100) delivering a 2x performance gain over the previous Intel 14th Gen Core Processor desktop solution. This differs from Lunar Lake which uses the Intel Battlemage Xe2 GPU.
Although the Intel Core Ultra 200S uses the LGA 1851 socket, LGA 1700 coolers (Fan/Liquid) are compatible with the new socket.
Intel Core Ultra 200S – Arrow Lake S (Desktops)
ASI has authorized partnerships with strategic manufacturers including CPU, motherboard, case, power supply, memory, storage, GPU and networking all of which are essential to system integrators, VARs, MSPs, and vertical resellers that build custom compute solutions. This unique breadth of products ensure that ASI can align with our customers to create more value, build markets, and strengthen our relationships.
The Intel Core Ultra 200 Processor family, Arrow Lake and Lunar Lake, incorporate new designed Lion Cove P-Cores and Skymont E-Cores, along with increased cache and a new cache hierarchy all of which result in improved IPC, lower power consumption, and a decrease in system temperature. Although Arrow Lake and Lunar Lake share some similarities, Lunar Lake is engineered to target the mobile market so it incorporates technologies specifically designed to optimize power efficiency and performance.
