015-Shreyas-Derashri.pdf

1、Revolutionizing RISC-V adoption Integrated Solution GPU+CPU+AINew Compute platforms Proven benefit from system approach Apple SiliconEfficient use of Memory Scaling challenge Optimisation at a system level Bandwidth and system data congestionSystem level design drives ecosystem Faster adoption of RI

2、SC-V87%New Compute platforms Proven benefit from system approach Apple SiliconM1 ChipArea(2 mm)PercentageTotal Die118100%Total CPU2118%Total GPU 23.520%Total AI6.66%SLC65%49%Compute Subsystem accounts for Half of the SoC areaGraphics is critical to compute platform In 2022,Semiconductors containing

3、GPUs contributed$96B to the overall Semiconductor market.This is 44%of the$216B market excluding memory,optoelectronics,Analog,connectivity,RF,discrete components,power management and sensors Source:GartnerGPU in 44%of Semiconductor TAMRISC-V will not disrupt the Android market without a capable and

4、 credible GPUGPU is the platform for AIGPU is the right architecture for AIGPU is critical to AI CPU does not meet the level of compute density needed GPU programming model is established and deals well with varied parallel tasks(e.g.NVIDIA with CUDA,UXL and SYCL)Algorithms move too quickly for dedi

5、cated hardware alone need some flexibility with a direct programming model Array-like approaches do not deliver a direct programming model and struggle with multiple workloadsYour one-stop solution for embedded ComputeGPU,CPU and AI#1Position in GPU for Automotive and Mobile13 BillionCumulative chip

6、 shipments with Imagination IPs84%Employees involved in R&D and engineering3300Patents and applications300mDCD GPUs shipped p.a.$94.5bnAutomotive(2025E)$96.8bnConsumer(2025E)$62bnMobile(2025E)4 Key markets33%2019-21 Revenue CAGR$153mn2021 RevenueGeographyCustomerProductSegmentMarket LeadershipAdvanc